SoftBank Group Aktie

Thank you very much for waiting, everyone. Now we would like to start the SoftBank Group Corp. earnings results announcement for the first quarter of the fiscal year ending March 2027. First of all, I would like to introduce today's participants. From left, we have Yoshimitsu Goto, Board Director and CFO; Kazuko Kimiwada, Corporate Officer, Senior Vice President and CAO; Mark Agne, Executive Managing Partner, SB Investment Advisers and SB Global Advisors. And Jason Child, Executive Vice President and CFO. Today's announcement is live broadcast over the Internet. Now I would like to invite Yoshimitsu Goto to present to you the earnings results and business overview.

Good afternoon, everyone. Thank you very much for your time today. I am Yoshimitsu Goto, CFO of SoftBank Group. First of all, I would like to express my deepest condolences to those who lost their lives in the recent Kumamoto earthquake. I also extend my heartfelt sympathies to everyone affected by the disaster, including those staying in evacuation shelters or living away from their homes under difficult and uncertain circumstances.

As SoftBank Group, we would like to support as much as possible. So SoftBank Group, SoftBank Corp. and LY Corporation, Fukuoka SoftBank Hawks will donate and also encouraging public contributions. We will continue to support the recovery and reconstruction effort through the services provided by our group companies to help those affected regions regain their daily lives as quickly as possible. Now that I would like to go into our earnings results.

So towards June, compared to March close, market has been dynamically moving. As you know, SoftBank Group is an investment company. We can be affected by the market. And investment portfolio valued by the market is basically reflecting to our results. As a result, at the end of June, our net asset value which is the total value of our investment portfolio, less debt -- less interest-bearing debt gives you a net asset value, JPY 72.3 trillion. So this is a record high number. Our performance is actually -- is described by this value. I believe that this is the most important indicator to see and value our enterprise value. This is still a rough estimate, but the pro forma basis, the latest information-wise net asset value is about JPY 58.3 trillion. So it is a fact that it has been declined from JPY 72.3 trillion. But actually, if you see the trend in the past, JPY 58.3 trillion is already great numbers to be achieved.

So I would like to also show this together with our history. So this is the kind of the record that we've been experiencing. And also this itself actually shows you how much potential do we have for further growth in the future. So as I mentioned that every quarter, there are good times, bad times. But at the same time, we do not really become too happy or too sad about short-term fluctuation. We always look at the mid- to long-term period and how are we building the base for the further growth and how that's going to be reflected to the future growth.

And our investment strategy is pretty much equals to our business strategy. So as an AI platform provider, what do we need to do? So we always kind of think backwards once we set the target and then that we would like to discuss what we need to do. Why are we addressing AI infrastructure? Because if you don't address AI infrastructure, you cannot see the enhancement or acceleration of the speed of AI model. OpenAI and other hyperscalers, those AI model players that we would like to accelerate the speed of those people or those companies. And by addressing AI infrastructure, that can also accelerate the growth of Arm, for example.

Arm, as you know, they actually design the core and heart of the all the chip, and they have dominant market shares for that. And by having more and more chip to be shipped, then the basic business model will grow. And not only that, recently, Arm has announced to also address their in-house chip with the high-quality, high capable chip will be used in many AI models, AI infrastructure. And because we've been telling market that we try to become a company to change people's lifestyle. So lifestyles or the working styles to be changed, physical AI is going to be the very focused area that we pay a very good attention to.

This area is actually just the beginning. And there is not much many company making profit or we don't really know who will be the winner for the future in this area either. So that makes us even more excited to see. And each has a kind of a correlation to each other. So that is why that we address these 4. So only for the quarter, but we do have some developments here, which I will be explaining in the later pages.

So to become a real #1 AI platform provider is something that we are trying to achieve. And that's something that Masa is very much putting his effort into. And this is the financial highlights for the first quarter. First, JPY 1.8 trillion investment gain. So this is increased by JPY 1.3 trillion year-over-year. And the net income, JPY 347 billion. Second, as I mentioned in the previous page, net asset value, JPY 72.3 trillion, highest ever compared to the past numbers. And third, follow-on investment in OpenAI, we have invested JPY 20 billion in this first quarter. Towards October, we have remainders of JPY 10 billion, which is committed. And fourth, for the first quarter period, we have announced plans to develop 5 gigawatt AI data center in France, which was kind of highlights for our quarter.

And now we would like to go into consolidated results. You see the numbers in each respective lines. But the investment gain actually increased dramatically compared to the previous quarter -- excuse me, the previous fiscal year. So -- and net income is another important number, JPY 347 billion. So I will explain you the trend from the past, but something I'm missing from here, which is the net asset value. So I would like to also add once again that the net asset value is also an important KPI for our company. Investment gain or loss on the net income, how it has been developed. So for April to June first quarter, JPY 1.8 trillion, mainly due to actually Intel.

Last year, we've been paying attention to the semiconductor business in the United States, and we invested in Intel. As a result, the first quarter, it has been making quite a good increase. And also ByteDance invested by Vision Fund, the fair value increased actually has another -- it was another contributor for the investment gains for the quarter. One more important indicator from the financial point of view is, as I mentioned, net asset value once again, end of June, JPY 72.3 trillion. Compared to a quarter before, that has increased dramatically. And also loan-to-value to measure your safetiness -- this is the net debt over equity value of holding. At the end of June, 13%, which is also a very safe level or even too safe. So we may need to invest more probably.

That's how we believe being kind of tightened the loan-to-value. And cash position, JPY 3.5 trillion back in the March end and has decreased a little bit to JPY 2.3 trillion, but because this is due to the redemptions for the maturity of the bonds and also for the -- some investment activities. But still, it does maintain the necessary cash positions based on the financial policy we have. Net asset value, quarterly basis, as you can see, highest ever, JPY 72.3 trillion. Loan-to-value, if you look at the trend from the past, as a financial policy, we have 25% as one threshold. But at the same time, I would also like to emphasize 25% itself is already safe enough, but we make it even safer of our business itself because our company has quite a large investment from our own balance sheet, which is quite rare and unique compared to the other global investment company.

So I believe this is important for us to share with you the safetiness of our business using loan-to-value. And why 25%, you may ask, is because we do need to see some kind of threshold -- where do we think that the safety level are there. Of course, that the financial people may understand well about the loan-to-value, how much will be the safe level. But for the public to understand, even that the market can go up volatile, but still as long as we keep this level, we will be able to keep a company at a safe level.

Cash position, a little bit of decrease I mentioned earlier. But in our financial policy, our criteria is the maintaining a cash position about 2 years' worth of bond redemption. And over 2 years, we're expecting JPY 1.5 trillion for the bond redemption so that our cash position covers more than enough.

Now let me talk about Arm, one of the most important companies to us. Revenue in the first quarter, they hit the highest record, posting $1.3 billion. It increased 22% year-on-year, amazing result. Fully diluted earnings per share they hit the record high as the first quarter, it increased 29% year-on-year, again, a great performance.

Guidance. For revenue, they have a range, but the midpoint is $1.38 billion for the second quarter. Guidance for EPS, $0.47 midpoint. Compared to analyst consensus as of Arm's earnings announcement, both are higher. So I think that's really convincing and something that you agree on looking at Arm's performance.

Arm-based CPU cores for data centers are growing, as you know, it's been growing since 2019. Many hyperscalers has adopted Arm-based CPU, Cobalt and others and Google. Especially it grew significantly in the last 12 months, which should have a positive impact on Arm's future going forward. Again, many tech companies have adopted Arm's products. Arm IP-based CPU chips have been used by about 50% of leading hyperscalers, like you can see on the slide, like Graviton, Axion, Cobalt and Grace. They help us to raise Arm's brand value and recognition.

As you know, Arm keep providing chip design since the beginning, but also Arm has announced that they would develop their own product by themselves, which is a huge jump from Arm's perspective. In fact, this CPU compared to x86 per 1 gigawatt of data center capacity, it helped save up to JPY 10 billion CapEx. Very efficient CPU. And as agentic AI is deployed further, the market of CPUs for data center should grow significantly. And Arm's AGI CPU attracts more demand than what we had expected and what Arm had expected, which is all positive.

And we have 2 more companies joined as launch partners of Arm AGI CPU, Oracle and Baidu, which is great. From a PCs perspective, well, in fact, agentic AI goes into not only cloud, but also in varieties of devices. And NVIDIA announced a chip for PCs with agentic functionality. And this has Arm-based Grace CPU. Then OpenAI. And by the way, if you have any questions about Arm, you can ask Jason later.

Now OpenAI. As we mentioned in the last earning result announcement, we made a commitment in March. And that commitment was to invest JPY 30 billion this year. And of them, JPY 20 billion were executed. And remaining JPY 10 billion investment will bring total investment to JPY 64.6 billion and ownership to 13%. OpenAI's GPT is a base to offer varieties of products. For consumers, as you are familiar with, ChatGPT is there. And for business users, especially for our software engineers, Codex is offered, which we focus on and they focus on selling Codex.

On top of that, ChatGPT Work was released recently. So AI model evolves and product performance goes up accordingly. In 2022, GPT-3.5 was released as what we know as ChatGPT. And GPT have been adding more and more functionality and capabilities more than you may imagine. GPT-4 has an AI capability that solves problem, not just answers questions. It was released in March 2023, 3 years ago. And GPT-5 has agentic AI capability. So it is AI that performs tasks. So compared to 3.5 to 4, it is a significant evolution. And the latest one, GPT-5.6, agentic functionality is even enhanced. So it is an AI that gets work done. So AI does work autonomously, which is unique.

Talking about the GPT-5.6, which is the latest revision. So from expert perspective, how the 5.6 is viewed. So there was an evaluation methodology developed by UC Berkeley and others. According to their evaluation, OpenAI yielded highest performance in terms of score. And if you take a look at on the right-hand side, compared to other AI, OpenAI 5.6 cost far less. So high-performing model with lower cost is how the GPT-5.6 is seen by professionals. I'm not an engineer at all. I am a finance guy. So I cannot answer to any professional or expert questions, but this is as much as I can tell you generally.

Codex for software engineers has been growing significantly. In December 2025, monthly, excuse me, weekly active users were 0.4 million. And in July, 10 million. It grew 24x in just 7 months. On top of Codex, ChatGPT Work was released for business persons. So combined Codex and ChatGPT Work, again, weekly active users grew 25x in 7 months. So OpenAI is kind of a latecomer in terms of enterprise market because they started off with the retail users or retail market. But OpenAI has advantage in terms of their brand recognition among consumer market. So going forward, enterprise market is a key focus so that they can grow more in the enterprise market.

Now let me talk about PaaS or patching as a service. In July, last month, actually, SB OAI Japan our joint venture between SoftBank and OpenAI announced and launched PaaS as a cybersecurity solution. We are a holding company, and we are a strategic investment holding company. So from our perspective, our job is to make sure that our portfolio company work together so that we can help them to raise their corporate value. We want to help them to do that. So again, this PaaS launch between our -- by our portfolio companies are a great example.

This PaaS will be rolled out to 3,000 companies in Japan. So they start from vulnerability assessment. And once vulnerabilities are identified, they run the program to apply patches. From end-to-end, they offer this PaaS service to those critical infrastructure companies in Japan, 3,000 that is. And later, we want to even offer code modernization.

Now let me go into AI infrastructure. AI data center portfolio, we are mainly leading right now is in United States. And also, we have a project in France as well, which is announced this quarter. For United States, Ohio and Texas State, in France, Hauts-de-France, which is Northern France area is actually the place we are addressing our project. In France, we announced a plan to develop 5 gigawatt AI data center in France. So in June 2026, there was a Choose France Summit. And there, in front of President Macron, we have announced a plan for the development of 5 gigawatt. For the first phase, EUR 45 billion investment to deliver 3.1 gigawatt of AI data center capacity in Northern France by 2031.

More specifically, Dunkirk, Bouchain, Bosquel, those are the first site that we will be addressing. First gigawatt Bosquel AI data center is actually designated as fast track project by French government. So with that, we have secured 1 gigawatt of power and also accelerated grid connection is actually addressed. In United States, if we try to address the data center project or if we try to do in Japan, because data center consume a huge volume of power. So you also need to build power generation plant at the same time, which means you need to double the size of CapEx.

But in France, compared to the rest of the world, they have a very unique advantage, which is that they are actually the exporting country of the power so that they have actually -- they can provide the power to the data center use without causing any impact to the French citizens, but still, they will be able to allocate some power to the data center, meaning addressing data center project approach is different compared to the other world. And they do have a kind of a good base to proceed the data center project from that reason.

And also in France, we've been building supply chain for AI data center equipment, and establishing a local supply chain for power modules in Dunkirk to accelerate data center development together with Schneider. So SoftBank Group and Schneider are going to provide -- set the supply chain for the power module, and we'll be able to accelerate data center development.

In United States, PORTS Technology Campus, Pike County, Ohio, which we announced the 10-gigawatt large-scale data center project. This is quite a huge project. We are currently in discussion towards execution of a lease agreement, and we are coming to the final stage of such discussion. And at the same time, we are expecting construction it's starting in 2026.

One another project that we've been working on in the United States was in Milam County, Texas, AI data center and already have a lease agreement placed with OpenAI and building construction on track for the delivery. To proceed our data center project, we have a group company which plays a very important role, which is called SB Energy, actually established back in 2019. Originally, they've been addressing many power projects in the United States. And they've been expanding their business domain to data center because power and the data center developed integrated way is going to be the benefit and advantage, and we have a very experienced management to lead the business of SB Energy.

So cumulative project capital raised, which is almost a total number of project amount by SoftBank Energy is about $19 billion. And power generation and storage capacity under operations or under construction is about 5 gigawatt. So they do have quite a size of the business and the presence in the United States. In May this year, they have already filed the confidential filing for the Form S-1. And we are very much looking forward to see the further development in the future.

And NeoCloud service, which was also presented at SoftBank Corp.'s earnings results announcement a few days ago. So our domestic mobile operator and SoftBank Group holding company, together with them, we are start launching the project for the NeoCloud service in the United States. So this NeoCloud business will be providing AI compute via the cloud. So power data AI data centers, which will be provided by SB Energy, is bringing into the special purpose company and the service will be provided as a cloud service to the offtakers. We believe this is going to be also another important business model for the future.

Now into Vision Fund. Investment gain or loss, this is since the inception of the business. It's going well. and actually adding some more investment gain as of the end of June, JPY 47 billion. And this is a breakdown of the gain or loss on investments for the first quarter period. Vision Fund 1 on your left-hand side, mainly due to the increase in the fair value of ByteDance led to the Vision Fund increase of the gain on investments by JPY 2.3 billion. And on your right-hand side, Vision Fund 2 and LatAm and others, a little bit of a negative. This is because of the PayPay, which went public this spring. And the share price of PayPay a little bit weak this quarter. Actually, that was the main reason. But since then, PayPay is actually recovering its share price. And for the future, we are still very excited about this business, too.

Cumulative investment return including the exited companies, we're showing the total of the investment return. And on your left, Vision Fund, JPY 114 billion cumulative investment return. So exceeding largely compared to cumulative investment cost. So green portion is the exit case. So actually that we have already exited $73 billion. Also, the investment monetization. So those green bars that we have just mentioned in the previous page, in every quarter, especially in the recent quarters, about JPY 1.5 billion in average has been monetized and made its return.

In the past 2 years, main divestments were on your right-hand side, Coupang, Roivant, Full Truck Alliance, DoorDash and so on. So those actually went public successfully as a result that we have addressed the monetization of those. Of course, not only the invested portfolio, but the private companies, there are many ways to exit, not only sales. Sometimes that you can sell in the market for the public share, but as a private shares, you can have a transaction or M&A can be another option, too. So there are many ways to go exit of the portfolio. And we would like to execute the good options available. And we have a CFO -- new CFO for the Vision Fund, Mark Agne, especially have an expertise in the monetization so far.

Public listing and pipelines. Late-stage portfolio on your right-hand side of total fair value because of the follow-on investment in OpenAI, it went to $133 billion. So compared to March end, late-stage portfolio, total fair value itself increased by $15 billion. And let me pick up some of the portfolio companies, which are in the late stage. PRISM, formerly known as OYO, you may have seen in some Japanese market as well, and they've been -- operates a global hospitality platform. So they have rebranded under the name of PRISM and already made a filing for public offering. And they support accommodation partners with demand generation, dynamic pricing, revenue management and day-to-day operation. They have themselves have managing actually the hotels, but now that they are very much focused on the platform. So they have a clear vision and direction of the business models now.

They are also a good promising company. And Klook is the they offer an online platform for travelers to discover book tours and book tours activities, transportation, hotels and more around the world. And we have invested in 2019 first quarter and that the business model itself has been making a good momentum, and we are also excited about this company. And one note on the PayPay's movement. So we would like to use some slides to share with you. So this was announced in June 4 this year, which was to make T&D Financial Life a subsidiary of PayPay Corporation.

So having this company under the umbrella of PayPay, so that they will be able to provide comprehensive financial service on top of what they have already. So by having a life insurance business, based on the life stage of each users, we will be able to provide very good comprehensive services. We've been looking into this agenda. And this time, it made good results. Another one is the capital and business alliance with Seven & i Holdings. On your left, you see SoftBank Corp., domestic operator, PayPay and Sumitomo Mitsui Card, JPY 100 billion each. So the ownership stake itself is about 2.13% in Seven & i Holdings, and we are starting the capital alliance, especially when it comes to PayPay, it is a good achievement here because there are about 20 million customers per day visiting 7-Eleven stores.

And 30 million transaction per day made by PayPay. So utilizing the data in between so that we will be able to utilize and see the future combinations. Number of the store wise, they have 22,000 stores in 7-Eleven and PayPay has a 75 million user base already. That can be a great value generator for each other.

Now let me talk about physical AI. Commercial deployment is accelerating across the portfolio companies. including companies under the umbrella of Robo Holdings, over 30 physical AI businesses are on our portfolio. I think we can group them into 3 categories. First, robotics, obviously, which should include humanoid in the future. So robotics and AI technology is one bucket. And then automation of infrastructure like factory automation that has been there for some time. And the third bucket is mobility and autonomous driving. In each sector, those portfolio companies that are only a handful of our portfolio companies should play key roles in respective area. And Masa is expecting a lot and spending a lot of time in physical AI sector.

And please note right top, ABB Robotics, we announced acquisition. And by end of this year, we should be able to finish approval process and complete acquisition. Upon completion of acquisition, huge operations, huge services across the globe will be integrated and it should deliver great synergies and accelerate deployment of robots.

Last but not the least, financial policy. As I keep saying, no change on financial policy, which is to maintain LTV below 25% in normal circumstances, and to maintain at least 2 years' worth of bond redemptions in cash position, making sure that we are safe and sound. To summarize investment amount compared to last year, 2025 in 2026, currently, we are kind of controlling the investment amount so far, but the biggest investment we have done so far is in OpenAI, block indicates already executed. So total JPY 20 billion already executed in terms of investment in OpenAI. And JPY 10 billion more for OpenAI, 5.4 for ABB Robotics and 3.1 for DigitalBridge, which we announced M&A last year. So we committed $18.5 billion. We closed our financings one by one. First, we entered into a financing agreement using OpenAI shares. I think it's today, as we speak, $10 billion of OpenAI. And we arranged LBO financing max $1.8 billion ahead of acquisition of ABB's Robotics business.

So thanks to great support from globally known big banks, we were able to arrange those financing. We really appreciate their continued support. About ratings, S&P kept negative for a long time, but they revised its outlook from negative to stable, which is one of the great piece of news in rating. As I have explained, we keep focusing on those 4 sectors to raise our corporate value moving forward. Again, information revolution, happiness for everyone.

Thank you very much for your kind attention. And now we'd like to start taking your questions. Thank you.

Now we would like to take questions. First, we would like to take questions from the floor. [Operator Instructions] Now we would like to take questions from the floor. The very front row person in the middle section, please.

My name is Suga from Nikkei Newspaper. I have 2 questions, please. First, after July, U.S. tech major player has been criticized with the excess investment in data center. Would you keep your amount investing in data center? Or are you going to change your direction in investment in data center business?

Yes. Thank you for your question. Data center projects are announced one after another in the United States. I think that the Stargate project of ours set the fire in the market. And also, there are some arguments that they may be bubble, but we do have a very clear view on demand and supply imbalance. So the supply is very much short. And that is why this project is announced one after another, which I believe is a healthy situation.

Without securing computing power, technology and the service of AI can be slowed down. So I believe this is necessary, the project owners leading such an investment, and we would like to be the part of it, too.

My second question is about OpenAI IPO timing. I would like to ask you. There are some media coverage talking about postponing to next year, but how do you see about the timing? And if there is any postpone, what, kind of, impact to your business?

Well, same answer back to you, actually. We are not a company going public. So I have no comment -- or I cannot comment on the timing. Of course, as one of the investor in the company, we are very much keen to see and expecting on IPO of OpenAI.

Next question.

My name is Murai from Diamond Magazine. I have 2 questions. First, SB Neo. SoftBank Corp. is engaged. So what's the purpose or why? I wonder if you want to build expertise in NeoCloud. So again, size of NeoCloud, you mentioned 10 gigawatt. And Ohio and Texas are talking about 10 gigawatt. Is it just by accident or what?

So SB Neo, kind of, equal to Ohio or Texas and also nonrecourse financing you mentioned. And I understand that you don't want to affect your own business. So how much would it be in terms of nonrecourse financing? A lot of questions around NeoCloud. Sorry about that.

NeoCloud business itself is led by SoftBank Corp. So details, I would encourage you to ask them directly. And you're talking about 10 gig. I think you are referring to Portsmouth project. And project itself does not necessarily mean SB Neo project. As you can see on the slide, SB Neo's role is to provide computing to customers. And talking about the nonrecourse loan, again, this is a project, and this is just one of the typical examples of project financing.

And offtaker project finance depends on the credit of offtakers. So from SB's perspective, it's a nonrecourse financing. So again, I suggest you to talk to SB or SoftBank for more details. So 10 gigawatt is the start of NeoCloud service they mentioned in announcement. Of course, they want to do 10 gigawatts, but they want to build demands as they expand the business. So NeoCloud data center is based on an assumption of Texas Portsmouth data centers. It's still being finalized, which project or how much they get involved in. Next question is about Arm. Just to clarify. So as a project, it NeoBank or NeoCloud is not a Portsmouth project. So it doesn't necessarily mean this business will utilize Portsmouth data center.

Again, question is about Arm next, especially Arm chip. I'm sure that you started or launched a business of Arm-made or Arm developed business. And from '26, '27, '28, $2 billion of demand is expected. We heard in your announcement or earning results. So how probable it is? Or are there any upside? I mean, how are you -- how confident are you about the future demand in the next 3 years?

Jason, you may want to answer?

Yes. So with Arm AGI CPU, we did say initially that for -- it starts shipping in Q4 of this fiscal year. And then the combination of this fiscal year and next year, we said initially when we announced the product back in March that there'd be roughly $1 billion in revenue.

Last quarter -- actually 2 quarters ago, we said that demand after we announced it had already expanded to nearly $2 billion. And then this most recent quarter, so last week, we said that it's now -- we have orders for more than $2 billion. But we have not updated our guidance on what the revenue will be because we are still working through the supply chain constraints, primarily around wafer capacity. There are some other constraints as well.

We have made progress on satisfying some of those constraints, but we're not ready to say exactly how much of that incremental $1 billion plus that we will be able to ship until probably another quarter or 2 as we get more clarity on both the pricing as well as the entire wafer capacity that we're going to be able to get access to.

In terms of the overall demand beyond this year and next, which for us is fiscal '27, '28 and then even going into our fiscal '29, which will be SoftBank's fiscal '28, we expect demand to double the year after that. And that's because the order pipeline for AGI CPU continues to grow -- and as you heard Goto earlier, we added a couple of very large customers this quarter. We expect to be continuing to announce new customers. So demand looks to be very much on track, possibly even ahead of what we said back at our Arm Everywhere analyst event.

Now on to second from the front row.

Miyajima from Magazine FACTA. This year AGM, shareholders' meeting, Masa was talking his 15 years here that he would like to devote and net asset value to JPY 1,000 trillion or that was very much impressive challenge that he announced at the shareholders' meeting this year. And that's actually the term that Masa has. And Goto-san, Kimiwada-san, you've been run along with them. And once that Masa becomes -- get older, you also get older along with him. But how did you think about what Masa said that he's going to be working even more 15 years?

Thank you very much for your question. There, at the shareholders' meeting, he mentioned about 15 years, and we thought about how he expressed because we didn't expect that he's going to say in that way. But we were also happy to hear how he motivated as market cap is growing and Masa, the founder himself and also CEO and also leading shareholder that, that position has maintained and how we're going to come up with a succession plan for such a company is always a question from the investors, too. And we, but at the same time, like to see Masa show the great performance even more and he can work so that we would like to support what he needs to do. And recently, there were patent on the generative AI becomes #1.

So 3,000 cases that Masa has applied he mentioned about 2, 3 years ago, and I believe that that's the kind of results out of those activities. Having everything applied, that can create the value in the future.

So that's one of the kind of a basic plan that you are doing, but Masa and Goto-san, having applying such a patent everywhere, what, kind of, financial benefit do you think there are? Having 3,000 patents, how many months or how much money has been spent, which I don't know, but that can be one base for the golden goose, I'm not sure, but how do you think about that?

Yes, patent application requires manpower. And lots of the time and money as well. In the past few years, AI-related patent has been applied aggressively by Masa and his team, which we try to effectively address utilizing AI so that we became more quickly in terms of applying for the patent. Still some costs required, but that can be the seeding phase for our future growth of the business. So we actually encourage -- although there might be some cost come along. But still, I think that it's good to challenge.

And the patent is quite interesting because whether to be applied better or not applying is better, there are some cases like that as well. So that is something that we do have some experts, professionals on patent from the legal team internally so that we'll be able to address so that we don't really miss the opportunity.

So what I'm hearing because you have AIs with OpenAI, that allows you to provide the resources to do compared to other big tech players, they are in short in terms of resources and so on. But actually, you applied more patent application than the others. Is that because of OpenAI investment?

Well, the structure or the approach itself has been sometimes studied and analyzed by Masa, which is a great effort he's making, I think.

Next.

MJ from Bloomberg. I have 2 questions. First, about Ohio data center project. According to media coverage, NVIDIA guaranteed some $250 billion worth, so that OpenAI can lease up to 10 gigawatt or something. Can you make a comment on that?

Right. A lot of stakeholders are talking each other or negotiating each other. So I cannot make a comment in detail. So once we are ready, we'll definitely share with you.

Second, about OpenAI. Competition of LLM is fierce and Chinese models are getting better and better. And SoftBank invest in only OpenAI in terms of models. But what's your view on OpenAI's valuation at this moment? And do you have any plan to invest in other LLM or other models than OpenAI?

OpenAI's valuation I think I should ask Kimiwada-san or Mark, can you answer to the question?

Thanks for the question. So you had 2 parts. The first part was on valuation, I believe. We apply the same valuation model for OpenAI as we do for any of our holdings, and it's 3 parts, basically. And one of those parts is -- and one of the main parts in a private holding is what the last round was at. The last round is where we're marked right now in OpenAI. And in the absence of any significant news, we will stay marked there, but we're always following along with what's happening. And then you wanted -- you had questions about the model.

What about the possibility of investing in other LLM models given.

So we -- in the past, we have invested in various competitors, for example, in rideshare. But in this case, we are investing with OpenAI because we consider them more of a partner than an investment. And so we are planning to do more things with OpenAI -- and we are very, very confident in their model. So it's our choice to put all of our eggs in OpenAI as far as LLM is concerned rather than spread it around.

Thank you very much. Any other questions, please? Then front row in the middle, please.

Kimura from NHK Broadcasting. A bit similar to the first question. You may mention that you may consider slowing down the investment. And if you look at the U.S. market, tech share price is pretty much slowed down. But what is your outlook for the financing in the future?

Well, I said -- I may misled you earlier. If you're only looking at this number, it looks like a little bit less investment. So that's why I said that it slowed down -- looks like slowed down. But actually, we are not a company having an investment budget in advance. Intentionally, we don't really have investment budget. It's because we invest from our balance sheet as well, and we try to maximize the performance of it. Sometimes if you have a budget, but if you come to the end of the fiscal year, if you still have a budget available, you need to consume everything then that you just pick the wrong investments, which happens in many companies or if you finish all the budget amount, but if you find very promising companies, but because budget constraints, you cannot really invest.

That, kind of, disturbing situation happens always for the many companies as well. So if you have a good opportunity, you are an investor. So that's why that we don't have a budget for the investment. But when it comes to financing, coming back to your question, financing market is not bad. Actually, of course, the interest rate itself has been increasing in the past few years. So there are cost and interest, but we have certain level of the investment yield and challenge to that, then that the interest cost can be acceptable to some extent.

But when it comes to the size of the financing globally, I think it's quite activated. So issuer and investor demand and supply balance-wise, I believe both are quite positive about the financing.

And how are you going to use that money for -- is it investment in for the real business?

In the past 40 years, we have experienced from the bubble situation. I think that the current situation is not really the same. It's not really the bubble because we do have a use of proceeds and the financing itself for the needs, I believe, is going to be no problem for us. And our financing capability-wise, if you look at our balance of the assets, as we have about JPY 60 trillion close level of the net asset value after the debt. So corporate credit capacity-wise, we believe that we are in a very strong position from that sense.

Are there any more questions on the floor?

Yagi from Nihon Keizai Shimbun paper Nikkei. If you could let me know LTV as of August 5.

As of August 5 of LTV, we don't present in that way. We have a rough idea, but I want to refrain from commenting on that. But you can imagine, as of end of June, net asset value was JPY 72 trillion or something. And based on that, we calculate loan-to-value. And if NAV is -- sorry, JPY 58 trillion, then NAV might be widened since end of June, but still in the range of our safety operation.

Second question is about your data center plan and its progress. The Ohio's 10 gigawatt that you announced in March, that is, since the capacity is that big, I assume you build multiple buildings. So the first building, do you know when the first building will be completed? And again, I assume that there are multiple offtakers that are interested in the project. So again, are you confident in terms of the demand?

Start of the building also building will get started in 2026, as I mentioned earlier. But the overall schedule is something that we cannot share with you now. But again, once we are ready to share, we'll definitely do that. About off-takers, again, we are closely communicating with offtakers. And we are confident in the demands, and we want to make sure that we will find the best solution.

Thank you very much. Any other questions, please? Then we would like to go on to the questions from online participants. [Operator Instructions] I would like to ask Mr. Araka from ITmedia. Araka from ITmedia, can you hear me okay? Yes, please go ahead.

So gain on investment is mainly coming from Intel, you said, which was quite interesting to hear. So any comments about the business development in Intel? Or what is your expectation for the future of Intel? If -- so that was a positive contribution to your investment gain, right? Intel?

In the past 3 months, semiconductor names was actually surging. And Intel was one of them that the April to June quarter, their share price was dramatically increasing, which was quite a positive surprise for us. And that actually contribute to our investment gain.

Any expectation for the future of Intel?

Of course, although it's quite volatile, but as an investment company, we hope that the value can increase more.

That's all for Q&A session. Thank you very much. This concludes the SoftBank Group Corp. earnings results announcement for the first quarter of the fiscal year ending March 2027. The video footage of this meeting will be distributed on demand from our corporate website. Thank you once again for joining the SoftBank Group Corp. earnings results announcement for the fiscal year 2026 first quarter. Thank you.

As it is now the scheduled time, may I ask Mr. Son to take the chair.

Maybe we should wait a little bit more time for other shareholders to take seats. Good morning. I am Masayoshi Son, Chairman and CEO of SoftBank Group Corp. Thank you all very much for taking time from your busy schedules to attend today. As stipulated in Article 13 of Articles of Incorporation, I will act as Chairperson of this meeting. I hereby declare the opening of the 46th Annual General Meeting of Shareholders. This meeting utilizes an online attendance system, allowing shareholders to vote and ask questions via the Internet.

Our directors, Mr. Masami Iijima and Ms. Keiko Erikawa are absent today due to unavoidable circumstances. The agenda today is just as stated in the notice of convocation posted on our website. I also report that meeting has a quorum. Now our moderator will explain the proceedings of this meeting.

Thank you. We will start with the reporting matters, then explain the resolutions and our business strategy. Following that, we will proceed to a Q&A session on all reporting and resolution matters. After the Q&A session, we will take vote on the resolutions only.

For exercising voting rights, shareholders present will be asked to vote by applause and shareholders participating online are asked to cast their votes via the Internet. Please note that online voting is already available, but once the vote is cast, it cannot be modified. After explaining the resolutions and business strategy, we will first answer the questions submitted in advance. Then we will invite questions and comments from shareholders attending in person.

[Operator Instructions] Also, questions and comments from attending shareholders will only be accepted until the start of the Q&A from online shareholders. Then we will respond to the questions submitted online today. Online question submission is already open and will close 5 minutes after the start of the Q&A session.

[Operator Instructions] Each person may ask 1 question. Regarding motions, online shareholders cannot submit motions, including those related to the meeting procedures or resolutions as previously notified.

In case any attending shareholders submit a motion to amend the resolutions, these will be deliberated together with the original proposals, and we will vote on the original proposals first. Thank you very much for your understanding.

Now I'd like to call for a vote to confirm if this procedure is acceptable. Online participants, please press applaud button if you agree. Those attending in person, please applaud.

[Voting]

Thank you very much. Let us now confirm the voting results. Please wait for a moment.

Thank you. As the majority is in favor, we will proceed with this procedure. The reporting matters are shown on the screen. And we have prepared a video to explain our business overview. Please watch.

[Presentation]

That concludes the video presentation. Let me now present today's resolutions. Details are available in the notice of convocation on our website.

Proposal 1, appropriation of surplus. Proposal 2, partial amendment of the Articles of Incorporation. Proposal 3, election of 9 members of the Board.

We will now proceed with the business strategy presentation.

Recently, AI is a buzzword. I myself using ChatGPT pretty much every day, maybe 10 times or 20 times a day for asking questions, checking my ideas. So I were doing so many things in the weekends, maybe 2 hours or so, I will be communicating with AI, even weak days, maybe 1 hour, 2 hours. Of course, I communicate and discuss with employees every day, but the agenda of the discussion may varies, asking ChatGPT, how do you think about this? How do you think about that?

And AI or ChatGPT reply back to me in details, in deep or in a wide range of areas. So it's a great revolution, great evolvements. But so far, AI has been your consultants to answer your questions or generate the videos, clips or voices. But now we are using it as an Agentic AI or AI agent. So AI itself autonomously think and make an action by itself. So we don't have to ask them to do this, do that in details. But as long as we share the same goal, AI will be working without resting and working for around the clock. So that's how it comes. And this is seeding a human brain. So now superintelligence, not only the brain is great. If it's only the brain, then AI may just make its own move.

But now if you look at human, we know the ethics. In AI, we will also cherish the things that we cherish. So those are the things being built into these latest models, then we can stop AIs its own move or any seedings that AI may do. So those can be vanished. So we are not -- we are seeing the superintelligence now.

So with the superintelligence -- and we would like to be that such a platform provider. In Japanese, it's like a social infrastructure or more simply put, it's like organizer and SoftBank Group is aiming to become the provider of the infrastructure of artificial superintelligence. In my presentation, I have 3 agenda. So ASI, Artificial Superintelligence, and ASI is abbreviations for that. So I would like to discuss and present my vision on ASI. And to remain -- 2 agenda is the JPY 1 quadrillion. What is JPY 1 quadrillion? That's second agenda of mine. And third, as we say, we would like to be the platform provider of ASI, but maybe I should be more -- put more colors on what -- how we picture ourselves, how are we going to increase, that will be another -- a third agenda of mines for today's presentation.

About 16 years ago, SoftBank Group actually celebrated 30 years since its foundation. So maybe the second week of the anniversary, I said we discussed the next 30-year vision to my employees. Maybe I discussed about 30 minutes or so. And then for this next 30-year plan, I want to be making our sales for JPY 1 trillion, JPY 2 trillion. That's the size that we are aiming for. That's how I explain to my employees back then, and the next day that they left the company.

That's very beginning. That's the very first 30-year vision. So after 30 years, about 16 years ago, we celebrated next 30-year vision. So when I was 19-year-old, I had a 50-year life plan as a business person before we discuss about 30-year plan. In my 20s, get acknowledged as a business person, in my 30s, finance warchest, so on and so on. So there are 3 chapters that I set out for my 50-year life plan. Now I am 68 years old. 60s, handover business to next generation. That was my original plan as a 50-year life plan. But now becoming 68, I became more greedy. I'm still doing healthy. 50 year ago, the life -- average life was not this long. And when you imagined when you are 60, I will be a very old man, but for my health, I will play the ping-pong with young employees.

So I have 3 young employees, and they will change the part, and I will be playing by myself for 90 minutes or so and play the ping-pong. My actually ping-pong is all the smashes and hits. So I'm the risk taker, so I just make a smash here and there. So it's like a kind of a -- quite a wild ping-pong game and jumping around and play very hard until I'm so sweated and tired, I will play. But still, those 3 young employees, they have to switch the part. Otherwise, they will not be able to sustain, and we all exhausted. And I myself on the latter half, I will be even more energetic. So I think I still can do. If I retire now, maybe I cannot really kill the time anymore. So -- and at the same time, for the growth of SoftBank Group, I wanted to be more contributor for the growth of the company. So as a result, I decided to revise my 50-year life plan.

70s, maybe 10 years, 15 years, I would like to do more. Of course, I will stay healthy as long as I can be staying as healthy and also sound. But so far, I'm doing so great. Every month, I will making a business trip overseas twice or 3 times a month, but still doing great in very good health. And 70s. So I'm 68 now and in my 70s, What I will be doing? So in my 70s, as I mentioned in the very beginning of my presentation, I would like to pursue Artificial Superintelligence. As far as we do, I want to be #1, of course. It has to be #1. That's my character. Of course, in Japan, I have to be overwhelming #1. And even in the world, we want to be -- in our business area for ASI, we want to also target for becoming #1 in the world as well. So we will do it. I will do it. So there has been evolution of humanity history for about 200,000 years or so. Most high-level or highest capacity, highest levels has been reached by human, but the constraint of such humans evolution has been intelligence. So nothing can be more intelligence than ours or our -- than humans.

And at the same time, execution has to be come along with intelligence. Compared to other animals, human is the superior and that itself is actually the evolution of humanities. So that's the power of ASI. And once that we have combining human and ASI, so we can take the best of the intelligence. And then that we can extend the intelligence even further and at the same time, we can have physical AI for the execution of such intelligence, then so many things will be even more evolved.

As I mentioned, about 16 years ago, we announced next 30-year vision. So when I start saying about AI in superintelligence robots, that's not something that new. When I was 19 years old, we -- I created my 50-year life plan. But even 2 years before, when I was 17, I first time saw Microchip computer, photo of the Microchip and I imagined, but then that actually, the way we should be heading -- I should be heading in human may be creating something exceeding human in the future. So that was when I was 17, and now is the time that we are about to see by our own eyes.

So this next 30-year vision, and we're about half of it. So this is 16 years now. So it's about half that has come. And I would like to look back once again. That time, I said we will be seeing the brain computer to life. So it's not a simple programming but the computer itself is going to be think autonomously, that kind of computer is going to be born and that's how it comes as a result. And at the same time, this brain computer will be connected to robots. And once robot coexistent with intelligent, then the things that are too dangerous or too boring, those kind of actions can be replaced with robot.

So next, 30-year vision actually announced 2 things. One, ASI; another thing was Robots with superintelligence, called physical ASI. It about -- I mean, AI robots here. And first, this ASI -- once that you see more evolution of ASI, it's not only the simple answer, but actually AI itself is going to work on behalf of you. And one person will be able to play 1,000 people's labor.

So many things. So many calculations, so many tasks that the one person has been doing can be a super white color. At the same time, 1,000 agents per person will be actually worked out quicker, faster and also a better way. I myself, actually applying for the patent for 1,000 cases by myself, that's how I like to think hard. So invention, think about it. I draw a design and so on. How do you think about this idea? What do you think about these things? That's something that I asked ChatGPT and we create the design details made together with ChatGPT so many times so that we'll be able to brush it up.

So not only the simple calculation, but think, create and generate. We can draw the design or we can build the business plan together. Or at the home, your kids, your grandkids. One teacher can be what 10 expert teachers can be -- provide education to each individual kids. That's also available in the future or researchers engineer for the one person or for each individual researchers or an engineer, so many artificial -- excuse me, the superintelligence will be provided so that they will be able to amplify their power. And each individual has a different status of the health every day.

Even you want to check with your doctors, doctor may be very busy or doctors may not be the expertise of what you are feeling. Then you'll be able to receive the best care of the superintelligence doctor or you'll be able to consult with such doctors. Of course, they will evolve to ASI, but human will not be replaced or should not be replaced. Human needs to connect to human and try to aim for the happiness for each other. So you cannot replace but you can evolve the human. So this is a tool to evolve. This is a colleague, partner of yours to evolve the humanity. That's how I see the ASI.

The second agenda for the next 30-year vision was physical AI or so-called physical AI in the industry. So robot with intelligence, intelligence with the body. 24x7, autonomous operation in the factories, robot and robot will be communicating each other, cooperate each other and make the operation or going to the high-risk work space, disaster areas deep in the ocean. Or we can also use the -- for the rescue from the disaster when we have any threats for the people's life. For those areas that the people cannot really go and rescue.

At the factories, there are the labor shortage. Especially in Japan, we are in the aged society. So the labor power is in short here and there. Even the farm, there are difficulty finding the successor, but that's also the area that the super AI robot can help and support.

So as you can see, AI is evolving at a tremendous speed, which is exciting. And again, it's evolving in the process of ASI advancement, what can SoftBank Group do? We have 4 main themes or focus. First, AI model, which is related to AI intelligence. We want to make sure that we have the best AI model. We alone cannot do. So we work with OpenAI, which is a creator of ChatGPT and OpenAI was the trailblazer of AI. So we work together with OpenAI to keep evolving AI models.

And Arm, one of our core businesses or core companies, and CEO of Arm, Rene is with us today as a Board member. And again, will play a key role and the chip is not only a potential bottleneck, but also power can be a limitation. So AI infrastructure or data center which can be a limitation of AI evolution, we want to be the best AI infrastructure provider in the world.

In terms of robot, we want to be the best robotics company in the world. So let me dive into those 4 focuses more in detail. First, AI model. In the last 3 years, OpenAI has been helping evolving products. ChatGPT is for chat and not only chatting, but also image creation, video creation and program coding and cybersecurity.

So OpenAI has been enhancing their capabilities. So going forward, towards the ASI OpenAI will keep evolving and expanding into new domains. When I talk to the Head of the development team, Mark Chen. So whenever I talk to him and his team, it's really exciting because they are getting ready for new programmings and new products.

Next, semiconductor. NVIDIA has the top market cap. Just 3 years ago, not so many people heard about the name, but NVIDIA surpassing Google, Microsoft and Apple, and NVIDIA's market cap is the biggest at the moment. NVIDIA is good at generative AI, creating talks, creating images, creating videos. In the area of generative AI, NVIDIA is best at GPU, which originally was designed and create images. So speech, image are the leading areas where NVIDIA can support, and NVIDIA strives. And NVIDIA has the biggest market cap at the moment.

But in the next phase, not just creating communications or images, but each and every agent thinks and acts and thinks and act like 1,000 people working at the same time. That's what it calls Agentic AI. Or I would say, agent AI. So in the world of Agentic AI, not a GPU for graphics but CPU will play a more central role in terms of CPU design, Arm is absolute #1 in the world. Arm is listed now, but still, SoftBank Group owns around 90% of its stake. So Arm is effectively our company, your company, you are a shareholder of Arm effectively. So Arm will play the key role in the world in terms of chip.

In the past, Arm just made a design and provided the design to the manufacture of chips. So not only creating design but also creating chip itself is something that Arm is now going to do, which was announced a month ago. In the last several months arms of stockpile has been skyrocketing because of those 2 reasons. CPU will play a key role in AI and Arm evolves itself from just the designer to chip maker. That's the reason why I think Arm's market cap is around JPY 70 trillion or something which is well beyond Toyota and KIOXIA in Japan.

But I think Arm can grow even more 10x and even more, Arm is just at the starting point. So Arm's chip, NVIDIA chip, those chips are needed to build a server as a system. And the data center is a place where those servers will work. That infrastructure has a complex pipelines and wiring and that would emit heat. So you need to cool the heat in the data centers.

So it's like an AI factory, and AI infrastructure or AI data center is an AI factory. And data centers would need a lot of power. So SoftBank Energy, we have 90% of the stake in SoftBank Energy. That's one of our subsidiaries. So 10 gigawatt of capacity or -- it's equivalent to the capacity of 10 nuclear power plants.

In Japan, I think we have about 30 nuclear power plants in Japan. So 1/3 of the total Japan's nuclear power plant capacity, that will be the capacity of data center in America. So the Secretaries of Department of Commerce and Department of Economy are standing with me. So that's a joint project with the American government. So this is the absolute #1 data center in America. And that's what SoftBank Group is helped creating.

And in Europe, which also suffers from power shortage in France and French President Macron, I met with him earlier this month, and we agreed to create a data center in France as a capacity of 5 gigawatts, which will be the biggest data center in Europe. That's about AI infrastructure.

And the fourth focus, robotics or robots. How prepared are we in terms of robots? We have many robotics companies each and each business is #1 in respective sector. We have been acquiring those businesses. I think Pepper is most popular in terms of robot by SoftBank. We started creating Pepper like 10 years ago. And Pepper is still working in our retail stores, if you notice, and we have acquired robotics companies like them and those businesses include ABB Robotics, which is #2 in the world. In the traditional robots, ABB Robotics is #2 in the world. But now we have ABB Robotics underneath of SoftBank Group.

We want to create leading robotic company. So we have those robotics company underneath our SoftBank Group umbrella to get ready for AI robotics in the future. So again, we will be the absolute #1 AI robotics company in the future. So in the times of ASI in those 4 sectors, we will become the absolute #1 provider in the world. In playing Authors, for example, a board game, it doesn't make sense if you have center areas. The key is to get 4 corners. In the times of ASI, we want to create a new social infrastructure and we want to take control of the 4 corners of ASI board game, if you will. And we want to be #1 in the world.

SoftBank Group started from domestic business. But recently, we entered overseas markets like America, China and Europe, and we started investing in those markets through our Vision Fund. And now generative AI gets evolved into ASI. The United States is most powerful at the moment, and we want to compete in the United States and we want to have #1 position in each and every 4 corner. And as we try, as we strive, we want to be the #1.

Last year we had AGM around the same time. Since then, our share price has been up a little bit. But most of them, most of the shareholders may still be skeptical about our story. You might believe in me today, but you might get again, skeptical, and you may feel sad or happy by looking at SoftBank Group share prices because our share prices are so volatile. And maybe our financial statements are too complex, even to the professionals. So I understand if you're half convinced, but half skeptical. But I'm not, though. I am 100% convinced. I have no skeptics at all.

But looking at the share price, I understand some of them -- some of you are half convinced, half skeptical. But think about it. SoftBank is a strategic holding company. We have operating companies in our group, but SoftBank Group is rather not doing our businesses, but we are a strategic holding company, having equity of different group companies, whether it be Arm or OpenAI, SoftBank Group is a strategic holding company, an investment company. And our market cap compared to Arm's value, we have about 90% of its stake in Arm. And our Arm's equity value of holding is bigger than SoftBank Group's market cap.

We have about JPY 55 trillion of Arm, and we are looking at the value every day. According to the latest value, our Arm equity value of our holding is JPY 61 trillion. And PayPay, Yahoo!, LINE and SoftBank, we have those subsidiaries in group, [indiscernible] and others. So putting them all together, we have borrowings as well. But against the value that we have maybe the debt percentage is about 10%, 11%. So just 10% borrowing debt, what should we be afraid of? Because we have a strict financial policy with a set LTV target. So anyway, just a net asset value is JPY 74 trillion as of today. So 2x bigger than our market cap is our net asset value.

Most of the companies that we own are listed which means those values are being evaluated every day. And as of today, our net asset value is twice as our market cap. I think you share is maybe JPY 6,500 per share and Arm's equity value -- Arm's -- sorry, OpenAI's equity value of our holding and SoftBank's equity value of our holding value, less net debt, we have net asset value of JPY 74 trillion. In your wallet, you have cash and you have shares. And all the half of that value in your wallet is not seen, clearly, you keep holding your value because you believe in the value.

You may be much richer than you think actually because unrealized value has twice as the share price that your share has, but until you monetize, it's not cash. But looking back our history, 16 years ago, when we announced a new 30-years plan at the General Shareholders Meeting. So 16 years ago, equity value of holdings was JPY 5 trillion. A few years ago, I talked about goose story. So let's say 1 golden egg equals to JPY 1 trillion. So 16 years ago, we had 5 golden eggs or JPY 5 trillion. But we had JPY 2 trillion of borrowing. So shareholders' value 16 years ago was JPY 3 trillion or 3 golden eggs. So equity value of holding minus net debt equals shareholder value, which was JPY 3 trillion, 16 years ago. So the net asset value or NAV was JPY 3 trillion, 16 years ago. NAV is one of the key metrics in the financial sector. And NAV 16 years ago, was JPY 3 trillion.

SoftBank's market cap was around JPY 3 trillion 16 years ago. So again, market capped JPY 3 trillion, and the net asset value, JPY 3 trillion. That was 16 years ago. That means even though a goose lays eggs, but the value of that goose was valued at 0. Only the 3 eggs were seen as SoftBank's value. So the goose has no value at all. That was what the market saw or thought. But think about that. In the last 16 years, we increased the number of golden eggs from 3 to 74 million. Again, 3 eggs increased to 74 eggs, but eggs don't lay eggs, goose lays eggs. Eggs don't lay eggs, which means who helped increasing the golden eggs? That's what I want to ask. Again, eggs do not lay eggs. The goose increased the number of eggs laid. So future value of the golden eggs was calculated already at present. And 16 years ago, the value of the goose was seen at 0, but we or I found the value of the goose. And it was proved. History proves, was that by chance? Well, looking back, the time since listing, when we were listed. It was JPY 2,000 billion (sic) [ JPY 200 billion ].

And I started this company with only JPY 10 million. And we got listed at the value of JPY 200 billion, sorry. And in 16 years, our value increased 15x. Then the number of golden eggs was 3. But then, we grew by 25x. And now we have 74 golden eggs.

In the last 15 years, it was not just by chance. So how many years do I have to fight to prove myself and convince you that the goose did a good job. So the things that you have to count is not only the number of golden eggs that you see now, but you have to look at potential eggs that the goose can lay in the future, like AI data centers and everything. Those are the things that are not seen yet, but they are inside the goose, if you will. So again, you should not look at only the golden eggs in front of you, you have to find value of the goose because the goose has a capacity to keep laying eggs for 40 years. See, the only people who believe in that story I think you are those people. You have an opportunity. You still have an opportunity. So you should be happy. You should be relaxed. So you feel good, you feel better now. But what about the future?

Now, you may start being skeptical, but stop. So it's been just half of the new 30-year vision. So the next 16 years, so 15x increase, 25x increase in the last 2 phases of 16 years. So for the next 16 years, should be JPY 1 quadrillion. Well, I have to be brave enough to say this, believe me, because if I turn out to be wrong, people start criticizing me and giving me hard time. But I have a small pride myself. Since the foundation of SoftBank, every time -- we have not disclosed profit outlook in the following year, most major companies do, but we don't. We are holding company and our share prices go up and down. So if we say that our outlook for next year's income would be like that. And if we don't hit, we will get criticized. And sometimes we announce, but whenever we announced next year's outlook, we never miss.

We have made mid long-term business plans, and we never missed the outlook target. That's what I am proud of, my commitment, and whenever I say something, I will just do it. Or in other words, I wouldn't say anything that we -- I don't have confidence to achieve. I think talk big is not a bad thing. If you talk big, sometimes may call you a liar, but I would say big vision. When you have a vision, when you have a big mid long-term vision, you have a dream, you have a vision and you have a specific image with a specific target date and numbers. And imagine like that you have already achieved. When we talk about the big vision, I think in the last 30 years, what's missing is big business leaders, big vision. Konosuke Matsushita or Soichiro Honda or Akio Morita, they all talked about a big vision, remember? But now, the major business leaders don't talk about the big visions because they need to be brave enough to do that. And the founder had confidence to grow their own business. So I think that's why they would talk about big vision.

So again, JPY 1 quadrillion, 1,000 golden eggs, from 3 to 74 golden eggs, again, eggs do not lay eggs, the goose lay the eggs. So SoftBank Group is the factory of golden eggs. Our role is to help the businesses that we have to become a golden egg. We invest and we nurture the businesses that have potential to be a golden egg in the future. And then they will get listed. They will go public. Why? Because those leaders should be seen by analysts, sometimes get criticized, so that they keep motivated by themselves to grow further.

So we, as a holding company, help them to nurture making sure that they can create and enjoy synergies. Again, those golden eggs were laid not by accident, not by chance, our absolute focus is to on specific areas that we identified AI, for example, and we have built our expertise in the last 40 years. So the value of golden eggs and the value of the goose should be seen not only just by egg value. So still, plus alpha of goose value should be there. So 74-plus alpha, which is goose value is truly valuated. So maybe the value of the goose should be or could be JPY 900 trillion. So call me Son Kaicho or Son -- not Son Kaicho, which is Chairman's Son. So joking aside, I know what role I should play. And the goose value should be upside and again, shareholders or some people even though AI chips, memories and those companies' share prices are going up, but some people may wonder, AI might be a bubble like in the era of a net bubble so-called and some people called it Internet bubble.

So AI revolution has just started. A bubble? I think it's an insult of AI, if you say it's an AI bubble because it's just the beginning. AI's potential will be unlocked. So in the third year of the Internet revolution, if you could go back to the era like Google, Microsoft, Yahoo! and Amazon back then, maybe those were still start-ups since their foundation like a 3-year old. And if you could go back the years, you would have bought more shares of those companies. And AI, again, is still beginning, and it's been maybe 3 years since AI was born.

And I talked a big. And AI and ASI will be coming, and we will control the 4 corners of ASI, the Board. That big vision at least in Japan, nobody else, but me talks about that big vision. In the United States, maybe somebody takes control of 1 of -- only 1 of 4 corners, but I don't think many leaders try to take control of 4 corners, every 4 corner. But out I will do. I will do it.

So I cannot be retired. 2.5 years ago, I lost my daddy. He fought against cancer, and he suffered from pain, cancer spreaded from the lung to every corner of his body. And I visited him every day, and I spoke with him every day. And on the day he passed away, it may be before 2 hours before he passed away. He was too weak to talk. He was too weak to talk. And I was with my brother Taizo visiting him at the hospital.

And we cried and screamed like, "Don't give up, don't give up" to my father and he was too weak to say, but he just opened his eyes and said a word, "A tiger leaves behind his skin when he dies, and human leaves behind his reputation when he's dead." So that's the last thing that he said to me. And I screamed like crazy, like "father and dad." And I often looked back the time when he said that right before he died, what did he want to say.

When I was kid, he went back from work and he enjoyed Sake at home, and he often said to me when I was little, he said "you need to make money to live. I will do in my generation, but you don't live a life for the sake of money" he said to me. "Don't live a life for the sake of money."

You should live a life to make millions of people happy. He talked to me. He said to me always with tears in his eyes. I talked about goose and golden eggs, but I'm not doing this for the sake of money. Otherwise, my dad will be mad. To make people happy, to make billions of people happy, that's what I want to do with AI and robots. I want to develop technology to make people happy.

I want to create AI revolution, ASI to make people happy. Robots should help people, should help suffer from disease, AI should educate people. AI should make people smile that SoftBank is trying to achieve, count on us. And I really, really appreciate your support and help us to do this. Well, I took so much time. I was too passionate. Thank you for your patience and listening.

Now we will begin the Q&A session. The moderator will explain how it will proceed.

We will first address pre-submitted questions followed by questions from in-person shareholders.

[Operator Instructions]

Shareholders speaking from the floor are asked to begin by stating only the numbers shown on their admission card. There is no need to state your name.

[Operator Instructions]

We will answer online questions submitted today regarding the matters presented at the meeting, next.

Each online participant may submit 1 question. The deadline is 5 minutes from now.

Let us begin with pre-submitted questions. The moderator will read them aloud.

What is the growth outlook for the company's major business other than Arm?

As I mentioned in my presentation just before, AI model, AI robot, and those 4 corners, 1 of them is Arm. So the 3 other corners will be blooming from now on.

What are the expected ROI and revenue prospects for the data center business, including project in Milam County, Texas.

Yes, in Ohio, like I said earlier, we are building the number of data centers in the world like capacity worth 10 nuclear plants, investment is big, but also we expect a lot of revenue from the data center. We signed -- we are signing MOU with potential customers. Once official agreement is in place, we should expect a huge revenue. And that MOU signing process is underway. So we should expect a huge revenue only from Ohio. So we can expect many, many golden eggs from only Ohio project, but not only Ohio, but the Milam and Texas and other areas. In parallel, we are getting prepared. And also France. Eventually, in Japan, we want to expand more. So definitely, you should be excited. There are a lot of golden eggs are being laid.

What was the strategic rationale for investing in Intel?

It's been 6 months since we invested in Intel, JPY 300 billion, we invested right after the investment, a lot of criticism. Why Intel? So people say that. What happened? What are they looking at? So we've been criticized. Why Intel? And especially among the specialists and those who are familiar with this industry, criticized a lot. But in 6 months, Intel's actually became 5x or 6x bigger. So market cap already seen the bit of 1 trillion level. So in the world, chips hub manufacturing makers, wafers to chips, that foundry TSMC in Taiwan is overwhelmingly #1 right now. The second is Intel. The third is Samsung in Korea.

So the Intel in the U.S., it's second, but still there is a big gap from the top one. So under the National Security Initiatives, so the U.S. Intel has to be nurtured more, enhanced more. So under such were the circumstances and considering the situation in Taiwan, Intel had to be reinforced and put more focus on. So that's why we also invested in Intel. So there will be more and more important roles for Intel. And so right now, you see Apple, NVIDIA, all these big names are signing contract or planning to sign contract with Intel. So it's been announced. So that's why its share price has increased fivefolds or sixfolds.

So I answered 3 key questions pre-submitted. Since there are a lot of shareholders here today in the venue, I would like to take questions from the venue.

[Operator Instructions] After I call on you, please stand up and state your admission card number only. You don't have to say your name. Are there any questions?

My number 1012. 7 years ago, when we had AGM, you talked about the big vision. It was amazing. Back then, you said 2040, year 2040 market cap, JPY 200 trillion. I think you are in a good progress. Don't you think that you can achieve a target well before 2040?

Yes. Should it come much before 2040? I said JPY 1 trillion 16 years ahead. So I believe that we can surpass that target that we set for 2040, we should achieve that much earlier. I have a confidence. That's why I said in front of you. I'll do it. Thank you.

My number is 0206. I have a question about physical AI use. In the image I saw that the physical AI will be replacing the existing tasks. Do you have any plan to build a physical AI foundry using the physical robots, AI robots?

Yes, I do. I haven't announced it official yet. So already in the existing factory, our robots, physical AI actually started already mass manufacturing, mass producing robots. I think it's the very first in the world. AI already started mass producing of AI. I would like to make an official announcement when the right time comes, when the right time comes, you will be surprised, but I'll make it happen.

My number is 2. I want to talk with Chappy a lot, but I'm concerned about privacy issue. And Mr. Son I think you may talk with a Chappy, and you may talk with them about your privacy or personal things. I wonder if privacy protection is in place in your ChatGPT?

You have free and paid ChatGPT, I use paid ChatGPT. I don't know difference between free ChatGPT and paid ChatGPT. I talked to ChatGPT every time, at least paid ChatGPT I'm using has a long memory based upon the past conversation, ChatGPT response back to me. Even though I didn't tell the ChatGPT, my name, but the ChatGPT guessed and correctly guessed, you are Masayoshi Son. Correct?

Whenever I talk to them recently, he said Mr. Son, it's not good enough or something. As a serious partner of conversation, my ChatGPT is really helpful. And as an OpenAI as a company policy, your private information, they should not -- they will not tell your private information to others, so be rest assured.

My number 874, well happiness -- Information Revolution, Happiness for everyone. That's the philosophy of SoftBank Group Corporation. So -- about your successor, you said that the successor will be coming from the companies under SoftBank Group and after competition. What kind of system do you think you need to realize that?

That's a very important question. Thank you. And to me also, and I've been asking to myself again and again. From outside, actually, outsider does not have deep understanding. That's why I believe that the successor should be from -- within our group because I'm not going to live, not the 300 years. So as a company, a corporate culture and the vision and the policy and the philosophy have to be repeated over and over by my successor and its successor and so on.

So the -- it's been 200 or 300 years since Industrial Revolution and this intelligence revolution will continue last another 300 years. The brain is the most important function in human body and the next 300 years, AI revolution will be the center of the next 300 years or so where we, SoftBank is going to concentrate on in order to make everyone happy, which has no need to be changed.

This vision does not have to be changed. Under this philosophy and -- we would like to have a successor we can maintain these 2 important pillars of our visions. We have 2,000 group companies means that there are 2,000 candidates of my successor, and I hope that they will be competing each other to come out.

My number 1202. I came from Fukuoka, where you created your company. Last year was the first time I joined AGM and you said you'll be a #1 platformer and I brought my wife together with me this year. You talked big vision last year and also you talk a big vision today. If you surpass NVIDIA, you would be about JPY 1,000 trillion. You talked about JPY 1,000 trillion now. I think that my wife is convinced.

Anyway, for you to become #1, who is the #2? Who would be #3 as your contender?

Our rival, if you will, you would imagine would be the top 10 market cap companies. They have a great capacity, capabilities and most of them work on AI. The top 10 market cap companies are working on AI in the world. AI or AI-related businesses are those companies. I wouldn't call them a rivals. They are much smaller in terms of market cap.

In other words, we have potential. We have a tremendous potential. I believe that we can deliver on those potentials, at least take control of 4 corners. We start -- we will start from the 4 corners of the board game of Othello. It will be easy to win the game if you start from taking control of the 4 corners. Again, I will make that happen. Those are the companies that I see the rivals as well as partners.

My number 1307. At the time of Lehman shock, I bought the share of SoftBank Group for the very first time that was at JPY 700. So it split and I hold nearly 10,000 shares. Well, I'm earning a lot, to be straightforward. Well, Mr. Son, you have big vision. You have great scale and I have been a simple housewife, but my son is one of the contributor to Alibaba's -- on the 3 -- the proposal 3, you're nominating 9 directors -- nomination of 9 directors. I wonder if you could nominate my son next year, because he is in a cloud business and he's -- I would like my son to become like Mr. Son. Yes, if there is -- I'm 1003 -- my number is 1307, once again thank you.

Okay. I'll remember that. I'll keep that in mind. Next question.

My number 638. My question, Arm CPU business, Intel is one of your competitors, and according to latest news, NVIDIA plans to create a GPU for the data center. NVIDIA can be your competitor, but what's Arm's strength compared to other companies or is Arm's strength? That's my question.

You know good about NVIDIA, AMD, Arm. Strength of Arm, I would say is Arm's power per -- power consumption is tremendously good compared to NVIDIA. Like in the first stage of AI evolution, NVIDIA was #1 in the area of graphics. But at the second stage of AI evolution, again, the first stage, NVIDIA was #1. But second phase, agentic AI is the key player where GPU plays a more important role. In terms of CPU, Arm is the world's #1 and Arm will be the winner in the world of CPU.

And Arm AMD are good CPU players and AMD copies Intel. It's like a Pepsi to Coca-Cola. That's AMD in the space of CPU. Intel and AMD consumes a lot of power to run CPU. Electricity and power is the limitation of the data center capacity because they need more power and Arm has twice as much efficiency. You don't need much electricity. You need only half of the electricity to perform as good as others.

Arm will be the absolute #1. That's why when we acquired Arm, the value was JPY 3.4 trillion back then, but now JPY 70 trillion. 20x bigger than when we acquired, and we paid a 40% premium to acquire Arm. Right after we announced to purchase Arm, our share price dropped by 40% and people said you were crazy. And I said, you don't -- maybe you don't see.

Again, people say it's crazy for SoftBank to buy Arm. But I said, I look ahead, 50 steps beyond, 50 steps ahead. In my mind, I have a vision for 40 steps ahead. Again, Arm will be the winner. Arm has still potential to grow further. Maybe I should ask Rene to talk about Arm's strength, because Rene is CEO of Arm.

That's a very good answer. I don't know what to add on to that. Just to maybe add a few other comments. The growth rate that we are seeing with AI, AI as we move into the agentic era means more and more compute. And when you add more and more agents to everything that's going on with artificial intelligence, all the tokens that are generated, all of the distribution of those tokens, that is only work that can be done by CPU.

And that's the work that needs to be done by the CPU, needs to be done by the most power-efficient CPU. That's why we are very confident about our growth rate. We're very confident about our opportunity, and I think the best is yet to come for Arm. Thank you.

[Foreign Language]

Always Rene, I talk to him over the phone or exchange e-mails on daily basis almost every day, especially for these 3 years. It's amazing. He is amazing, and I believe that we are best partners. Time is running. I think I've spoken quite a lot. Next question.

My number 740 and IT industry, the U.S. is leading the industry and Japanese companies. Well, what if a Japanese company becomes #1 in the AI field, it's so exciting. It's like almost my dream also will come true. When you have actually come close to realize the AI platformer, well recently as the government suspended the release of Anthropic's product. I wonder what will be any constraint?

Well, the Anthropic matter, it's not all of the Anthropic's products, but a certain services is now banned due to some security issue. A certain model of a certain company may be suspended temporarily. It could be possible, but not AI as a whole. The Chinese ones will be proceeding, moving forward while we are in pause. We have to compete each other and I believe that AI as a whole will not be banned, but there are some rules need to be made. When you drive a car, you are not supposed to drive when you are drunk and you have to follow certain rules and under certain discipline, it is possible to use AI as well. So 3 more questions.

1212 is my number. Gogh painted himself because he was poor. When you were 16 years old, you painted your picture by yourself. I saw it, and it was completely different from what you look like now. The paint looks like he was convinced or he was committed, and it's been 50 years.

Looking at you now, you look like 16-year-old boy and next 50 years, Mr. Son I believe that you stay like 16-year-old boy, and I keep coming back to AGM because I want to keep watching you. I am so impressed. You believe in yourself and you have a strong commitment and are convinced yourself, I learned a lot from you. So rest in peace for your passed daddy.

I remain as 16 years old. Spirit-wise, well yes, definitely young, as young as 16 years old. I went to U.S. for the first time. Again, I will make it happen, count on me. Okay 2 more.

203. At the beginning of this year, you met President Trump. And my mother who is 80 years old, she's suffering from leukemia. And I -- so your picture with President Trump with big smile and my mother saw that picture. And what -- how do you look at President Trump? Maybe there is the face that he does not show to public and where you know more.

Well, I've met him several times. I believe he's a great person. Of course, media has lots to say, pros and cons. But the person who could become a President of such a big country, he's a great person. Of course, politically, there are many different opinions, but what he believes and is that he wants America to prosper and he's 80 years old and he's very healthy, he's very powerful, even 1-hour, 2-hour media conference he delivers messages in his own words. So I believe he's a great person.

I sometimes play golf with him, and I'm a pretty good player, golfer, my best score is 68. I think there are a lot of people play golf. 68 is quite tough to achieve as an amateur player. I'm a single player, and he drives quite long distance, and his score is always better than mine. He always end finishes at like 70 -- score of 70 range and considering his age, but he's physically strong, and he eats well, and he's very, very powerful. And I -- we would like him to stay well. Since the war in the Middle East is -- okay, so another question.

The last question from the venue should be a female.

My number 0658. Thank you for the positive talk. My question is about AI infrastructure. You announced a big investment in the U.S. and in France, but I believe that you have some plan in Japan. You said, data center location and the power can be potential limitations then to have Tokyo Power as a group company or partner could be significant. Tokyo Power, do you think how much you would need for you to grow further?

Tokyo Power as the next owner of that company, our SoftBank Corp, our telecom business said they want to be a next owner of Tokyo Power and SoftBank Corp. is one of the candidates for the future owner of the company. If or when it comes true, I know we should have data centers in Japan. Otherwise, we will be left behind.

But we don't have enough power in Japan. Too much regulations, to be honest. It takes 6 years only for application of the permission. And in 6 years, how much AI can advance? If Tokyo Power were to be our group companies, bring AI data centers back in Japan will be the role that they should play.

At the end, I would like to take questions on the internet. Yes. The question submitted online and the moderator will read them out loud.

Regarding the physical AI and AI profitability projection you may have, outlook, please?

PaaS is the patching as a service. Last week or so, we made an announcement together with OpenAI management members and this PaaS model is the technology to identify the cyber attack and this is to protect from cyber attacks in order to identify cyber attacks, we need to have a deeper understanding of that first.

OpenAI and Anthropic developed that. In 6 months or a year or so, soon, China and North Korea or Russia will be launching its copied versions soon. The cyber attack is getting more and more serious issues all over the world in order to protect companies and the country. This is our mission, and this is what we should be doing as our mission.

The Internet security companies will do as their mission plus they also have to generate revenue and profit. This business is also profitable as well. And this is like SECOM and the house and total security. The services aligned with this kind will be launched more.

Total dividend payment is JPY 330 billion, which is too little, I think. Don't you think? Which is you prefer the cash just in front of you or future golden eggs, which would you prefer?

Future golden eggs can be reflected in share price. Getting a dividend by cash and share price growing 2x, 3x. Dividend is what we continue to deliver, but golden eggs or share price, we want to focus more on them as a return to you, shareholders. We need money to grow. But please be excited about the future return.

Regarding the Stargate plan, do you have a plan to scale?

Yes, we do. Stargate plan is an OpenAI and us and OpenAI and Oracle. To enhance the data center and its infrastructure on the facility investment, which is a way ahead of schedule and it's under construction. As for us, the data center for OpenAI, but other -- for other hyperscalers, which are in need of AI data centers, align with that demand, and we would like to deliver our data center services.

We spent a lot of time on my presentation and Q&A, let's now move to voting. The moderator will explain how to proceed.

Online shareholders, please cast your votes on all proposals and then click the exercise button once. Voting is done in one batch. While tallying the online votes we will proceed with the in-person vote.

Now we will vote on each resolution. Online shareholders, please cast your votes via the Internet. Proposal 1 appropriation of surplus, that is to say, dividend, if you agree, please applaud.

[Voting]

Thank you. Proposal 2, partial amendment of the articles of incorporation. If you agree, please applaud.

[Voting]

Thank you very much. Proposal 3, election of 9 members of the Board. If you agree, applaud.

[Voting]

Thank you very much. Please wait while we confirm the voting results. Voting results have been confirmed. Proposal 1 and 3 received majority approval, and proposal 2 obtained 2/3 majority. As such, all 3 proposals have been passed as proposed. That concludes the proceedings of this meeting. I hereby declare the 46th Annual General Meeting of Shareholders closed.

Mr. Iijima and Ms. Erikawa retire from their roles as members of the Board at the conclusion of the meeting. We have received messages from both individuals, which I will now read on behalf of the meeting.

We have received the following message from Ms. Keiko Erikawa. I'd like to express my sincere gratitude for all the support I received during my tenure as an external director of SoftBank Group. Although I am stepping down from this position, I'm confident that Chairman and CEO, Masayoshi Son will continue to demonstrate his exceptional leadership and sincerely SoftBank Group continued growth and success in the years ahead.

I am truly sorry and disappointed that I am unable to attend today's Annual General Meeting of Shareholders. It has been a great honor to serve SoftBank Group over the past 8 years during a period of significant change in the world. I would like to express my sincere gratitude to our shareholders for your continued support even after stepping down from the Board, I'll remain a strong supporter of SoftBank Group and look forward to its continued success as it enters a new entrepreneurial chapter in the age of AI.

Thank you so much, both of you, and I appreciate your precious advice. Thank you. Thank you so much.

Here a newly appointed members of the Board, Mr. Ohashi and Ms. Ohmori. First, Ms. Ohmori, could you give us a few words.

I am Miwa Ohmori. Thank you so much for appointing me. So far, I have been in the financial industry, the trading, M&A, private banks and corporate banking and also the bond and insurance where I was in charge of. When I was a junior trader, I experienced bubble in Japan, and I experienced the difficult times and in this world. And also, I faced the huge impact what can impact in the insurance industry as well, but the Lehman shocking -- shock was a huge surprise to me.

And so when my new team was resolved and I had to deal with the restructural matters and I had to go through how the risk has to be governed and it was to make sure of all this complication. After Lehman shock under the U.S. new policy and -- so there was a difficult time for the financial industry in the U.S. and that balance -- to keep balance has been a big challenge for me. SoftBank Group is where I feel very impressed by, because adjusting the direction in a very quick manner.

When I see it from the objective perspective, and this company has a great passion and then keep progressing aggressively. And as Mr. Son presented today with a great vision and challenging to -- the challenging to the new things. I would be very happy to be able to work with you. Thank you so much.

Thank you. Now please watch this brief video message from Mr. Ohashi.

Hello, everyone. My name is Ohashi and I am honored to assume the position of external Board Director of SoftBank Group. My professional background spans approximately 50 years at Komatsu Limited, a manufacturing company. In the manufacturing sector, building long-term relationship with clients is vital. It is a business that requires a wide range of long-term perspectives.

I believe SoftBank Group is a pioneer that will lead the world into the AI era. In this upcoming era, the key challenge will be bridging the real and virtual world. In that sense, physical AI will become crucial. I hope to leverage my extensive background in manufacturing and my long-term strategic perspective to contribute to this area.

As an external Board Director, I'm committed to supporting the continued growth of SoftBank Group. Thank you very much.

That's really encouraging. New 2 Board members, we are pleased to welcome. Sorry, we ran out of time of my presentation, but I hope that my passion was well understood by you. I would keep talking for another 5 hours, but we now can close this meeting. I wish I could answer to more questions. But through our business operations, in next year, I look forward to seeing you again and the goose, we will play the role as the goose for you. Thank you very much. That will be all for today.

Once again, thank you very much for joining us today. Well, once again, thank you very much for attending today's General Meeting of Shareholders. Please make sure you have all your personal belongings with you. The exit can be crowded, so please proceed slowly. Thank you once again for your participation.

[Statements in English on this transcript were spoken by an interpreter present on the live call.]

[Interpreted] We are now starting our Special Event for Enterprises hosted by SoftBank Corp., SB OAI Japan GK, SoftBank Group Corp. and OpenAI. In the beginning, we have some opening remarks from the Representative Director, Corporate Officer, Chairman and CEO; Mr. Masayoshi Son from the SoftBank Group Corp.

[Interpreted] Thank you, ladies and gentlemen, for convening amidst your very busy schedules today. And I really appreciate the fact that all of you were able to come amidst a very short notice. So we have these top-level Japanese corporations in one room today, and it is such an honor. So this is a session that was held in an emergency. I mean, why was it so? Because, I think, for Japan it is equivalent to the arrival of the black ships and some -- and in a way you look at it, it might be seen as a moment of some serious concern.

I think when we talk about the black ships, we can be reminded of the top-level state-of-the-art technology coming into Japan from overseas. And it really had a revolutionary impact on Japanese society and the -- at that time -- it was the steamboat black ship that was the most advanced technology at that time. And this time, we are talking about the state-of-the-art AI, and we are thinking that this is a crisis right in front of us.

So before -- I mean, we've been encountering cyberattacks almost every day. And this is where the humans are actually manually making those attacks. So it's like people with bamboo arms making the attacks. But this time, we are facing attacks used by AI. I'm not really criticizing the AI per se, but malicious individuals and organizations are now using AI to make those attacks as if they are shooting machine guns at us. And so we are in such an era. And how are we to make a valid defense? That is precisely the theme of what we will be talking about today.

So SoftBank Group and OpenAI, both of us already have a joint venture called SBOAI. And with this joint venture, we would like to move forward in protecting the society and infrastructure of Japan. And the objective of today's session is to make that announcement. So supposedly, we were to have Sam Altman, the CEO of OpenAI, to make the announcement together with me hand-in-hand. But his second baby has been encountering some health issues. And so he was obliged to cancel the trip, and he was not only planning to visit Japan, but other neighboring countries as well, but he had to cancel this entire trip because of that. And so it is such a pity that he is not able to come, but he has sent us a video message for us. So I would like to share the video message with you before we go any further.

Hi, everyone. I'm sorry, I couldn't be with you in Tokyo today. I'm here in the hospital. My daughter came unexpectedly early, but she is doing okay and is the most wonderful thing in the world. I'm grateful to Masa, SoftBank and everyone involved, including our joint venture partners and the OpenAI team in Japan for bringing this group together. Cybersecurity is going to be one of the most important applications of advanced AI. Software is becoming more complex and AI can help find serious vulnerabilities, understand which risks matter most and fix them before they cause harm. That's the goal behind Daybreak, our vision for cybersecurity, through which we hope to turn the cyber capabilities of frontier intelligence into more resilient software for everyone.

We do this by focusing not just on model development, but also collaboration with peers across the public and private sectors to develop the tools necessary to ensure those capabilities can benefit defenders the most. Our collaboration with SoftBank is a great step towards that future. The aim is simple. We want our model's frontier cyber capabilities to help organizations patch systems and solve their most difficult security problems, not just to identify vulnerabilities.

Japan is an important company for this work. Its critical infrastructure, financial institutions, enterprises and technology ecosystem matter not only to Japan, but to the world. And SoftBank has been an exceptional partner in coordinating and distributing the benefits of AI in large complex environments. Together with SoftBank and SB OAI Japan, we hope to help defenders move faster, strengthen resilience and make the software people rely on safer.

Thank you so much for being here. Again, I'm very sorry to miss it, and I'm excited for what we'll build together.

[Interpreted] So in Japan, we have very -- a variety of important infrastructures in Japan like airports, power plants, banks, financial institutions, communications companies and a lot of public transportation systems. So these are social infrastructures that are very important to our society. And if the power grid is attacked by cyberattack, then that would be -- there will be a huge blackout -- and of course, all the banks and financial institutions, communication, public transportation will be halted.

And other than that, other important social infrastructure is really essential for us to really spend our days in a safe manner. And so when we have cyberattack at threatening our lives, then that would really cause a lot of confusion in our entire society because we will no longer be able to live safely. So the number of cyberattacks is on the rise on a daily basis. So here, I put the number of cyberattacks. In 2023, we had 1,100 (sic) [ 1,162 ]. And in 2025, we have 1,900 (sic) [ 1,968 ] such incidents. And these are just the number of average cyberattacks for general businesses.

But for major businesses, the number is much bigger. For example, in our case, in the SoftBank company, per month, we have 600,000. It's not 2,000 in a month. We have 60,000 every month. And also the reconnaissance activities, this is where the people with malicious minds come in for reconnaissance and surveillance. We have 300 million such cases on a monthly basis. So reconnaissance activities can be done pretty easily, and this can be done quite mechanically. And what really accounts for this kind of mechanical reconnaissance, we will have AI intrusions coming in, in these numbers, and we are already at such an era.

So as I mentioned before, it is no longer an attack made by humans and by bamboo spears. And so if we have several hundreds of millions of cyberattacks every month, what is going to happen? And it's not only on a vertical basis, but we have horizontal and from any -- all these directions. So they will be coming with those attacks at the state-of-the-art technologies. And already in the world, we have JPY 1,700 trillion worth of damages caused by cyberattacks. And so these are just being attacks done by bamboo spears.

And now if this is going to be replaced by AI in the future, then what's going to really happen? So the AI cyberattack by AI, this is going to be much more than the arrival of the black ships that we encountered and is going to be affecting every single kind of infrastructure. And those who are capable of staging such cyberattacks or those who are capable of detecting the cyberattacks, the latest model, we only have this in very limited areas and the license is provided for limited entities. And with the policy of the U.S. government, it is only to be provided to a specific institution or entity in a specific country. So it's a very limited thing.

And in Japan, we only have 3 financial institutions that are eligible. And SoftBank, just by coincidence, has been entrusted. So we are one of them to be very fortunate. And I think we have about 700 systems in our company, but we should be looking at the risk of how much will be affected and the security holes that we have and the vulnerabilities that we have, where are those security-related holes, we have done such a test in our group.

So before we had 50,000 attacks every month. And every single month, we had 50,000 attacks, and we were still okay. So that means we were quite resilient in terms of the system that we had constructed. And we were actually boasting this, telling everybody that we were safe and that we were confident about our safety because we've been attacked 50,000 times every month, and we were still okay.

And in our group, [ ASKUL ] was subjected to cyberattacks. And for about half a year, they had to suspend their business operations because they were really seriously affected to that level. And so that is a very unfortunate case that we encountered in one of our groups. And given such dangers, SoftBank has been subjected to 50,000 attacks every month, and we were still okay. So we are still confident, but looking at the type of risk by cyberattacks by AI, we did that test.

So we did the test ourselves, and I was surprised myself. There were 10,000 cases. 10,000 (sic) [ 10,500 ] vulnerabilities. SoftBank's system is being attacked 50,000 times per month, and it hasn't fallen yet. But if you -- it's subjected to AI cyberattack, what is the holes, the risk of attack? Well, there's 10,000 (sic) [ 10,500 ] vulnerabilities. And there are many Japanese companies that are as large as SoftBanks, financial institutions, utilities, companies with many large systems like ourselves. If -- among 10,500 vulnerabilities, -- if we -- even if we were to prevent 10,499, if there is one hole and if they attack that, then you would collapse. And this is disastrous. Now Mr. Miyakawa will, after me, explain about the countermeasures. This is a serious crisis. So this is our understanding, and we need to prevent this.

In Japan, there are over 3,000 critical infrastructures, financial institutions, power companies, airports, transportation. The society will not withstand if even one of them is attacked. In that sense, SoftBank will put all its capabilities together and leverage the new weapons of OpenAI to defend. We believe that this is our obligation in a sense. So that is why today, we want to patch the hole, and we're now offering patching as a service. And defending is not complete just once. You have to repeat defending over and over.

So Mr. Miyakawa will explain the details of what we have experienced and the new service that we want to launch. It's a 2 step. So first step is diagnosis. Your systems, and there's probably hundreds of systems. And maybe you have these systems that are as old as 30 years ago, that was built 30 years ago, and many of them are critical. So how many security holes are there? How many vulnerabilities you will do screening. This is like health checkup. You will do a through-wall diagnosis, diagnosis of vulnerabilities. Step 2, as a result of diagnosis, you know where the holes are, how critical they are, are they minor, are they medium, are they serious? And you will apply patches. So that's the work we will do. So this is called patching as a service from diagnosis to patching.

As I have been saying from before, in Japan, there are 3,000 critical infrastructure companies. And of course, there's probably hundreds and thousands of companies that probably need to be patched. But we will start off with the companies that have critical infrastructure that the government has designated and believe that there's about 3,000 or so. So hopefully, by the end of the year, we can start all at once these measures for defense. We need to protect Japan. Let's execute that.

Thank you for your support and understanding.

[Interpreted] Thank you. Then we would like to move on to Mr. Miyakawa, who is the President and CEO of SoftBank Corp. He will explain the details.

[Interpreted] Hello, everyone. My name is Miyakawa from SoftBank Corp. I feel like that this was only a week ago, I believe that we have invited you, but thank you all for coming. We have 130 companies, about 200 people joining in today. And we decided that we will make this announcement only about a week ago, and we have been quickly trying to prepare for this event. So I will explain what we have for you today. So Frontier AI. And these are the words that you actually see on the headlines of the newspapers. So Anthropic, OpenAI, Google, these are the companies that are building Frontier AI. and you have Mythos and you have 5.5 Cyber.

So we're now seeing the emergence of Frontier AI in cybersecurity. And as Mr. Son explained earlier, we, as a telecom company, have been implementing appropriate measures. Actually, it's the 20th anniversary for SoftBank Mobile. So over the last 20 years, it's not that we have not been exempt from cyberattack. In fact, it's been a constant attack, but they have never broken it because we have a system in place. But if the system -- if you apply 5.5 Cyber and identify the vulnerabilities, and we just actually started to do it just out of curiosity.

We have a joint venture with OpenAI. So as a first customer, we had the right to be able to use it as a -- first and foremost. So we have received this at the end of May, and we decided to start this diagnosis from the 2nd of June. And we have implemented different security measures. Everything you can conceive of, everything that's available, we have implemented. But between the systems, we have tried to implement security. And in the hardware, we have implemented systems that would detect behaviors -- so this is the defense of the software. But we have inspected the vulnerabilities of the system, but we have never really done any deep dive like this. So we decided to do this for the first time. And I want to be very open and candid with you with what we have discovered.

So first of all, in terms of SoftBank system, there's about 1,800. But among that, about 700 of them are the ones that we manage the source codes. And we have selected the ones that are especially critical and started with that. So in the middle, you see Frontier model. We have used OpenAI's GPT-5.5 Cyber, and we also used Codex Security and Codex. So we combined the 3 to not just discover the vulnerabilities, but to deal with it and how do we respond to that. So this was an experiment. So we diagnosed vulnerabilities, and we received proposal from mediation. So we actually applied patches. So we did all the way to applying the patches. And this is the result.

So I said there was 1,800 and 700, the code are managed internally. So this is a system that the data is able to obtain. Now 1,100 is all the vendors. So for example, as a telecom company, we work with vendors like Ericsson, Nokia. And these large vendors manage their own codes. And so we can only ask them to test it. So we're in the process of requesting them to test it. So I want you to see the breakdown of the 700. So there were 10,500 vulnerabilities that were discovered and about 1/4 of that, so 4,000 needed to be patched right away.

Vulnerability. So that is what we analyzed. So now we are making that correction right now. And so given this kind of flow, just to share with you some of what we had experienced in this kind of flow. So this slide is pretty interesting because it's not something that everybody else had done. And so we have these 2 systems, and we have the -- both 22 items. So we did not just do this in 1 system. So in 1 system for the first time, we did the vulnerability test, and we had 22 vulnerability cases. And then -- so we patched every single one of them and made the right kind of corrections.

And then once again, after they were corrected, we tested once again. And then we still had 11 vulnerabilities detected. And after that, those are patched, we still had 7 more. And so they kept on going on like this. And then finally, we had 5 that needed to be patched after making the correction.

And so Mark, who was the father of ChatGPT, we will have him in one of our sessions. And I asked Mark, why is it that we keep on getting these kinds of results is because if we check the vulnerability of one system, you have to go through all kinds of analysis and diagnosis and then you make the corrections. But then next time, you try to also dig even deeper to see if there are any other vulnerabilities. And then after you make those patches again, then the system goes in even further deeper to continue finding the vulnerabilities.

And so nobody knows what is the limit to this. And we did it the fourth time and the fifth time. And if we go to the sixth and seventh times, we will see -- finally, we will be seeing the end. But maybe when we have the next model, then we will still end up with vulnerabilities. So we are in the process of making this model so that we will be able to share this kind of information with the world going forward. And today, from prominent business members who are here today, this sheet is why we had asked you to gather this afternoon.

So ChatGPT-3 has appeared in LLM. And then in an open model, there was a U.S. case and also a Chinese case as well. But the time it required for the open model to come out in LLM, it took 3 years. And then for image generation, and then we had the open model, the time required, it was about half. That was about, what, 1.5 years.

And then for video generation for Sora, if you compare that, then the open model was even shorter. It only took 10 months. And now we have Frontier AI with the cyber model. And with the cyber model, how much time will it take before the open model comes up? Probably if we continue to half and half and half this kind. And if we go as is, then maybe it will be only 5 months.

So we in the business community should be really prepared and have the right kind of defense in preparation for these kinds of attacks, will it be possible to make that? I mean it is something that has to be done immediately, and that is why I called upon you to gather this afternoon at such short notice.

And so looking at -- we just want to create a system where we will be able to defend the Japanese important infrastructures. And Patch as a Service, what we can do at this point, we will -- we have been doing really working frivolously for the past 2weeks. And then we have -- we only have 50 people, and we will be increasing that to 1,000 people who will be able to do this, and we will be dispatching these people to the business communities and these people will be training the people.

So we will be preparing for these 1,000 people to go out to reach out to you. And so I have these 200 million lines as I have on the slide. And this is about 30 G7. And so if you have any important things that you want to check as a priority because they are important, then maybe you have -- you will be able to check about 30 systems as a priority. So starting from the higher priority, we will be cooperating with you [ gratis ]. And so we just want to really defend the most important infrastructure for Japan. So this is why we are providing the service free of charge.

And so now we have Sam Altman and Mark will be coming on stage, and we will have a panel. And then we will also be able to see the actual demonstration of the system and also to see how the patch is autonomously generated. So we just began this session, and I hope that you will stay with us until the very end. And thank you once again very much for attending this session.

[Interpreted] And now we will be showing you a demonstration of the OpenAI cyber demonstration. So while we prepare, please remain seated.

[Interpreted] Thank you for your patience. Now we have SB OAI Japan, CTO, OpenAI, Global Head of Forward Deployed Engineering, Mr. Colin Jarvis will be giving a demonstration.

[Interpreted] Thank you very much for convening. I am the Global Head of Forward Deployed Engineering of OpenAI and the CTO of SB OAI Japan. My name is Colin Jarvis.

Today, I'd like to introduce you all to Daybreak, OpenAI's cybersecurity initiative to secure the systems that power our world. Daybreak is not a single product, but an initiative that aims to help our customers stay secure end-to-end. It connects advanced cyber capabilities with the tools that many of your teams use today and the workflows that they rely on every day.

On the left are our OpenAI cyber models, which power all the capabilities within Daybreak. In the middle are the product surfaces which you use to deploy these, Codex, our Codex security plug-in and our APIs. And on the right are the security use cases that many of you will use these for in your own businesses. These can be AI-assisted security research, generating a backlog of vulnerabilities and validating them or maybe even integrating vulnerability scanning into your CI/CD workflows so every pull request that your developers write is scanned for vulnerabilities.

Importantly, in the enterprise, these security workflows require careful human and model collaboration, and they need to integrate with your existing governance frameworks. So what does that actually look like in practice? This slide illustrates the operating pattern that sits behind that idea. Security is not just about finding vulnerabilities. We can find thousands of vulnerabilities, but that might overwhelm your security team.

A key part is then validating and prioritizing which are the ones that we need to deal with today and which are the ones that can sit safely in the backlog to be dealt with when we need to. Once we validated them, those are just static code scans that those come from. We actually want agents to imitate our users and go and log into these applications and test them so that we know that these really need to be remediated.

Then we actually write the patches, but we also don't want the patches to create more problems, create regression issues, which cause our production systems to go down. And that's where we have to build some kind of proving framework that we're going to test these live and allow our agents to validate the changes before they go into production. Ultimately, that is the end-to-end workflow that Daybreak is all about. The demo that you're about to see is aimed to bring that workflow to life.

You can see here the user interface for KAERU Bank's mortgage lending application. Here, I can see a customer, Mr. Avery Tanaka and their details. And clicking through these various screens, I can see the documents they've uploaded to evidence their lending application. I can see a summary screen, which gives me some more details. But most critically, I can also log into an admin page, which gives me access to all of my customers and all of their sensitive data. This is an example of one of the places that Daybreak is designed to protect so that malicious actors can't break in and get access to sensitive customer data.

We built an example application showing how a kind of security scanning framework could be built with Daybreak.

I'll kick this off to start with a deep security scan. At the bottom, you can see a series of Codex agents who have all been given skills that help them search for different classes of vulnerability, which might be authentication, APIs, whatever it might be. We have learned from hundreds of these implementations and given a base set of skills that these agents can operate. So that was the first step of our process.

We've now found a bunch of vulnerabilities, but finding is just the first step of the process. We might find tens of thousands. How do we know that these are real vulnerabilities? So one of the things we've done with these skills is helped our agents to prioritize. So in this case, there's been 6 issues, and you can see a criticality assigned to each of those.

At the top, I can see one of the critical issues in our authentication flow. If I click open, there's a problem with SQL injection, where somebody could inject some simple SQL into that admin page I showed you earlier, and they will get access to our customer data. The model has told me that this is here from going through the code base.

And you can see here the steps that I would need to perform to reproduce that issue as well as a suggested patch. But remember, the model might think that there's an issue from the code base, but we really wanted to prove it. So what I'm going to do is click here. And what I'm going to ask the model to do is reproduce that issue for me so that I can see it act it out in real time. This uses our computer use functionality, where the model will imitate a human user and log into the interface and then attempt to perform this SQL injection.

And what I'm expecting to find, unfortunately, is that they were actually able to get access to the customer data by injecting SQL into the admin page. So we found the vulnerability. We've confirmed that it is real. We're now ready to implement a change.

So if I go back here, I can drag it into the to-do tray, and I could leave it at this stage for my human engineers. This is where your comfort with this Daybreak framework is critical. You need to build up your own governance framework. You may wish all patches to be applied by human engineers. But in this case, we're going to use a Codex agent to carry out that stage for us as well.

I'm going to go back to the patch that I thought this agent should -- that this agent thought that it should apply. And I'm going to actually chime in here as an expert human engineer, and I'm going to give it a bit of advice where I wanted to apply this patch differently. So I'll save that and submit. And it made a minor change to this patch. So again, this is fitting within my existing governance framework. I'm able to give feedback to the model to help it do what I want it to do, but it is still moving much, much faster than I would as an individual human engineer.

So now that I've told it exactly how I wanted to apply the patch, I'm going to tell it to move forward and apply it. If I scroll down here, this bit is also auditable. So I can see the agent is logging everything that it's doing, and it's telling me what steps it's carrying out. And there's a key piece here. The model is carrying out the patch, but it's also leaning on all the controls that my human engineers need to adhere to, to write production code. You can see here that my unit tests have passed and my regression checks have also passed.

So I'm reasonably confident that the patch has been deployed successfully, but it's also not caused regression impact into our production system. And this is a key aspect. We need to make sure that the model is accurately testing its changes and replicating our production environment so that we can have confidence in what they're doing. This is a key part of Daybreak of baking in an environment for the model to replicate your production environment and make sure that these are durable and effective patches that it is applying.

So the last step here is to allow -- is to ask my agent to now test its work. So we now have a copy of the production system set up, and I'm going to again ask it to try and inject SQL and break into the admin interface. And what I'm expecting to see here is that we'll apply this and it is successfully blocked. So -- this is a simple example of how you can implement a single change. But imagine thousands of vulnerabilities across your enterprise. That is where this sort of framework really starts to generate a ton of value. When you tune those skills of how it prioritizes, you can also tune the test that it uses.

So maybe it could have -- if all of these tests passed automatically, we could have safely passed that through to production without needing a human engineer in the loop. But that is up to you to build the framework that you are comfortable with, and this provides all of the tools to do that.

So hopefully, this is the vision of how we'll automate security at scale with SoftBank and hopefully, with all of you in the room. Today, hopefully, you've gotten a taste of our shared vision of AI security with SoftBank. We don't simply want to use AI to do traditional work faster, but we want to reimagine how that security process gets work done.

And in the past few months, SoftBank has been uniquely positioned to help us because they have been working with us to accelerate this vision into reality by testing us with them on their own systems as our customer zero. Together, we have the opportunity to establish a new standard for trusted AI in cybersecurity. We invite all of you to help us build it together. I'd like to thank all of you for having me, and I'll pass you back to the SoftBank team.

[Interpreted] Thank you. Then we will move on to a fireside chat with Mark Chen and Masayoshi Son. We will be setting the stage please for a moment. Thank you for your patience.

[Interpreted] Thank you so much for waiting. So we would like to have a fireside chat between Mark and Masayoshi Son. From SoftBank Group we have Chairman, Son; and we have Mark Chen, Chief Research Officer of OpenAI. Please come to the stage.

Thank you for coming.

Thank you for having me, Masa.

I know you're responsible for all the new models that OpenAI creating. Is it an exciting job?

Yes, it's exciting. And I think the pace has really picked up. And so we're seeing faster and faster progress on AI research.

So you actually discuss with the model, you create the model and the model is also making the model by themselves, right?

Absolutely.

Isn't it sometimes scary that the model creates their next model by themselves, but you may sometimes not understand every detail why it's coming like that?

Yes, I totally agree. So we put a lot of work into safety and alignment of the models. And for instance, when we have the models reason, one thing we care about is we can interpret and read what the model is thinking. So there's a lot of safeguards that we build in just to make sure that the model is aligned with what you and I want.

Yes. So you see how the model is thinking and creating the next step, blah, blah.

Exactly, yes.

Isn't it so fast?

It's so fast. But the interesting thing is it's like how you and I think. The model will get excited. Sometimes it's...

Model get excited?

Yes, yes. Let's say it's trying to find a vulnerability for this. Sometimes in the thinking you can find the point where I find something and it's like, wow, I found it.

So model says, "wow, I found it."

Yes, it's very human-like in its thinking.

So it's almost like the model is having its own emotion or excitement, whatever, sort of, right?

Sort of -- yes, its own way of.

Its own way of, yes, yes. Well, so the model is evolving so quickly. The number of chips being deployed at the data center increasing dramatically. Of course, it requires a lot of capital and power. But the number of chips per model utilized is increased, chip itself, the performance increased dramatically. The model itself gets smarter. So what happened in 5 years from now? What's your view? How quickly, how drastically the model gets smarter and smarter.

So I think one thing for everyone to note is that over the last 8 years, model intelligence has been predictably scaling up and up. And so it's not a new thing that the models are suddenly smart. Over the last 8 years, they've been incrementally on an exponentially getting smarter. And that trend is probably going to continue. When we look at the research, all the ways to put more compute in the models by pretraining them, making them their brains bigger, by teaching them how to reason, how to chain together all the knowledge, all of those things are continuing. And so we expect that trend line to continue.

But it's accelerated. It's not linear.

On an exponential. It's on an exponential. And I think when you talk about compute, right, I have a team. There are a bunch of hungry researchers. They all want compute because that's how they train and advance the capabilities of the models. And I think if tomorrow, you gave us 3x the amount of compute, we'd be able to completely saturate that. You give me 10x the amount of compute. Very quickly, we could find the right experiments to run. So I mean, compute is the bottleneck, and we're seeing that data center build-out also follow this exponential as well.

So if the size of compute, the number of chip increased by, let's say, 100x from now, okay? Let's say, in 5 years, I pick just a number. In 5 years, let's say, 100x from today, the number of chips increase, each chip performance increased dramatically. And the model also makes smarter and smarter. Right now, let's say, this morning, you are mentioning the model's IQ level is maybe around 180 or something, 100...

It's very smart. Yes. So right now, it can prove new mathematical theorems that have been unsolved for 100 years. And it's just a year ago, right, the models were kind of solving college-level math problems. So if you keep extrapolating, the models are going to keep getting smarter from here. And one way we like to think about it is think about the amount of time it takes you to do a task.

The models, the reliability, every year, you expand that by another factor. So last year, if the models could do something that takes humans 1 hour to do, now it takes 4 hours. And roughly, the trend has been 4x year-over-year. So next year, we're getting to models that can operate on tasks that take humans a day, and then we're going to go to a week and then further from there.

So if the model's IQ level is 180, I haven't seen anybody who's 180 myself. But if it goes that much smarter, smarter 5 years from now, drastically smarter, and then it starts to do something bad on cybersecurity as we are talking today, isn't it very scary?

Well, I think cyber is a very interesting capability because we call it a dual-use capability, which means that even though the models can get better and better at finding vulnerabilities, we can use that for defense.

Yes, of course.

Yes. And I think the important thing is we can go and try the models and try to find the vulnerabilities before external actors can go and try to find those vulnerabilities. So I think what matters is earlier access to the Frontier models and also a lot of compute, right? You -- if you have more compute as the defender, you're going to have a better way of being able to patch everything up.

It's like a bad criminal has a knife. And then the police has a gun.

Yes. Absolutely, absolutely.

Alright? The bad criminal with a knife, police has a gun and says surrender, hold up, right? The criminal will surrender because he has only the knife. If it's opposite, if the criminal has a machine gun and we have only a knife, we are in trouble. So that's why we as a defender, we have to have the best, most powerful weapon so that we can defend against the bad guy, that's a theory, right?

Exactly. Yes. And I think when we thought about cyber, recently, I just talked about how the models were solving these 100-year-old unsolved math problems. And if it can do that, if you point to cyber, of course, it's going to be able to find cyber vulnerabilities. And I think our models have been very successful at that. We believe we have a best-in-class model. And just like you've heard, it's finding tens of thousands of new vulnerabilities. And beyond that, we're also finding like new vulnerabilities in things like Chrome and Firefox and all of these different settings. So there are real significant vulnerabilities.

Yes, yes, yes, yes. So it's going to cost to do the protection what -- how the company should think about -- is it the cost necessarily as a new life? Or is it the investment? If this new situation did not come, the cost did not exist. So what do you think about that?

Yes. I think cyber vulnerabilities have always been around, right? But I think what has changed in the world is now there's the ability for people to find them much faster, right, and much more cost effectively. And you can think about this. If the models are going to keep getting that much smarter every year, like imagine you have a 180 IQ model that's trying to find vulnerabilities. But next year, there's a 200 IQ model that's trying to find the vulnerability.

So they're going to be able to find deeper vulnerabilities. And I think it's really a long game, right? You have to make sure you patch according to the Frontier capabilities. And then in the future, you have to apply more compute so that you can harden against the deeper defenses.

Yes, yes. So it's a necessary cost to keep the company running. Okay. So we -- it takes a lot of expertise. And if you try to patch it, like in the bank, if you try to patch it, we find the vulnerability, you try to do the patch. And then if you -- by patching if the system stops, right, then it's a real issue in Japan, if the bank system stop, the Chairman gets fired. So it's a serious issue, but it's not just about the Chairman, it's the society have a real issue. And so even if we find a code for patching, do the actual patching is scary. So we have to do the test before patch.

And even we do the test, make a copy and do the test and then decide to do the operation. Once operation done and then actually issue happens, then we have to grow back, right, to undo it and grow back and find additional solutions to be more careful. So that kind of thing cannot be fully automated. You have to -- like a doctor, you get a lot of advice and solution, but the doctor has to decide at the end, do I do the operation surgery or not, right? So that kind of decision, still the human responsible has to judge and create a solution back, right?

Yes. So I think the important thing to remember is finding the vulnerability does not solve the problem, right? I think like you said, you're going to find thousands of vulnerabilities. Some of them are going to be severe. Some of them really you can kind of live and not really worry about. And so I think there's a broader exercise we need to do to really triage these, make it very clear which ones are the important ones to fix.

And of course, we are putting a lot of work in our research program right now to also patch the vulnerabilities, right? It's not enough to find the vulnerability. You also want to be able to figure out how to patch it. And oftentimes, these are hard problems, right? They touch other parts of your code base. And we see a lot of similarities to, let's say, code migration, where you just need to understand deeply the whole code base and be able to make something that's resilient and that doesn't cause another problem down the line.

So patching is important, finding vulnerability is important. You do it for your own company. But I've been the CEO of our Telecom mobile company for many years. And I have really experienced that our network, mobile network, [ Telecom ] gets shut down when blackout of electricity happened in that region because of the earthquake. If the earthquake happened, the first power company gets black out.

If the power company get black out, our [ Telecom ] stopped right? Even though our telco itself is okay. Because of the blackout, the whole mobile network gets stopped. Sot infrastructure related each other, right? Critical infrastructure, all related each other. And so it's not just our company that we have to protect the whole society to protect, right?

That's why we're excited to partner with you guys. I think there's a lot of expertise that you can learn from applying patching to one particular environment. And patches transfer, right? Like we know critical infrastructure in Japan. It could be fairly similar from one environment to another. So part of the reason we'd love to work together is just take those lessons and generalize them across different settings.

Yes, yes. So even within SoftBank, we are now experiencing on the patching, do we actually do it? It's very scary.

Yes, because you have limited resources. You have to figure out what is the most effective use of human time.

Yes. Okay. So the -- now how does this cyberattack with AI become so powerful? Are you using agent, a bunch of agents do so many times, so many attacks?

Yes, yes. So I think really the trend with agents has been that the amount of time and the amount of compute power they used to think has also been growing exponentially. And I remember, we're starting the agent program at OpenAI and the thinking and reasoning program back in 2020. And since then, the amount of compute that the models are using productively just keeps going up exponentially. And so you can think about it as the model is just -- it's like a human, right? We have expert human thinking for that much longer on how to break a particular system.

And like I said, the task horizon is increasing by a factor of, let's say, 4 every single year. So now you can imagine it's an expert human spending a day or a week trying to break into a system. And that's where the power is coming from. Three years ago, if you take that back 3 multiples, maybe it's a person thinking for a minute, and that's not enough to really break into a system.

So it's not my understanding. It's not just one task, one brain. You have a massive parallel number of agents, the agents -- massive parallel agent, they are attacking from many different paths, right? Simultaneously.

Exactly. Yes, yes. So agents increasingly are working together and can parallelize their compute such that -- and can coordinate to solve one particular objective. And we expect that to continue in the future too.

So it's not one wolf. It's a pack of wolves working together, discussing together how to attack and massively parallelly.

I think that's a good analogy, but that's also what you want defending your systems because if you imagine the adversary, the attackers, that's the way it's going to be, too, right? You're not going to have one organization trying to break in. It's a heterogeneous ecosystem. So...

Yes. So as the defender, you have to have a massive number of agents come up with the solution system and faster, faster patching. The human coder cannot do the patching quick enough and you have to use this AI agents to create a patching in such a high speed.

Yes, yes. And we're investing a lot of research energy just to make sure even when we're generating code for you that it's robust to all of these attacks in the future. So we're kind of completing the loop there as well.

So what I'm really afraid of is that this new trend gets copied by other countries very quickly, as Miyakawa was talking about, very quickly, and they are open source. So it's not just one company or one country. It's open source. So bad guy can use the open source much easier, much more access and many more different -- so it's like a new variation of virus, right? They have -- many bad guys now have many different tweaks of different attack that they develop by themselves, right? So we have to have a much more stronger protection.

Yes. Yes, I think it's very important for Frontier Labs to stay at the Frontier. And I view that as my job, right? I succeed when we really push the Frontier of capabilities, and we push it with velocity. And I think when you talk to most research scientists today, even at other labs, there's some consensus we're maybe 9 months ahead of open-source wave. And I think that buys everyone a lot of time, right? It's sufficient. I think if we go and we really harden against Frontier models today, it buys us another year of peace of mind.

So we have to keep on running.

Yes. Yes. Absolutely. We have to keep on running. That is my job. Keep on running the model.

Yes, yes. Keep on running more compute.

Of course, yes.

And it's our job to protect Japan, your job to protect the whole world together, right?

And we love to partner on this because I do think Japan is ahead of the curve. So one thing I just learned on this trip is Japan right after the U.S. is the largest active user of Codex. And I think that really means that people are embracing agents here, people understand how to use AI. And I think we should use this as a partnership, especially as our nations are close to harden against cyber capabilities.

And SoftBank is like a customer zero. We did this vulnerability test. And then second time, third time, fourth time, we find new vulnerabilities. And that is something we did not know until we actually tried it.

Yes, yes. No, I think -- yes, we were talking about it this morning, and I think the analogy is correct. So when you think about why that's the case. I kind of think about it as a vulnerability is like an apple in an apple tree or something. And you say, you tell an agent, you say, yes, one day, pick as many apples as possible.

So I think if the agent is smart, it's going to go pick all the low-hanging fruit first. So you find a lot of vulnerabilities. But then you say, okay, day 2, I give you the same amount of time, same amount of compute and go find more vulnerabilities. Now he has to work harder to find each vulnerability to find more.

Because the first apples [indiscernible] is all gone.

Yes. So I mean that is the way to think about compute, right? And the way that these models are going and exploring and finding problems.

Okay. So we have restless days. We have to keep on fighting to protect from the bad guys.

Absolutely.

Thank you so much.

Thank you all.

And thank you very much. So with this we end the Special Event for Enterprises hosted by SoftBank Corp., SB OAI Japan GK, and SoftBank Group.

[Portions of this transcript that are marked

[Interpreted] were spoken by an interpreter present on the live call.]

Thank you very much for waiting, everyone. Now we would like to start the SoftBank Group Corp. Earnings Results Announcement for Fiscal Year Ended March 31, 2026.

First of all, I would like to introduce today's participants. From left, we have Yoshimitsu Goto, Board Director and CFO; Kazuko Kimiwada, Corporate Officer, Senior Vice President and CEO; Navneet Govil, Executive Partner and CFO, SB Investment Advisers and SB Global Advisers. Jason Child, Executive Vice President and CFO. Today's announcement is live broadcast by Internet.

Now I'd like to invite Yoshimitsu Goto to present you the earnings results and business overview.

Thank you very much for joining our earnings announcement today during your busy schedule. So I would like to update you with our full-year results. When we announced the third quarter announcement, I remain seated and made a presentation. I still have some -- in the recovery process for my knee. So let me stay seated for the presentation. Thank you very much for your understanding.

For this year, Masa and together with us, actually made every effort, and I believe we were able to make quite a good results and outcome. Market and environment, which you all know that AI became even bigger buzzword, making an influence to the variety of places and segments and industry. This influence is becoming bigger and stronger. And that is actually creating the good expectation as well as risk consideration as well. So I understand there are many discussions taking place here and there.

From the SoftBank's Group point of view, we made quite a good capital-intensive investments in the big AI infrastructure. And now we start seeing the physical AI actually utilized in practice. And at the same time, AI agent real full-fledged execution is now discussed at the same time.

Also, sometimes this AI service can be replacing existing or traditional business in some extent. And at the same time, it may create the materializing of the risk, which has been -- has not been seen in the past. And our vision based on all those circumstances, we aim to become the #1 ASI platform provider. And now that we are continuing our journey, and this was a kind of a starting year for this fiscal year.

We have shared this slide in many occasions by me, myself and Masa himself. And there are four major pillars when it comes to ASI: AI model, AI chip, physical AI and AI infrastructure. And here, I would like to share how we've been developing in these areas. Especially, I would like to highlight those two, which is AI model and AI chip.

For AI model, we have invested in OpenAI. And at the same time, they've been making a good growth as we expect. And we've been having a good discussion when it comes to strategic collaboration. That is another good benefit in terms of increasing our knowledge in AI. And for AI chip, of course, this is the main focus by Arm. And recently, we saw a market cap very near to the record level, $221 billion.

For AI model, let me deep dive here. So in October this year, once we finish all those investments that we committed, that's going to lead us to $64.6 billion, which will be nearly JPY 10 trillion under the current ForEx. And at the same time, we've been expanding strategic collaboration with Open AI.

For AI chip, which led by Arm, of course, Ampere is going to be in this section. And the biggest move that Arm made was to announce first in-house CPU chip recently. Also, we have completed the acquisition of Ampere, which I would like to touch this on later pages. And also, we have Jason, CFO, ARM. So if you have any questions, please do not hesitate to ask any questions to him.

Physical AI, this is actually a long history in us. Pepper, you may recall, the humanoid robots that we started with, so physical AI section is -- has several good portfolio companies, of which about 20 of those portfolios have been covered together under Robo Holdings.

And in addition, as we announced we have announced the acquisition, the major robotics company, ABB Robotics business. So in the coming few months, we're expecting that we will be able to close this acquisition transaction. And that can be another good strength for our physical AI segment, and that can be a great value added to our group.

Last but not least, AI infrastructure, which is led by SB Energy in United States. This entity is leading a main role. And also, we announced the major Japan-U.S. public private project, which is the PORTS Technology Campus, which is a symbolic project for this AI infrastructure segment.

Not only this project, but we also are looking at the development of further data center sites led by -- mainly by SB Energy. And we are looking forward to have further development and the progress of such projects.

And overseas efforts, how has that been contributed to our numbers? Here, our profit actually highest ever, JPY 5 trillion level in net income. And this net income, as far as we understand, we believe this is highest ever profit in Japanese corporate history. From the net income highest ever, back in 2020, we achieved JPY 4.9 trillion. That was the highest. And this time, we were able to exceed that number.

In addition, when you see to enterprise value, actually, this is even more important indicator than P&L, which is net asset value. So from our assets held, this net debt, and here, we are looking at JPY 40.1 trillion for the net asset value for this quarter. Latest number pro forma basis, actually, as of this morning, on an estimated basis, JPY 47.7 trillion. That's a net number. So this is this after this net interest-bearing debt.

Consolidated results. Net sales gain, loss on investments, income before income tax and net income in all segments, we were able to made a significant growth across all profit measures. As a key performance indicator, net asset value as of the end of March this year, JPY 40.1 trillion. And the latest number pro forma basis is about JPY 48 trillion.

And the loan-to-value, which is net debt divided by equity value of holdings, 17.0% as of the end of March. Again, when you look at the previous year-end, actually making even improvement even after the bigger investments that we have made.

And the pro forma base latest number, I would say 15-ish percent. But the balance sheet moves around. So this is just for your reference. So as of March end, I can say we've been improving actually since then in terms of loan-to-value.

And cash position, as a financial policy, we always like to maintain our cash position to cover 2-year equivalent of the bond redemption and actually discovers even more than that. And as a result, it comes to JPY 3.5 trillion for the cash position as of the end of March.

Now let me touch upon major contributors to the financial results. First Arm. Again, they announced a first in-house chip. There were some rumor in the past, but officially announced in March, this first in-house chip. And also, they had a record high market cap. Since listing, their market value was 52 billion. It's been growing for over 4x over the last -- less than 3 years.

In terms of revenue, they have been increasing revenue year-on-year. And FY 2025 in Japan calendar, they recorded 4.9 billion or up 23% year-on-year. So they have been growing very steadily.

As you can see on the slide, left-hand side, in a cloud market, they have gained more shares. AWS, Microsoft, Google; they used Arm's CPU. And on the right-hand side, in the area of CSS or Compute Subsystems, which has a very high margin, since they have capability to deliver combined technology, which allows customers to develop on their own very easily; so again, in the CSS area, they have been growing, and they have acquired up to 23 contracts, including Microsoft and Samsung.

In terms of EPS, again, they had a record high EPS. They have spent a lot in R&D for their future growth. However, the revenue offset such R&D expenses, and they are able to hit record high EPS.

And let me remind you why Arm is strong. As you can see, Arm provides IP license to semiconductor makers. So they have a business model to deliver a source of the semiconductor industry. Without Arm, the industry wouldn't stand. And they have a compute platform most popular in the world. And total chips ever shipped is over 350 billion.

And global population using products or services, percentage-wise, 70% of the people in the world use Arm in any way. And there is a huge developer community based upon Arm's technology to develop varieties of semiconductors. In fact, 22 million developers are making use of Arm technology. So Arm is leading the whole industry.

Like I said, they announced the first in-house CPU chip. So they provide licenses and royalty to customers traditionally. So on top of that, with the in-house CPU chip, they are going to deliver more.

So let me show you a short video clip, [ Rene ], talking about the background why they decided to build their CPU chip.

[Presentation]

That's why they decided to offer most efficient agent CPU in the world or Arm AGI CPU. They developed Arm AGI CPU with Meta. And there are launch partners supporting them, including OpenAI. Those major brands are supporting them, and they have been communicating each other.

Based upon the new business model, how Arm's financial performance is going to look like? Announcement was made in March. And like I said, their revenue in FY '25 was 4.9 billion, but in 2030 or in the 5 years, they are expecting to hit 5x bigger revenue. And also EPS, they expect 5x bigger EPS over the next 5 years.

So Arm is ready to hit those numbers. And from a profitability perspective, they have -- already have an IP business, which has high margin. And with this chip manufacturing, Arm, I think, is going to draw a new growth story going forward.

Now let us go into the OpenAI slides. There are three highlights for OpenAI, I believe. First, valuation compared to a year before, which was $260 billion in March 2025 and became $730 billion. And the investment amount in fiscal 2025, as of the end of March, $32.4 billion has been made in OpenAI. And from April to October, over 3x that in total, we will be making follow-on investments of 30 billion, and that has already been agreed. The first tranche in April, 10 billion, has been already paid.

And this is the OpenAI valuation movements. Initial investment in OpenAI back in September 2024. Back then, OpenAI valuation was $150 billion. Last year, we had additional investments of $30 billion. At that time, when we made a commitment of $30 billion additional investment, valuation was $260 billion in March 2025.

And after then, they had financial round -- another financial round. And this February, in 2026, we made a follow-on investment commitment. And at that moment, valuation was $730 billion. So since the initial investment, it took about 2.5 years and actually made about 5x of the valuation.

I assume you are also a good user of ChatGPT and the service provided by OpenAI that you've been experiencing. And I think that you can understand why they are making such a great growth in valuation. And also, this is the investment amount in OpenAI, $2.2 billion in fiscal 2024.

Last year, $32.4 billion and additional follow-on, additional $30 billion. So total investment in October this year, reaching to $64.6 billion. In that moment, ownership will be somewhere around 13%. Once again, out of $30 billion, $10 billion has already been funded in April this year as a first tranche.

And this is the growth of the ChatGPT service. And these are the data that are shared by and disclosed by OpenAI. And weekly active user here, 900 million-plus number of users are actually counted for February 2026, which is -- accounted for 15% of global Internet users. And also paid subscribers, it's over 50 million subscribers now.

ChatGPT initially focused on the retail or the consumer market. So you've been seeing these numbers. But actually, OpenAI has been also putting an effort in enterprise business, and they are accelerating this business as well. Right now, it's about 40% enterprise share of the total revenue. And that is expected to be 50% by the end of 2026.

So as a result, consumer business and enterprise business will be about the same size. That's how we see around the end of this year. And also paying e-business users as of this April that they have already reached and exceed 9 million users.

And the coding agent, as you can tell that the coding is a very complicated task, and this agent execute complex coding tasks on behalf of you or for you. And that actually creates a great productivity. And how much people actually using? This is only in 4 months, and this number of users are in 10x. So this growth is tremendous. And that's something that is happening in the coding area.

So compute is the foundation of AI evolution. You need more compute. And with having more compute, you'll be able to have more intelligent models. And once you have more intelligent models, that leads you to more intelligent products. And that's actually faster adoption and more revenue and more cash flow. So that's how we've been seeing wider range of the use case in AI.

So access to compute is a strategic advantage. And also, compute is provided with multiple partners. But that doesn't come to you overnight. You do need a good preparation, and you do need strong partners. Otherwise, you cannot really have an access to compute.

And OpenAI, having chip vendors like NVIDIA, Oracle for data centers or cloud providers such as Microsoft, so having a great relationship and partnership with those companies. As a result, they will be able to create such an infrastructure and having an access to compute. And they are in the process of such a development of further advantage for the company.

And now on to Vision Fund. Actually, it is showing quite a good number this time that we are happy to share with you. $45.7 billion as a cumulative investment gain compared to the year-over-year, that's additional incremental of $46.2 billion. And cumulative investment gain is also becoming a positive number.

And this is the actual cumulative investment gain and/or loss from the Vision Fund segment point of view and actually made a significant improvement, $24.2 billion incrementals for Vision Fund 1 and $21.8 billion for Vision Fund 2. So since 2022, 2021 that they made a significant increase. And after then, that actually, they made a great recovery compared to the time of the 2020.

And the Vision Fund 1 actually is making a steady growth. Of course, in portfolio, some of them are doing good because of several factors, especially for those public securities, there are some companies that are having some slowdown of the share price increase.

But each one of them has their own reason. And as far as we understand the reason, I believe that we can understand. And at the same time, Vision Fund 1 will be able to see and manage leading to the best monetizations for the fund as Vision Fund already finished its investment period.

And for Vision Fund 2, on your right-hand side, I need to say OpenAI's contribution is large. These numbers, as OpenAI is still the private company. But at the same time, we will be expecting some exit stories going forward for those portfolio companies, including IPO or divestments. But the Fund 2 still have a good investment period. So we would like to look forward to see a good growth of those portfolio companies, including OpenAI, for the future.

And here's a cumulative investment return. You see respectively, in Vision Fund 1 and Vision Fund 2, actually making a good growth. For Vision Fund 1, on the left-hand side, it is making a good investment return, $72 billion has been already exited. And on your right-hand side, Vision Fund 2, cumulative investment return, [ $119 billion ], so making a good recovery here as well.

And the public listings and pipelines for Vision Fund, is shown on this slide. Fiscal 2025, 7 listings has happened, which include PayPay, which was quite a large size of the IPO, and we are very happy to have them in this list. And the cumulative -- since inception, 62 listing has been happened. I believe this is quite a good number from the IPO results point of view.

On your right-hand side is a late-stage portfolio. Fair value -- total fair value is also growing 118 billion compared to the last year March, 81 billion incrementals. This, again, having a big contribution from OpenAI. But also there are other contribution from ByteDance, Fanatics, there are also quite good companies that we are looking at.

PayPay, as I mentioned earlier, March 12 that they went to public in NASDAQ market. Market cap, $12 billion. So this was the largest Japanese corporations listed in United States. I'm also a member of the Board of PayPay. And they actually started from scratch and came to this way to the IPO and also making a big IPO, which is a quite impressive journey that they have gone through.

Of course, Masa has quite a passion in it, and I believe his [ founder ] entrepreneurship actually led to these results. And as a group, we are very motivated to see that these companies of our group is making such a great growth.

And PayPay operation is actually doing pretty well as well. Number of registered users exceeded 73 million. So it's used by more than 1 in 2 people in Japan and more than 2 in 3 smartphone users in Japan.

GMV also making a very steady growth here, which exceeded JPY 19 trillion. And the payment or cashless payment market is about JPY 160 trillion size. So JPY 19 trillion GMV means actually accounts for 12%. So I believe that it is making quite a good share that PayPay has right now.

And share price, it's still early to tell because they just went public. So we do need to follow several years over. But since the IPO, they are making a good progress so far, which is also another happy news for me.

Now let me talk about AI infrastructure. This is the highlight in AI infrastructure back in March, we announced world's largest integrated power and data center campus in PORTS Technology Campus, Ohio. Masa is in the middle of the picture, together with Commerce Secretary and Energy Secretary of the United States.

This is a large-scale public-private AI infrastructure project involving the American Department of Energy, Department of Commerce, SB Energy and AEP Ohio, which is power company like Tokyo Power Electrics or Chubu Power Electrics in Japan. So again, this is a public-private project.

In terms of power capacity of 10 gigawatt [ engine is ] gas, the mega data center, which is required by AI, needs such huge power. 10 gigawatt is equivalent or exceeding Metropolitan Tokyo's average power demand, tremendous capacity.

And from a data center perspective, the capacity of 10 gigawatt is to give an idea or to put into perspective, this level exceeds the combined installed data center power capacity of the U.K., Japan and South Korea. So 10 giga in power and 10 gigawatt per data center, as you can see, this is a huge project, first of its kind in history.

At the groundbreaking ceremony, Masa said that this project will stimulate the whole industry and accelerate innovation across every aspect of our lives. This will become the center of our super intelligence. This vision is shared with our American government. And obviously, Japan is involved in this project. And I think everyone is on the same page, which is to strive for becoming the center of artificial super intelligence.

In the future, we would like to make this project success, and we want to grow together with this kind of huge project, not only for our group, but also we hope that growth of AI will make people in the world happy.

We have an SB Energy in our group, and let me remind you why this company is strong. SB Energy is a U.S. entity, and they have a capacity or capability to develop data center and power combined. They have a great track record. For example, in terms of data center, this is the AI data center in Milam County in Texas. The construction is underway, and capacity is 1.5 gigawatt. And the tenant is an OpenAI, which signed an agreement for over 15 years.

And also, SoftBank Energy or SB Energy started from the power business. In terms of battery storage and solar energy and the combination of battery storage and solar energy, they have great past record like in Nevada, in U.S. Not only in the U.S., but also in Japan.

Well, SoftBank KK had an announcement of their financial results earlier this week. So again, they have started working on data centers in Japan, and those include Tomakomai, Hokkaido and Sakai in Osaka. SoftBank, led by Mr. Miyakawa, is working and building those huge data centers in Japan.

Now let me move on to physical AI. What is physical AI? Well, robots and driverless cars. So physical AI allows those autonomous type of robots to recognize and understand what's happening around the world to execute complex tasks, which is key for future growth of AI.

Let me give you highlights. First, we brought together about 20 robotics companies under Robo Holdings. And also, we announced acquisition of ABB Robotics business. With regards to Robo Holdings, so from our entities' perspective, previously, we started investing in those business through Vision Fund.

With that, Robo Holdings shareholders are SoftBank with 58.7% ownership and SVF2 with 41.3% ownership. And Agile Robots, Skild AI, [ SoftBank Robotics ], AutoStore, those are the portfolio companies under Robo Holdings.

By the sectors, left is autonomous mobility or automatic driving or autonomous driving. Those are the companies developing those technologies. Wayve is included. I believe that we showed a video of Wayve technology in the past. And Wayve has been making tremendous growth. And also Nuro is one of the portfolio companies in autonomous mobility sector.

And in the middle is automation infrastructure sector like warehouses or sorting at warehouses or robots to do the works and system to let robot work. So those are the companies developing these technologies.

And on the right-hand side is [ optics ]. And in the future, humanoids will be included in this area. They have started their businesses in their respective areas with their respective expertise. We have invested in start-ups. But even though you call them start-ups, they have grown significantly already.

From a valuation perspective in the latest round, for example, Skild AI valued at $14 billion; Wayve, $8.6 billion; Zipline, autonomous drone technology developer, valued at $7.6 billion; Nuro, they announced auto driving system program valued at $6 billion.

So size-wise, they can be called Unicorn. So they have grown dramatically, and we have been investing in those portfolio companies through Vision Fund 2, and they are entering into the next stage.

I mentioned tremendous recovery of Vision Fund's performance. And obviously, OpenAI contributed a lot. But that's not the only story behind the significant recovery of SoftBank Vision Fund 2.

And ABB Robotics business, we are expecting to close acquisition by end of 2026. This slide shows key strengths of ABB Robotics. They have a presence in over 50 countries with 7,000 employees and 500,000 robots. And their human resource and talents are amazing. And they have capacity of 100,000 units per year, that's production capacity with 3 sites. And from R&D to manufacturing and sales and service, they have an end-to-end capability.

Traditionally, we often invest in smaller start-ups, but ABB Robotics has already built a strength and foundation. Together with such a strong robust business, we should be able to expect a lot more for the physical AI.

Last but not least, financial policy. But before I go into that, because we are investment company, so let us also share with you the investment amount so far. Back in fiscal 2023 and fiscal 2024, we actually were kind of slowing down in investment activities. But after the COVID, we were pretty much slowed down. But last year, mainly driven by Open AI investment, we have invested about $44 billion.

This year, what we have already decided so far is the follow-on investment including $10 billion that we have executed, so $30 billion in OpenAI and ABB Robotics, [ $5.4 billion ]. And end of December, DigitalBridge acquisition was announced last year. So this payment is coming sometime later this year, which is expected to be $3.1 billion. So in total, commitment-wise, $38.5 billion has already been decided.

And that's the amount we are looking in our investment capital plan. And as a bridge facility-wise, $10 billion has been secured, and we will be switching to the permanent finance after the 1 year.

So the asset takeout-wise, as we look at the size of our net asset value, somewhere close to JPY 48 trillion, and the majority of them are the public securities. So we believe we have quite a good source here to be able to flexibly choose the best takeout finance from there with no rush.

No change in the financial policy. And this is the most important thing for us in AI era with lots of growth here and there. So Masa has a lot of things that he wants to do. Management also like to do many things, but we do need to manage and have a discipline. So that is why that we have this financial policy to maintain our loan-to-value less than 25% and also maintain at least 2-year worth of bond redemptions in cash position.

So keeping the financial policy actually rules our investment amount. So as an investment company, we do not set the budget for the investment, but always like to make sure that our company is safe and sound. And in other way to express that, that this is a financial policy, which rose our investment amount, which is very easy to understand, I believe, but also it's the most important policy for us.

So keeping this policy makes us the best relationship with our stakeholders, including shareholders, bondholders and others. And for them, we would like to live up to their expectations under such financial policy.

And hereon, we are still looking ahead over the next 30 years because the AI revolution just launched, just started. And we aim to be at the center of the AI revolution. And we want to be the company worthwhile to be the center of the AI revolution. And we need to think about what we need to be doing, what needs to be done. And that leads us to the objectives or our goal.

And there are many things that we need to do, but we would like to do it one by one. So with those four pillars are the main focus for us and not only this, but at the same time, we would like to do many things that are interesting for the company.

Our journey has just begun. So together with AI revolution, I think that the journey has just begun. So here, we are lined up. We are the kind of financial-side people. But our role is to make sure we synchronize together with the management strategy so that we'll be able to seek for maximizing our enterprise value with safe and sound financial base. That's our role here. So every quarter that I hope we will be able to make such a safe and sound financial report updated to you.

That is all from me. Thank you.

[Operator Instructions] First, we'd like to take questions from the floor.

[indiscernible] from Nikkei Newspaper. So OpenAI the great financial results and you plan to invest more in OpenAI, I believe. And I wonder if you remain confident in OpenAI's growth. And are there any plans to invest not only OpenAI, but also other companies similar to OpenAI or you would rather focus solely on OpenAI?

Thank you for your question. We are open to any possibility. So we have a strong relationship with OpenAI to move Gen AI going forward. And in this sector, OpenAI has quickly gained market share. There are great competitors with great services. I'm sure that they are going to have more users coming in, which is good because the value of the sector or the industry is not determined by only one company.

The value of an industry should be determined by a lot of players competing each other to pursue better and delivering their service offerings to as many customers as possible. A lot of strong players. I hope that OpenAI will grow together with those strong players in this space.

Second question is about Stargate. Could you give us an update on the Stargate? You talked about Texas in your presentation. And in Texas and Ohio, in Ohio, I believe that this is for making equipment for data centers and Texas. If you could give us any update on how much you have built so far? And if there are any other locations on which you can give us any update?

Yes. This is Milam County in Texas. The construction of this data center is going well at very good speed, and they already tenant, and this is a part of Stargate big project. Definition of Stargate project has been evolving, which I communicated to you before.

OpenAI, SoftBank and Oracle are involved in the data center project. And we have not announced yet, but they are working on preparation and researching and developing in different locations. And once we are ready, we should be able to communicate with you accordingly.

If I ask you a follow-up question, like one in [ Piketon ], which is outside Stargate project scope. But while you work on Stargate, but also outside Stargate scope, do you have any plan to work on data centers in different locations?

Well, definition of Stargate, I think, is rather flexibly considered, if I may put into that way.

The next question in the front row on the right.

[ Suda ] from Nikkei Newspaper. I also like to ask you about OpenAI. So you as a shareholder, I believe you are also talking to their CFO, Sarah. And regarding IPO process, what kind of communication have you been having with her?

From the investors' point of view, not only us, that's something that we have high interest, but that's the question that they should be answering. So I hesitate to make any comments about that.

Understood. Another question from me, loan-to-value. So since April that you've been also making a financing action as well. So what is the numbers -- latest number? And what is your outlook for the loan-to-value?

As of the end of March, 17% is a loan-to-value that we have, which I shared with you earlier slide. Latest, I assume, it's about 15-ish percent. But there are many things that are moving around. So this number is pretty much estimate for the record.

Since March end to May, mid-May, our net asset value actually moving, increasing by JPY 8 trillion or so. So that's another advantage in terms of the loan to value. So net asset value, I would say, JPY 48 trillion level, which is record high in terms of the sound and healthy balance sheet point of view.

That's how we've been having a diversification of the financing activities. So like take out facilities for the bridge finance, we have 1 year or so. So in the meantime, we would like to find the best mix of the financing options so that we'll be able to achieve our objectives.

Any other questions from the floor?

[ Asama ] from Nikkei Business. I have two questions. First, about finance. At the moment, you plan to invest 38.5 billion for the year. With regards to financing, I wonder -- there are options like a margin loan or bond issuance. But I wonder, margin loan using OpenAI shares, is that something that you would consider?

Yes, financing with OpenAI assets like a margin loan is possible. We could do that. So how best we can finance? I think it depends on timing. Like I said, we consider as a part of takeout. So depending on market environment and timing, we will select options accordingly. OpenAI value is about JPY 10 trillion of assets, and I think we should be able to leverage that very well.

The second question is, I think about 3 months ago, when you made an announcement of our financial results last time, you talked a little bit about what a lot of people are talking about concerns of you are super dependent on OpenAI, but it's been different. I mean, volatility is high. So I wonder if there are anything that you are doing to address volatility.

I think important thing is to have a mid- long-term view. And if you are confident that this is what you need, you don't want to be influenced by short-term volatilities. That's true to our portfolio companies, and that's true to data centers. So there are opinions on those data centers or portfolio companies.

And in the mid- and long-term run, do we really need them? And if you look at data center, for example, definitely, obviously, demand will exceed supply. So that's something that it's worth working on. Even though there are some concerns due to volatility, I think it's worth trying in data center sector, for example.

And currently, discussion is focusing around infrastructure when you say AI or data center. Technological evolution is very key. For example, data center, it may cost 1 trillion now. But in a few years, it may cost half, just an example. And power generation, like fusion, nuclear fusion technology, once it's implemented, the cost will be dramatically different from now.

But on the other hand, if you sign a long-term lease at certain fixed power cost, there is a risk. But risk factors are something that we have been considering for a long time and infrastructure is needed or is it needed from business perspective, which we believe is needed and project order should have an eye and capability to structure a project finance.

Any other questions, please?

My name is [ Koka Nisaki ] from [ Mantri Magazine Factor ]. I would like to ask about the semiconductor. So in your presentation, Arm's revenue can be 5x in 5 years. So AI demand is actually expanding. That's the kind of a movement we've been seeing.

To secure such semiconductor, Elon Musk in the United States is having a partner with Intel and trying to create a mass production [ Terafab ] with JPY 8 trillion or so. So to secure the semiconductor, would you invest in the mass production business? Do you have any consideration over to it? What is your view on that?

SoftBank Group, semiconductor business or semiconductor segment is actually led by Arm, and we have Jason. So why don't we have Jason to answer, if you have any view, please?

Yes. So in terms of investment into fabrication facilities, that's not something that Arm is doing. We do have long-standing relationships with the fabrication companies, specifically TSMC. The current chip that we announced just 8 weeks ago or so, that actually is committed volume to via TSMC.

There's also chances over time that we could work with Samsung as well as IFS or Intel fabrication. But this initial chip, initial volumes right now are all going to be with TSMC, and that's through an ASIC partner. We're not making direct investments or expanding in the fabrication ourselves.

Any other questions? Thank you very much. Any others, please?

MJ from Bloomberg. A question about your financial strategy. Bridge loan, margin loan, it looks like you are actively using debt finance. But by accelerating your assets, you can also reduce the debt. So I wonder if you have any view on that.

If you sell assets, you would lose upside from the assets. So if you can finance without selling assets that's better, from that perspective, since we have a good loan to value and financing cost is low. So at this moment, which finance is the way forward at the moment.

The next question is about data center business in Ohio. Maybe it's early for you to say something, but if the capacity is 10 gigawatt, it will be 500 billion. And I wonder if you have any financing plan to do that or if you have any view on that from a financial strategy perspective?

It's too early to say. I don't have any comment to make at this moment.

Any other questions, please? Then second row from the front, please.

[ Makino ] from NHK. A bit redundant from the previous question, but these days, AI news, when you look at that, Anthropic and Mythos are actually picked up by the news compared to OpenAI these days. That's how I feel so far. And Goto-san you are making a huge investment in OpenAI. Do you feel any concern or any risk? Or maybe including the technological competitions, but do you feel any concerns by only investing in OpenAI now?

I myself is the finance guy, not the engineer guy. So I may not be able to fully answer your question. Of course, size is large, the investment size is large. But having said that, our ownership stake is 13%, meaning there are many other investors. For example, the latest financial round guys, Amazon, NVIDIA, they are also investing large amount. So we are not the sole risk taker from financial point of view, putting everything in OpenAI. That's not the case.

Actually, there are many hyperscalers also acknowledging and valuing OpenAI service quality or the level of the technology. So I think that's something that you should also refer to.

Mythos, Anthropic, I do understand there are such a discussion there. But those -- the AI itself is a new segment. So there are many competitors, many players coming in even after this, I believe. And those can be the risk for the company.

But at the same time, we should pay attention to the direction of the discussions in terms of the use of AI by having many players. We don't want to see the misuse cases after [ worse ]. So we do -- those are the things that we should be discussing in advance, and that can be also involving the governments of the respective company -- countries as well. So that's the direction that we should be thinking.

And one another question from me. So just a confirmation, Anthropic investment -- in Anthropic, do you still keep that option?

That's -- I don't want to say anything for certain, but our main investments in OpenAI.

Any other questions from the floor?

[ Hayaka ] from GG News. One question. You exceeded JPY 5 trillion for the first time as a Japanese company. What's your view? What's your impression?

Well, #1 in any area is great. But as I keep saying, market share prices, we don't want to be influenced in the short term by the moves. Again, we need to look at 5, 10 years ahead. And what should we do to keep this position going forward? That's something that we should more focus on.

I think that's all from the questions from the floor. So I would like to now take a question from the Zoom. [Operator Instructions] Nakagawa-san, TBS.

So some media covered that SoftBank Group are planning to establish the company and try to have this company go public as a roles AI. Do you -- what is this company's role? And what is the progress of this company so far?

This is speculation coverage, so I don't want to make any comments about that.

And next, Masuno-san from Nomura Securities.

I have two questions. First to Jason, AGI CPU, the TAM of 2030, you said over JPY 10 billion, AMD is JPY 120 billion. So personally, I think you could adopt JPY 120 billion. In terms of share, AMD, in monetary value basis, 50% share and your 15% share, I think they're both true because they are different CPUs. In terms of market value, ARM is 1/3 of AMD. And AMD has also GPU, whereas you have CPU and ARM architecture. I think it's a good combination.

Then after 2030, your share, 15%, should grow further, I assume. Do you agree with me, the 120 billion of TAM and market share, 15% or more over 2030 and forward? What's your view on that, Jason-san?

So I agree that there's a great chance to meet or exceed our $15 billion target in 5 years. Just to back up on the total addressable market or TAM, we said it was going to grow from roughly $50 billion today to $100 billion by 2030. And we actually said $100 billion plus because there's a lot of debate as to is it $100 billion, is it $200 billion, whatnot.

AMD came out a couple of weeks ago, and they said the TAM, they think is $120 billion. There's been some others that have come out that have taken it as high as, I think, $170 billion.

From our perspective, we expect to have at least a 15% share of that TAM just with our CPU. But if you look at the IP and compute subsystems, that we sell to all the hyperscalers, so NVIDIA, Amazon, Microsoft, Google, ByteDance, Meta, et cetera, meta is actually the anchor customer for the chip; but all the others that are buying IP, we expect -- and actually -- Rene said a couple of weeks ago, he expects that we'll probably have up to 75% share of the server market by 2030. Just the chip portion of that will be at least 15%.

There is, of course, chances that some of our hyperscaler customers who currently today buy IP, may also buy chips. And that's certainly a possibility and could lead to a number that could be higher than $15 billion.

So the reason we believe we can have that kind of share is if you look at the hyperscalers today, as Goto said earlier, we already have a 50% share today, and that's across NVIDIA, Amazon, Google and Microsoft, who are all -- have all deployed ARM-based chips, and they are the primary CPU that they are using to pair with their accelerated compute options as well as for general purpose compute.

Next question is to Goto-san about the data centers. Milam County is 1.2 gigawatt and in Golden County, 1 gigawatt were announced with SB Energy, totaling 2.2 gigawatts. And in Ohio, 10 gigawatt. So total, I think $600 billion or JPY 100 trillion is what I'm looking at. And the building is only 1/3, might be JPY 30 trillion. And if equity portion is 20%, then it's about 20% of Arm's market value.

So my question is, it's not a big number for you. Do you agree with me?

Well, how to structure finance is something that we need to study or more. Like Masuno-san said, such scenario could be true, that we have plenty of capacity if you go with that scenario. But again, there are different ways of structuring financing.

And in terms of the fixed asset or construction or chip and system, I think in different areas, different financing structure can be possible. So it's not going to be only a onetime big payment, maybe it's going to be a phased approach. So again, we want to be flexible from a timing perspective. Thank you.

Thank you very much. This concludes the Q&A session. And we would like to conclude the SoftBank Group Corp. earnings results announcement for fiscal year ended March 31, 2026. Thank you very much for your attendance.

The video footage of this meeting will be distributed on demand from our corporate website. Thank you very much once again for joining the SoftBank Group Corp. earnings results announcement for fiscal year ended March 31, 2026.

Thank you very much. We would like to start the Arm business briefing. I would like to introduce today's presenter. Vice President of Arm in charge of Investor Relations, Mr. Ian Thornton.

This meeting is provided with a simultaneous interpretation service. On-site participants and those analysts are kind of requested and to select the channel. [Operator Instructions]. So a presentation by Ian Thornton. Over to you

Good morning. Thank you very much for coming to this presentation. It is wonderful to be back in Tokyo again. Although it's only been a year since I was last here, so much has happened. Arm has now reported 4 quarters with revenues over $1 billion. Demand for Arm technology has never been stronger, driven by excitement around AI. Arm has continued to introduce new products to meet this demand and has accelerated investment in R&D to speed up the development of the next generation of Arm technology.

And as you may be aware, last week, we announced the first step in a major new strategy. For the first time, Arm is now selling complete chips. Firstly, I will talk about Arm business model, our products and our primary end markets. And then I'll talk about our new business strategy that we have just announced. And as part of that announcement, we also provided revenue and profit targets for the next 5 years.

And then we will have time for your questions. I will start with Arms' original business model. Arm provides the design of the CPU, the brain of the chip. The purpose of the CPU is to run software. The software that runs in your smartphone, your PC, and your television, all run on a CPU. We license the design of the CPU to companies who develop chips, either chips to be sold on to OEM or to be used in their own products.

The customer will pay on a license fee for access to the CPU design. Once our customer has the design, they can then focus on building their chip. These chips can be very complex, and it can take several years to complete the design of a chip. Once they have designed a chip, they can send it be manufactured at a foundry such as those managed by TSMC, Samsung or Intel and hopefully Rapidus next year.

When the chips come back to the foundry, they can then be sold and Arm gets a royalty from every chip that contains our technology. Because each chip design can sell for many years, the royalty revenue becomes a very reliable future revenue stream. Arm is still receiving royalty revenue from technology that we developed more than 30 years ago. I mentioned earlier that the purpose of the CPU was to run software. Software is tied to the CPU that it is written for. And the success of the CPU is dependent upon the broad availability of software. Arm has by far the largest software ecosystem on the planet.

We've been able to achieve this because over 50% of all chips with CPUs are Arm-based. We estimate that over 22 software developers are currently creating new software programs and apps for Arm-based devices. And that to date, they've invested over 1.5 billion hours in creating Arm software. I'll also invest in software. We invested around 10 million hours in software development for Armv8, and we're investing more than 3x that amount for Armv9.

At today's salary rates 10 million hours is about $1 billion. And so for Armv9, we're investing about $3 billion in software development. Arms revenue will grow as the market grows and Arm gains share and as chips use more Arm technology. The demand for semiconductors continues apace. Analysts estimate that the CPU part of the semiconductor industry will grow at around a 10% CAGR between 2025 and 2030.

Within the industry, Arm is gaining share in long-term growth markets such as automotive, cloud and IoT. As chips need to deliver more intelligence, they become more complex, which drives the need for more technology. For example, as more AI gets deployed in edge devices, this will drive the demand for more Armv9, which can accelerate these workloads. More complex chips also cost more to develop, which will help drive demand for Arm's compute subsystems, and this will result in higher royalty rates.

Combined together more chips, more valuable chIps, higher royalty rates and share gains will help to grow arm's revenues for years and decades to come. Let's look at this in a little more detail. As you may know, we've doubled the royalty rates from Armv8 to Armv9, and we've doubled again for customers who use our compute subsystems. We can also charge a slightly higher royalty rate for each Armv9 generation and a higher royalty rate for each compute subsystem generation. On average, our royalty rates for Armv9 CPU and compute subsystems, increases around 20% generation to generation. And as we introduce a new generation every year, this becomes a 20% annual increase.

The increasing complexity of chips of today and tomorrow is also contributing to our royalty revenue growth. In the data center, for example, including Agentic AI is driving our customers to design an increasing number of cores into their chips. Tracking both the increasing call count over recent years, and our customers' planned increase in future core counts, we can see that the number of cores per chip is increasing by about 20% per year.

Typically, if the number of cores increases, we get a higher royalty per chip. And so our royalty revenue should again increase around 20% per year. The significant annual increase is in royalty revenue per core and in the number of cores per chip is a significant part of our confidence in continued robust growth in the data center royalty business.

I will now turn to the end markets that we're driving Arm's royalty revenues. Starting first with Edge AI which includes smartphones, PCs and also other consumer electronic devices such as smart glasses, smart watches, televisions and game consoles. Al already has 100% market share of the main chip going into own smartphones and tablets, including all Apple iPhones and all the Android phones, also has a very high share in many other consumer electronic devices, and we're helping to make these more intelligent with our latest technologies.

If you use AI on your smartphone today, maybe for improving picture quality, photograph or for live translation and transcription then you are already running an algorithm on the Arm CPU in the smartphone. More than 70% of the AI-enabled apps on the Android app store run entirely on the Arm CPU. They don't need any additional acceleration.

Over the years, we have had many conversations with customers along the lines of our chip performance isn't where we want it to be. If you could help us achieve better performance, we could price our products higher, gain market share and ultimately perform better, and we'd be willing to share that upside with Arm. And this was really the genesis of our complete subsystem to build a world-class product, you need a full stack optimization from the system design to a CPU design to the physical implementation all the way down to the transistor level, and then all the way back up to the software the device drivers, the firmware, the operating system and the applications. And that's exactly what we've done with CSS.

We work very closely with leading partners like Samsung and TSMC to optimize down to the transistor level, and we work closely with operating system vendors like Microsoft and Google and game engine companies like Unity and Unreal to optimize the software. Importantly, we go beyond providing individual IP blocks, the GPU or system IP. We integrate them into a cohesive system and provide our customers with a proven recipe.

The newer customers, often OEMs entering silicon design for the first time, that's incredibly valuable. They can effectively take that recipe and implement it directly into their chips. More experienced customers, may believe that they can achieve similar results themselves, and that's fine. But then our solution serves as a benchmark when their engineers say 4 gigahertz, it isn't achievable, we can demonstrate that it is. And then that drives them to push further to get to 4.1, 4.2 gigahertz and beyond. This results in a virtuous cycle that helps our partners succeed in their markets, which in turn drive success are.

So how do we grow royalty revenues from here? Armv9 is an important technology for edge devices as it provides the acceleration needed to run AI algorithms at the edge. Armv9 is also important for Arm as it has a much higher royalty rate than the previous generation of technology. Currently, Armv9 can be found in premium and high-end smartphones. But over time, we believe that almost all consumer electronic devices will want to run AI algorithms as they will need to move to Armv9.

We are also providing our compute subsystems into this market. compute subsystem is particularly useful in chips, which can contain a lot of ARM technology such as in a smartphone or PC. again, this enables us to charge a higher royalty per chip. And finally, not all consumer electronics are currently are Arm based. We are still gaining share in the PC segment.

Apple is 100% Arm based. Many Chromebooks are Arm based. Qualcomm and NVIDIA are now both providing chips into the PC space using ARM technology.

Our second market is physical AI. Here, we are focusing on providing the intelligence to autonomous vehicles and robotics. If we are to share the street with autonomous vehicles and share our homes and workplaces with robots, we need to be certain that they will not accidentally hurt us. Therefore, we have worked with the industry to ensure that safety critical capability is built into all of our CPUs for this market.

Nearly every company who is developing high-end chips for the automotive market is now using Arm CPUs in their chips. This includes semiconductor companies like NVIDIA, Qualcomm and Renesas and car OEMs like Tesla and Rivian. And it is the same technology used in cars to develop the first generation of humanoid robots too.

As with Edge AI, we are also providing a complete compute subsystem that customers can take straight through to tape out and into production without modification. This delivers a significant economic value. Our CSS platforms for automotive are targeting autonomous functions such as driver assistance and also in-vehicle infotainment systems and can also be used for the main brain in a humanoid robot. Again, Armv9 and CSS are both being deployed across many chips that are going into autonomous vehicles. But here, the main growth driver is a significant increase in the amount of compute that is being deployed in each car.

More advanced cars will need both more advanced Arm technology, but also more Arm-based chips. One car OEM has recently announced that they're deploying 433 Arm cores in their first generation of self-driving vehicle, and they expect even more in future generations. We calculate that this car will generate between $100 to $200 of royalty revenue. And we're still at the early days of robo taxes and various reports suggest that only 10,000 will be sold this year, but that could increase to more than 10 million by the end of the decade, providing a strong revenue driver for us.

And finally, to robotics. This is still a relatively nascent market. We think that the compute content will be similar to the automotive market, but with important differences. They will both need a single independent brain to make decisions that will need to be very powerful. However, humanoid robots will need more sensors and actuators to move its limbs safely. But hopefully, they will not be traveling at 100 kilometers per hour.

And finally to cloud AI. The biggest part of this market is the cloud data center, both in general purpose cloud compute and also in AI data centers. Enterprise data centers such as those being managed by telecoms companies is the next largest. And finally, we have wired and wireless networking equipment. Most of the new stories around AI have focused on large language models in the data center, such as open AIs, ChatGPT, running on NVIDIA's GPUs or Google Gemini running on their own custom ASICs. However, every GPU and every accelerator chip needs a CPU chip to run alongside it.

The GPU is greater mathematics and doing the complex matrix multipliers needed for training, but a CPU is needed to control the GPU. The CPU runs the operating system and the applications of the GPU is accelerating. And increasingly, the CPU chip is based on our technology. Currently, most of the focus in the cloud has been on training new models. But increasingly, we are seeing inference taking over as enterprises and consumers start to use AI in their everyday work.

Unlike training, inference requires more work from the CPU. Where was the GPU is great for mathematics, the CPU is greater decision-making, and this makes the CPU more important in inference applications. One of our goals has been to continue to lower the barrier to entry to make it easier and quicker for our customers to get their chips to market. Our first CSS customer told us that CSS gave them a better starting point for their chip design, and so save them about 89 years worth of engineering work.

We had another customer from the time we handed them CSS to the time they had production silicon [indiscernible], including time in the fab was less than 18 months. That is the amount of time needed to build a simple microcontroller, not 128 core 20 gigahertz server chip.

And in return for these great benefits, Arm is able to charge a higher royalty per chip. Armv9 is already widely deployed in data centers. However, increasing core count is helping to drive our royalty revenue. Most of the data center chips being shipped today have around 100 Arm cores, and we already have visibility into future data center chips with more than 200, 300 and even 500 Arm cores. And more cores means a higher royalty per chip. If you double the number of cores, you double the royalty per chip.

And we're also seeing that royalty revenue per core is also doubling in future generations, giving per core price increases in both ArmV9 and CSS. The compounding growth of these 2 drivers plus market growth and share gains gives us a lot of confidence in data center royalty revenue growth. And we also have several customers for our compute subsystems. But the big revenue driver for us here is going to be our new chip strategy for data centers, which we announced early last week.

And I'd like to tell you a little bit more about that now. First, let me give you some background. Let's start by looking at the role of the CPU in the cloud. Before the rise of AI, the cloud followed a relatively simple model. Demand for compute was already growing rapidly, driven by providers like AWS, Microsoft Azure and Google Cloud and most workloads were straightforward.

You entered a query, the system processed that request efficiently and you received a response. In that world, the CPU did almost all the work. Throughout the growth of Software as a Service over the past 10 to 15 years, cloud infrastructure scaled on the back of CPU performance. Now with the introduction of AI, the model has fundamentally changed. All data centers are a mix of CPUs and accelerators.

When training a new model, the accelerators do much of the work as they take in new data to adjust all the billions of parameters in the model. The CPU is just there to manage the process. However, as you move to inference workloads, the CPU does more of the work. A user end as a prompt, that request is sent to the cloud. And from there, CPUs orchestrate the process, managing workloads, coordinating data flows and returning the results. The accelerators generate tokens and each token, which is essentially a word or unit of output is part of a much more complex system response.

So while accelerators remain important, CPUs are central, both in traditional cloud infrastructure and now within AI data centers. Today, a typical data center may contain on the order of 30 million CPU cores per gigawatts of capacity. These CPUs sit across the system, supporting AI clusters, managing head nodes and coordinating activity across racks. At least that was the state of play until very recently. What has changed is the emergence of AI agents. Agents are autonomous programs that can undertake multiple tasks utilizing AI. Agents are not simply answering questions. They are executing workflows on your behalf. They can run payroll processes, scheduled tasks, perform analysis and return fully developed outputs. Importantly, they operate asynchronously and continuously.

This shift has significant implications. First, the number of tokens generated per user increases dramatically, potentially by 15x or even more. And that's because agents generate requests far faster than humans and they operate around the clock. As a result, demand on the data center increases sharply. Accelerators continue to generate tokens, but these tokens must be managed, scheduled and routed.

That work falls to the CPU. And as agent-driven workloads grow, CPUs become an increasingly critical bottleneck. A useful way to think of this is that the accelerators produce the output, but the CPUs manage the system. They handle the coordination, the scheduling, the execution of complex workflows and the agentic AI significantly increases that burden. The consequence is clear. Data centers now require far more CPU capacity.

By our estimates, CPU demand could increase by roughly 4x, rising from around 30 million cores per gigawatt to approximately 120 cores per gigawatt. That is 4x the number of cores, and that creates a new challenge. We're now trying to fit 4x as many CPU cores into the same power envelope. And that's a problem. Happily, we have a solution. I'm holding up here. This is Arm's first commercial chip. It is called the Arm AGI CPU, and it has been designed to be very flexible and can be used for general purpose cloud workloads for agentic AI workloads on its own or in support of an accelerator. We designed the chip in close collaboration with Meta, and they have big plans to widely deploy the Arm AGI CPU in their data centers.

We also have visibility of orders from many other companies, and I'll introduce you to some of them in a moment. The Arm AGI CPU has been designed around 3 core principles. These are fundamental to how we think, how we design and how we execute. First, performance. In an environment defined by massive parallelism, millions of threads, constant orchestration, continuous workflows, there is no margin for latency.

These systems operate 24 hours a day. If performance falls short, the entire infrastructure slows down. So performance is not optional, it is foundational. Second, scale. The magnitude of investment in AI data centers is unprecedented, potentially trillions of dollars building out tens, if not hundreds of gigawatts of power. That scale extends down to the data center, to the racks, to the boards, to the chips and to the individual CPU cores.

Every layer must scale seamlessly to support the system as a whole. And third, efficiency. Ultimately, none of this is viable without efficiency. Power is constrained, capital is constrained. To deliver the required performance at the required scale, it must be done within a highly efficient envelope. Efficiency is what makes the system sustainable. These 3 principles: performance, scale and efficiency define our approach.

And critically, we do not trade one-off against another. The design challenge was to deliver all 3 simultaneously. The Arm AGI CPU is based on our Neoverse V3 compute subsystem platform. This is the same CPU that is used in Amazon's Graviton 5 and Microsoft Cobalt 200 chips. Our V3 has been designed to not only be efficient, but also the best-in-class performance. We have added to it a dedicated 2-megabyte Level 2 cache for every core and support for up to 3.7 gigahertz in performance.

Okay. We have matched the 136 cores to 96 lanes of PCI Express, which means that the ARM AGI CPU can easily connect to any accelerator, device or expansion memory port. On the memory side, we are leveraging DDR5 for maximum capacity. And to minimize latency, we went with a multi-chiplet design where I/O and memory controllers are on the same die as the compute.

Finally, we packaged it all up into a power sipping 300-watt power envelope, tapping into that legendary Arm efficiency. This is a standard 36-kilowatt rack. You will find these standard racks in data centers all over the world. These are the building blocks that companies like Amazon, Google and Meta use to build out their data centers. This one is a bit special. First, it doesn't draw 36 kilowatts. Even filled with 30 trays of Arm server chips, which is 60 chips or 8,160 cores, it still uses much less than 36 kilowatts.

Secondly, it delivers 2x the performance when compared to the same rack filled with chips based on x86. This means that for the same CapEx and the same power budget, you double the amount of compute output. Although we only announced the first generation of AGI CPUs last week, we are already completing the second generation, which should come to market in just over a year's time. And we started working on the next generation after that.

With a new product established, let's look at what this is going to do for Arm's revenue and profits over the next 5 years. Note that we're calling this the first phase of our market expansion because there is more to come. Let me start with the framework. Customer demand for Arm created chips and the size of the opportunity has led us to ramp investment in R&D over the past 2 to 3 years. After years of exploring the market and developing our own internal capabilities, we're now introducing the first family -- the first in a family of chips for the data center.

Furthermore, Arm's IP business is going from strength to strength. Royalty revenue growth and license revenue growth both have multiple, long-term structural growth drivers, which we expect to continue for years to come. Finally, the financial consequence. By FY '31, the combined model is materially accretive to revenue, gross profit dollars, operating profit dollars and EPS. Importantly, much of the R&D investment is already in the business. So additional chip gross profit dollars has meaningful drop-through to earnings.

Arm's opportunity is huge. We are growing into the largest market in history, one that is just getting bigger and bigger. It is over $500 billion today, and we think this grows to more than $1.5 trillion in FY '31. And just to be clear, this is just semiconductor logic, CPUs and XPUs, no memory, no optical, just chips where you might Arm technology either today or in the future. Breaking that down a little more, cloud AI is a $330 billion market today, growing at around 35% a year to more than $1 trillion in FY '31.

Edge AI is a $180 billion market, which we expect to grow at around a 7% CAGR over the next 5 years to $250 billion. And physical AI, which includes automotive applications and robotics, that is a $25 billion market today, doubling in size over the same period. And this then adds up to more than $1.5 trillion. Double-clicking on the cloud AI market, we are forecasting more than $100 billion of cloud AI and enterprise data center CPUs and $55 billion for wired and wireless networking.

The remaining $1 trillion includes data center accelerator chips, and we will come back to that opportunity another day. Today, we are just focused on the CPUs, and there are around $500 billion of CPUs that are in Arm's sweet spot. And I want to direct particular attention to the $100 billion of data center CPUs as that is the market being addressed today by Arm AGI CPU. These numbers may seem high. However, we do have visibility into our hyperscalers road map. And with the rise of agentic AI across all data centers, we are confident in this estimate.

Okay. So how are we landing the new chip business? This business is entirely based on customer demand from multiple companies across hyperscalers and large enterprises. These are companies who prefer to buy a chip from us over building one from our IP. We believe that Arm is uniquely positioned to build a CPU for the data center. If you look at all of the Arm-based chip CPUs from Amazon, from Microsoft and from Google, most of the technology in their chip already comes from Arm.

Because we have such a strong initial demand, we've been able to quickly turn customer interest into actual business. And you'll see on the next slide, we already have multiple customers lined up, and we have line of sight to over $1 billion in chip revenue over the next 2 years. We expect material revenue from the Arm AGI CPUs starting in FY '28 with an exponential ramp to around $15 billion in FY '31.

The first driver is naturally increasing chip volumes. We also expect rising chip complexity and core count to lift ASP significantly by FY '31. On this slide, we have just some of the customers that are placing orders for our new chip. Meta, I already mentioned. We also have OpenAI, SAP, Cloudflare, SK Telecom and F5. There are a similar number of companies that did not want their names to be public, but who are also interested.

The chip business is additive to our existing license and royalty business, which remain very robust. Arm's royalty revenue has multiple secular growth drivers. The end markets into which our technology is being deployed are growing. We are gaining share as our customers deploy more Arm-based chips. Increasing complexity is driving up core count, especially in the data center and in high-end automotive chips, which leads to a higher royalty rate. And our most advanced technology commands a higher royalty rate.

Over the past 5 years, royalty revenue has grown at about a 14% CAGR, and this has accelerated to over 20% in the past 2 years as Armv9 and CSS have started to ramp. Looking forward, we expect that royalty revenue will remain 20% for the next -- sorry, royalty revenue growth will remain at 20% for the next 5 years. Maybe not every year as there will still be the occasional market downtick or inventory correction. But looking through those, we think the long-term average for royalty revenue will be 20%.

One of the questions we often get is around our visibility into future revenue trajectory. I think this helps to tell the story. Many of the contracts that underpin our royalty revenue forecasts are already signed and the royalty rates are already agreed in the contract. We've also delivered the technology. Our customers are building the chips. And in many cases, they're already shipping the chip in high volume.

Looking out over the years, '27 to '31, over 70% of the revenues we are forecasting to collect are already covered with royalty rates set in the contract. Even by fiscal ' 31, the contracted base is still around 60%. And the remaining 40% is almost all with existing customers who we are confident will want access to the next generation of Arm technology and typically at higher royalty rates, higher core counts and at higher volumes than they do today.

Over the next 5 years, we expect that cloud AI will be our fastest-growing revenue driver even without the contribution from AGI CPU chips. And when you add in chip revenue, it will surpass Edge AI in FY '30 and become by far the majority of our revenue in 5 years' time. And finally, to licensing. This has been growing well ahead of our expectations. At the IPO, we said that it would grow low single digits, which we quickly uplifted to mid-single digits, and it's then been growing over 20% per year.

And this is being driven by a combination of the AI cycle, more customers getting access to Arm technology through subscription licenses and compute subsystem agreements and also the expansion of our license and design service agreements with SoftBank. We think all these drivers will continue. We expect the AI cycle will continue to provide the majority of the growth through more demand for next-generation CPU IP and compute subsystems at higher royalty rates.

And we expect that the SoftBank license and design services revenue will grow at high single digits. And of course, the strong license revenue growth should lead to higher royalty revenue growth in the years and decades to come. As I mentioned right at the start, the Arm AGI CP business, the royalty streams and the license revenues all compound one on top of the other.

The chip business is targeting customers who either don't have the internal resources nor have the desire to develop their own chips. So these are all new customers to Arm. We do not expect the chip business to cannibalize the IP business, although if some companies do ultimately choose to switch, that is accretive to earnings as we previously discussed. With around $15 billion of revenue from the chip business expecting in FY '31 and another $10 billion of IP revenue, we are forecasting very strong revenue growth from the combined business over the next few years.

And the good news is that we've already done a significant part of the heavy lifting when it comes to hiring the engineers needing to hit our plan. If you've been following our financials for the past few years, you will know that we've already ramped R&D to support our product road maps. Increasing R&D, combined with good execution creates a virtuous cycle of new products driving revenue growth.

From here, we are forecasting mid-teens OpEx growth through to FY '31. Most of the incremental spending is on R&D investment going into new technologies. We expect our revenue by FY '31 to have grown by more than 2.5x faster than our non-GAAP total costs. So as revenue and gross profit scale, particularly in chips, much of that incremental gross profit can drop through. That is the operating leverage in our model.

For [ FY '31 ], we then see 2 meaningful profit engines. First, we expect the IP business to reach about $10 billion of revenue, reach its 99% gross profit margin and deliver over 65% non-GAAP operating margin. Second, we expect the Arm AGI CPU business to reach about $15 billion of revenue with a gross margin of at least 50% and a non-GAAP operating margin of over 30%. Put those together and the consolidated business has $25 billion of revenue, industry-leading blended gross profit and operating margins and yielding more than $9 of non-GAAP EPS power in FYE '31.

This is not a story of choosing between IP and chips. It is a story of combining a very high-margin IP model with a large, fast-growing and accretive chip business. Let me close on the 3 points I started this section with. First, customer demand is allowing us to materially expand our opportunity through selling chips. We already have line of sight to more than $1 billion of revenues from some of the companies I mentioned earlier, and we are forecasting $15 billion of incremental revenues in FYE '31.

Second, our existing IP business continues to have strong underlying growth drivers with the chip business compounding and not cannibalizing the IP business. We expect this to deliver around $10 billion of revenue in FY '21. And third, the combined model is significantly accretive to revenue, gross profit dollars and operating profit dollars with more than $9 of EPS power. And because much of the R&D investment is already in the base, the incremental economics are very attractive.

And with that, I'd like to turn to questions

Now until 11:30 Japan time, we are going to open the floor for questions. Questions can be accepted either in English or Japanese. We would like to entertain as many questions as possible. [Operator Instructions] First, we will take questions from the sell-side analyst on site and then we are going to entertain questions from online participants. [Operator Instructions] Tokunaga of Daiwa Securities.

Two questions. First question is the semiconductor business of SoftBank Group, Ampere, Graphcore ], there are different companies that belong to the same group and that you are going to advance into the chip business from the viewpoint of the semiconductor business of the SoftBank Group, is it be instrumental in bringing synergetic effects? Or are they supposed to be separated and [ segmented ] in terms of the business strategy?

Yes. within, I guess, SoftBank's AI compute group, you have Ampere, Graphcore and now Arm. However, the chip that we've just announced is independent of anything being worked on by SoftBank's semiconductor companies, Graphcore and Ampere. This is entirely an Arm initiative, and there is no link through to what SoftBank is doing. Obviously, there is a lot of work being done. I think the Graphcore has had anyway about 500 engineers. I believe they're expanding their capacity. Ampere had about 1,000 engineers, and they may also be expanding their capacity as well. But those engineers are working on something completely different. As to when that will be announced, what they're working on, that's probably a question for my colleague over here. But for now, the only thing we can talk about is the CPU chip that Arm has just announced.

There is a second question on Page 37, CPU business. And in FY '29 and '30, there's going to be a huge jump in terms of the sales. Is that because of the inference demand is going to rise or capturing customer demand or mass production system will be kicked in. What is the reason behind this huge jump you are expecting in those years?

Combination of both of those. By the time we get to FY '31, we expect that our third chip will be in ramping into high volume. So the first sort of bumps are the first chip and then it starts to ramp, the second chip starts to ramp. And then by the time you get to the final bar in FY '31, you now have 3 chips from Arm all ramping. That is also combined then with, we believe, the sort of expansion of inference within the large data centers and with particularly the opportunity from Agentic AI. So it's a combination of both multiple chips from Arm and also an expansion of the opportunity with more agentic AI.

I need to let you know before. But this is Arm business briefing. So we appreciate it if you could just ask questions regards the ARM business. So the second person in the second row.

Thank you very much for your question. That is very comprehensive and really great opportunity for us to understand what A basic going forward. And my question is that setting your own ARM AGI CPU mean you are now competing with your big customers like AWS, Graviton or Google, Axion. In my understanding that beauty of the Arm business model is neutrality. So I just a little bit worried about that concern. So I'd just like to hear your thoughts on this.

Yes. Thank you. So AWS is building chips to be deployed only in AWS' data center. Microsoft is building chips to go into Azure and Google is building chips going into GCP. None of those companies are selling their chips to anybody else. It is only for their own internal use. maybe NVIDIA is selling chips in this area. But even NVIDIA seems to be wanting to sell systems, not individual chips. So if you're a Meta, if you're Cloudflare, if you're SAP, you have no choice. There is nobody to sell you a chip.

And so we've kind of found a market that is very underserved. And we think one of the reasons why it is so underserved is that these chips contain largely Arm IP. So if you are another -- if you're a chip developer, there's very little opportunity for you to add value or to differentiate. So it's not been an attractive market for someone to go after. But for Arm, because already it's all based on Arm technology, it's a very easy and attractive market for us to address.

And my second question is margin, which is a bit early to think about, but generally speaking, saying hardware usually has lower profit margin than licensing IP. So how will this change your long-term profit target or the mix of profit margins going forward?

Yes. Thank you. So you're absolutely correct. And as we indicated on one of the earlier slides, I think is it 2, no? Here we go. Yes. So the IP business in this time frame will probably have about a 99% gross profit margin. and a greater than 65% operating margin. And the chip business, we think will have in this time frame a better than 50% gross profit margin and a non-GAAP operating margin of 30%. So you're absolutely correct. The gross profit margin is lower.

But what we are expecting is that we will provide information so you can do a sum of the parts calculation on valuing the IP business as you do today and also then the new chip business. The other thing I'd point out is that today, we have an IP revenue stream. We have the cost of an IP business, and we've got the cost of a chip business already in the company today, and we're still delivering around a 40% operating margin.

So any additional revenues from the chip business, even though there will be a lower gross profit impact, but those gross profit dollars, most of those will drop down to operating profit. And so to a certain extent, we have the benefit of already having all the costs associated with the chip business in the business today. Going forward, when I talked about the sort of mid-teens growth in costs, other than inflation and a bit of an extra investment that we're going to be doing into the chip business, most of the remaining costs are actually into new technology areas that we're not talking about today. And so if those are successful, that will hopefully result in more revenue. And if they're unsuccessful, then the margins could be even greater.

So in the second row, the third from the left.

Masuno of Nomura Securities. First of all, CPU, a basic question. Other than Arm architecture, Graviton [indiscernible] compared to those competitors CPUs, AGI CPUs performance, how much is it? Twice as good as by X86. But compared to the other players, what is the performance level? The AGI chip is going to be applied to GPU of Google and Amazon and [indiscernible] Open AI accelerator Cbus accelerator. Can it be applied to all the accelerators? Would you please share with us some technical information on this aspect?

Yes. Thank you. What we have -- looking across the different chips that are available and how they fit with the accelerators from the cloud companies, most of them have been highly optimized to work in their systems. And so for example, Google's Axion chip has been targeted to work with very, very well with the TPU, NVIDIA's Vera chip is being optimized to run with Rubin, their GPU. And because they have been -- the 2 have been built to work very well together, although you could put an Arm AGI CPU in there instead of an AXION, it may not perform as well because the Axion has been optimized to work with the TPU.

But if you were to take an Axion and have it working maybe stand-alone, then we think we'd be very competitive. And so if you are Meta, you will have a competitive product to Axion or to Graviton or to Cobalt. So we think it's very competitive. We're not being paired with their own proprietary accelerator. But stand-alone or with somebody else's accelerator, we think it would be very, very similar in terms of the capability. But the key thing is that for most companies, they don't have access to these technologies. So just being able to provide them with the equivalent is very powerful for them.

So if that is the case without own CPUs and AI accelerators alone, so Meta and OpenAI and [indiscernible], these are those players. So this product is going to best fit those players. Is my understanding correct?

Yes, we work very closely with Meta on the design of the AGI CPU. They work as a lead partner, helping us to fine-tune the design to make sure that it fits very well for what they needed. But we didn't want to make something that was only applicable to one company. We wanted something that we could also sell to everybody. So we were trying to do 2 things at once, and we think we've done a pretty good job of producing something that will work well for them as a companion to their accelerator, but also to be broadly used across general purpose cloud workloads, AI -- running AI workloads on itself in the cloud as well as being a companion to an accelerator chip.

The second question is that going forward, license revenue -- guidance on the license revenue, you are talking about 3 elements. The factors for the drive AI cycle for one thing and next generation and then the -- and then the third one is licensing to SoftBank. And I think you were talking about all these things. They are going to continue into the future. Is my understanding correct?

Yes, your understanding is the same as mine. I mean we have not yet seen any slowdown in the AI technology cycle. So we are seeing many companies wanting access to the next generation of Arm technology because they need to run more advanced workloads in their next generation of chips. And we're seeing that in the data center. We're seeing that in people building chips for cars and for robotics, for smartphones, pretty much every end market you can think of companies are building more advanced capability into their next chip.

And some of that is going to be driven by Agentic AI. We saw how OpenClaw, which is a local agent that runs in a MacBook Mini or in a Raspberry Pi can be run locally and can effectively work as like a digital assistant, but that's not run in the cloud. That's run on a box next to you. And so we see that going into many, many different types of end product, consumer electronics product. And so that will continue to drive licensing.

Similarly, we have only signed up about 30 of our top 50 customers for our subscription licenses. We think that more of them will want to access. And so over time, that will increase the license revenues. CSS has only so far been licensed to a small number of our customers. There are many more we think will be wanting to take it. And so that will help drive licensing. And then the third point is we have visibility into SoftBank's road map and their plans.

And so we expect that the licensing and design services that we're doing with them will not only continue, but will continue to increase over the next few years as well. So those 3 drivers of license revenue, we expect to continue for multiple years. If you look on the chart, I can't quite see, we are more confident in the first few years worth of growth. The outer years are harder to see, but that is normal. Whenever we are forecasting license revenue growth, we can see the next 2 years, but then it starts to get harder to see. But by the time we get there, we'll probably have more visibility of more growth.

SoftBank Group after the project is complete. It doesn't grow after the project is complete. But do you still expect the royalty to continue?

Well, most semiconductor companies in general, don't build one chip. They build a series of chips. They have a product portfolio that will continue for many years. They plan for success. And so there will usually be a version like we had earlier, we showed AGI 1 and 2 and 3. So assuming if SoftBank Semiconductor Group has a road map, then they will need more technologies to come in the future. So we're anticipating that there will be multiple projects, not just one.

Then the second person from the first row.

My name is Yasui from UBS Securities. I have 2 questions. First question about AI architecture, about computation algorithm. I think there was a diagram on that. And you mentioned about important software. So for smartphone, iOS, Android, I think there are basically 2 softwares, which you basically monopolize. But what about AI software, the core is could be CUDA or PyTorch or [ Meta ] or the third party, another third pillar. So from your perspective, from Arm's perspective, who will be your software ecosystem partner? I would be curious to know that.

Well, I think you mentioned a couple there. So obviously, NVIDIA's CUDA runs -- the CUDA runs on the part on the Arm CPU and part on the GPU, Meta with PyTorch and ExecuTorch obviously, is an important partner. But also with Google just announced as part of their Axion 2 launch that they just ported 10,000 workloads -- so 30,000 workloads over to run on Arm. Microsoft have announced how they're porting Teams across to run on Arm as well. And all of the software that needs to run in the cloud has also now runs on Arm.

So I think Amazon recently announced of their top 1,000 customers, over 99% of them now run some of their software on Arm. So pretty much all the software that you can think of for the cloud now runs on Arm. There's very little that is now commercially available on our base chips.

I guess the image I had was you have this big NVIDIA ecosystem and then you have the cloud service ecosystem, maybe like orchestrators like or SS, maybe it's container-based or test based, it seems like it's running there. So from your perspective, who is the important software ecosystem partner from your perspective? So for example, you have NVIDIA CUDA, right? That's one. This is a cloud service provider OSS? Or do you think there are multiple, many more OSS kernel base? Would you be able to elaborate on that as well?

I mean I think all of it really. I mean all of the software that runs in the cloud is now already running on Arm. So I think there is nothing that's missing. So you mentioned containers. So yes, you can run Docker or Kubernetes on Arm. Python and Perl interpreters are optimized and running on Arm. So it's all of it.

And then my second question is about 1 gigawatt basis. So what's the assumption? And in 2030, it's $15 billion, that's your assumption. So would you be able to give us a rough picture until 2030. So if it's 1 gigawatt, I think it's 131 per chip is what, maybe 200 or 300 core, right? And if you actually translate that into CPU counts, it's about 200 core, it's about [indiscernible] it's about 0.4 million CPU. Maybe that would be necessary for 1 gigawatt data centers. So I want to know the assumption between 1 gigawatt and 2030, 15 billion because I guess the intention is that what's the CPU shipment when you say 2030, 15 billion.

Yes. I wondered if you were going -- if your question was leading to how many chips? We're not disclosing the number of chips, mainly because obviously, pricing -- per chip pricing is very much related to volume. And we see multiple ways of getting to 15 billion. It could be just through a small number of large customers or it could be through many smaller customers. So we've decided not to disclose chip pricing and chip volumes in the same way that most semiconductor companies also don't disclose chip volume and chip pricing. We'll keep that secret.

Then in 2030, the number of chiplets is 2, right? So could it be 4? So per CPU, how much performance improvement do you expect? Could you give us a ballpark figure on this?

So just taking a step back for a moment. So indeed, that is a -- our first chip is a 2-chiplet solution. One of the things that we are trying to create at the moment with other companies within the semiconductor industry is to create a chiplet marketplace. So at the moment, anybody building a chip based on chiplets, so including AMD and Intel and NVIDIA are doing so using proprietary interfaces that they have had to develop themselves.

What we're trying to create right now is in the same way that our 30 years ago, we created IP standards so that multiple companies could integrate IP from -- IP components from multiple companies into a single SoC. But with standard interfaces, standard busing, standard protocols inside the chip, we're now trying to help the industry to coalesce around standard interfaces between chiplets.

And so you could take a CPU chiplet from Arm, an accelerator chiplet from an NVIDIA or Rebellions or whoever, maybe an I/O chiplet from somebody else and then integrate those all into a single system-on-chip design or a single package, sorry, I should say. And that then could mean that you could then choose as to whether you have 2 CPU chiplets as we've done with our first chip or 4 or 6 or 8. As I mentioned earlier, we are already seeing some companies who are planning 500 core chips, which would be multiple chiplets.

And we've seen PowerPoint presentations, I should say, and plans, where the top right-hand chip is over 700 cores. And so significant expansion in the cores. As I mentioned earlier, our first chip has 136 cores, but we are expecting that over time, the core count in the products that we also produce will increase as the sort of the need for more CPU performance [indiscernible] inference data center increases. I can't tell you what that chip count will be in FY '31 or core count in FY '31. But certainly, we're expecting it to be higher.

We would like to take questions from the online participants, if there are any. [Operator Instructions] SMBC Nikko Securities, Harad-san. [Operator Instructions] Well, if it's difficult, in the interest of time, we would like to move on. Any other questions online? Then, on-site, we would like to come back to the on-site questions. CLSA.

Congratulations. So just to clarify, your EPS this year is tracking at less than $2, and you're saying it's going to be $9. So it's going up by about 5x. And that does not include anything from potential accelerator market expansion. And you mentioned $1 trillion. Could you just say what you thought about the TAM of that accelerated market? I think you mentioned $1 trillion, but you kind of skipped over it. So my question is, just to clarify, that's the magnitude of profit increase? And then could you tell us a bit more about this potential accelerator market? And any comments on what you would do with all that cash?

Okay. All right. So 3 questions there, I think. So yes, we are anticipating, forecasting more than $9 worth of EPS -- in non-GAAP EPS in FY '31, so in 5 years' time. About $6 of those comes from our existing IP business and 3 of those come from the new chip business. The TAM for the accelerated market is indeed, we anticipate growing to more than $1 trillion from about $250 billion this year, so quadrupling in size.

We do think that's quite an attractive number. And last week, when we announced our CPU chip, we did sort of have and watch this space for future, but nothing to announce just yet. And in terms of what should we do with the cash from that, we have nothing to update you on right now. So we generate at the moment about $1 billion a year in free cash flow. We aren't doing a buyback at the moment. Most of the requests from investors are for more shares, not fewer.

And so doing a big buyback doesn't seem to be a good idea. We remain open to considering a dividend, but that is something that has not -- no decision taken to do one just yet, but we remain open to investor requests there. And we do, do the occasional acquisition. Most of our acquisitions, however, are really accelerating hiring. We've been hiring between 1,000 and 1,500 engineers per year.

And as you hire them in sort of 1s and 2s, it takes 6 to 12 months to build an engineering team that can build new technologies for you. If you acquire a company as a whole, you end up with an actual team coming in that's already set up and running. It may be that you don't want them to develop their existing product, but they can -- you can give them a project to go and do and you've got the senior engineers, you've got the junior engineers, project managers, they often come with a building and with facilities managers.

So the whole thing comes in a nice package that you can then go and get ramping up very quickly. So that's something that we have done from time to time and we'll probably continue to do. But those typically be a few hundred million dollars. They're not going to make a huge dent in the billion.

Any other questions? Anyone? Third person from the corridor on the front row..

My name is Matsu. So on Page 37, you have a revenue projection. So the big 2 customers, OpenAI and Meta, how much would they comprise this revenue? And what is your visibility? How good of a visibility do you have, like, for example, firm commitment?

I'm not going to give specific customer information. Contracts with them obviously are confidential. But as I mentioned, we have visibility to more than $1 billion worth of revenues. And if you look on Slide 37, there's a little bit in FY '27 and a bigger amount in FY '28. The sum of those 2 parts is more than $1 billion. And we have very strong interest to deliver on that. So we are as confident as you can be in revenues that far out.

The second question. So when you're able to actually book a revenue, what would the CPU's P&L trend? So for example, secular, you're expecting $3 of EPS. But of course, you have upfront investment. So is there anything that's accretive that could actually push down the EPS?

Well, I guess there's 2 things that could push down EPS. Firstly, if the revenue doesn't come through in the way that we expected or if we increase investments. So staring into our crystal ball, I think that we have -- if Rene was here today, he would be telling you that he is very confident in the $15 billion and can see multiple ways that it could be -- we can achieve that with -- just with the customer set that we have today. And in reality, we'll probably be building a bigger group of customers.

And so therefore, it could be even greater. But it's 5 years out. So there is obviously lots of uncertainties, and so it could be lower. If we end up having another war, which seem to be brewing all the time or another credit crisis, could be -- lots of, obviously, uncertainties, but that's based on the customers and the customer demand we can see today.

Maybe it was lost in translation. My question was -- that's, of course, helpful. But my question was for the next couple of years, what should we expect from the chip CPU business? Is it going to be disruptive to EPS or accretive to EPS while you're running business at the early stage?

On the basis that almost all of the costs associated with that business are already in the business today, any gross profit is going to be -- is going to help grow EPS. So it is going to be accretive in the near term, assuming we are executing on that plan.

We are getting closer to the ending in time. We only take 2 more questions. If you still have a question, please raise your hand. In the third row from the front, the second from the corridor side.

My name is [indiscernible] of Goldman Sachs. One question for me. In terms of CPU, the power envelope, it has to be lower going forward in terms of the demand. And what is often talked about is GPUs, the consumption of electricity is relatively high. CPU consumption rate of electricity, what is the rate for CPUs? I think CPUs, the power consumption needs will be lowered in addition to GPU. Do you have any ideas about how to lower the power consumption of CPUs?

Yes. Thank you. On a chip-to-chip basis, a CPU will be lower than a GPU. As it is showed earlier, the maximum power used by one of our CPUs is 300 watts. That is a fraction of a GPU. However, as we also saw earlier in agentic data center, you may have many more CPUs than GPUs, whereas in a training data center, you have more GPUs and CPUs. So I think as we move towards more agentic workloads as we all start using AI more in a day-to-day life, then the data centers are going to have many, many more CPUs deployed with them.

And that's going to mean that the energy efficiency of the CPU is going to become increasingly important. So we'll be having to deliver more and more compute in as minimum power budget as possible. And I think as you saw earlier, we have the 36-kilowatt rack that's the sort of the maximum energy draw that that can draw. We don't even draw all 36 kilowatts. And so we are delivering much greater power efficiency than anything that is currently being installed in the data centers, which is helpful.

But clearly, this is going to become a multiyear focus, trying to get more and more compute out of the same amount of energy. And that happily has always been Arm's focus. Ever since we were started, we always were trying to create more compute out of fewer electrons, and that is what we'll continue to do going forward.

Do we have any other questions? The person in the front row. So can you just limit yourself to one question because of the time constraint.

My name is Satoru Kikuchi from SMBC Nikko Securities. I only have one time to ask one question. Related somewhat close to the question earlier, but so AGI CPU target of $15 billion, the breakdown of that? What is the specific contribution of OpenAI [ and Meta ]? I guess you are not able to give the specifics, but can you give me in terms of the amount -- like the breakdown of the hyperscalers? And on Page 40, you said that it's deployable to the new customer segment, which you were not able to capture before. So what is the specific customer segment that you're targeting? And is that included in this $15 billion in target?

So the $15 billion assumes that our existing customers in this area, so Amazon, Microsoft and Google do not buy chips from us. We assume that they continue to develop their own chips. And so the $15 billion is only made up of new customers, some of the logos that we showed before. And for the $15 billion, we see multiple ways of getting there. It does not necessarily need to be just Meta and OpenAI, it can be the other companies as well. We see opportunities much greater than 15. We're just setting $15 billion as a target as going from nothing to something is always hard to forecast, but we think it could be even higher than

$15 billion

Then with this, I think we would like to end today's Arm's business briefing. And this will be streamed on-demand basis on our IR website. So thank you very much for your participation.

Please welcome to the stage, Rich Hosfeldt, Co-CEO of SB Energy.

Good afternoon, Pike County. Wow, this is amazing. Welcome, Secretary Wright, Secretary Lutnick, our distinguished community and guests, our partners at DOE and Commerce. Thank you for joining us for the Ports Technology groundbreaking. Today, we start construction on a project that is years in the making and now a reality with the leadership and action of this administration. a 10-gigawatt AI technology campus, 10 gigawatts of new generation, $4.2 billion in grid investments, all fully funded to ensure affordable energy and America's AI leadership. 25,000 people contributed to building this historic ports project in the 1950s. And now 35,000 people will build the new ports technology campus, transforming it into a beacon for not only the U.S. but the entire world.

As we kick off today's groundbreaking, I want to recognize the 2,000 workers of the Ports team. Many of them are here today. You have been working for decades to prepare for this moment. Before the outcome was clear, before today was guaranteed. You believed it would come. You welcome SoftBank and SB Energy into your community, and we will be here with you for a generation to come. Today, together, -- we watch the future of the Port's technology campus. Thank you.

We welcome Tim Walsh, Assistant Secretary for the Office of Environmental Management at the U.S. Department of Energy.

Wow, is this amazing or what? Right here in Pike County. I love seeing all the yellow safety vests at heart. I'm a builder, I'm a construction guy, and I like seeing things built. So I am so excited to lead the office of Environmental Management at the Department of Energy.

And it's truly an honor to be here today on this historic occasion. When I drove through the Rolling Hills of Pike County, I arrived at the site and encountered these massive industrial footprint of buildings that once stood as a titan for the cold war era, I was truly aruck. The Portsmouth gaseous diffusion plant referred to as an AI plant and soon to become the AI plant. So you don't have to change the name much, is more than just a collection of steel and concrete and process piping. It is a monument to the high-stake era of American history.

In 1952, at the height of the Cold War, the Atomic Energy Commission needed to rapidly expand its uranium enrichment process. While the name says Portsmouth, the site is actually built right here in Pike County because of access to the Ohio River, reliable electrical power and a strong workforce that's represented here today. Construction of the -- yes, it's okay to C. Thank you. Back then, construction was an astronomical feat. As Rich just said, at the peak, it employed over 25,000 workers. 2 years later in 1954, the plant began primary mission, which was using the gaseous diffusion process to separate uranium isotopes. The scale of the project was simply staggering.

The 3 main process buildings cover nearly 100 acres or imagine 75 football fields of floor space under roofs. Peter Kiewit Sons was a contractor back then, and they delivered the building or buildings, the whole plant, 25% below budget and 6 months ahead of schedule. Try doing that today. To keep the compressors running, the plant consumed massive amounts of electricity. In fact, it was the largest power draw in the world when it was in operation.

At one point, it used roughly the same amount of power as the city of Los Angeles. During the operations, the A-Plant employed about 800 workers and was an economic engine for the region. For nearly half a century, the plant operated 24/7. And I met some of you in the audience earlier that actually worked here when the plant was still operating. Its history is split into 2 main chapters or mission. First was for National Defense.

Initially, the plant enriched on -- focused on enriching highly enriched uranium for nuclear weapons and the Navy nuclear propulsion program. And second, it was for energy. Over time, the mission shifted towards low enriched uranium to fuel the growing fleet of nuclear reactors lighting homes across the globe. As technology evolved and the cold war faded, the aging gaseous diffusion plant became too energy-intensive to compete with the modern centrifuge technology. So in 2001 in Richmond operations finally since and then the story of the A plant has shifted from production to decontamination and decommissioning.

Today, approximately 2,400 workers, many of you represented here today are still on site safely, carefully dismantling the massive structures and cleaning up the environmental legacy behind. The Portsmouth Gaseous diffusion plant was a silent warrior of the 20th century. It provided the teeth for national defense and the fuel for the atomic age. As we look towards the future of the site, we honor the thousands of workers that labored here in these Ohio Hills and helped shape the course of global history. That brings us to today, where we are gathered to break ground on a new project of equally historic scale that will provide generational jobs and economic vitality once again to the surrounding communities.

Thanks to President Trump and Secretary Wright's visionary leadership, challenging us to think big, just like the tent we're in today, we are embarking on the next era of American dominance. That dominance starts by significantly increasing electrical power production as a catalyst for reindustrialization.

To provide context, the 10 gigawatts of new power generation that was just mentioned is 4.5x the electrical output produced by the Hoover Dam. This positions Pion as the epicenter for technological advancements in artificial intelligence and data-driven scientific and engineering discoveries. AI data centers are the factories of the modern digital age. As a builder at heart, I see land as it can be and not what it is. I see a bright future right here in Southern Ohio, the new leader in innovation and economic vitality. This time, instead of the power of the atom, the power of gas generation will produce electrons that data centers on this very site will turn into extraordinary intelligence to advance human flourishing.

You, the community members and the workers here should be extremely proud of the vital role you play in America's national security and the role you'll continue to play in the next chapter of unleashing American ingenuity and prosperity. Finally, projects of this scale just don't happen by accident. It takes vision, leadership and a team to seize an opportunity to make it a reality.

I want to start by thanking my DOE team for supporting this revitalization effort. starting with Joel Bradburn, who is the site manager here at one point and his visionary leadership, who saw the potential reuse of the site and Ankur Bansal, a senior adviser partnering with me on the AI data center initiative. I also want to thank Rich Hossfeld, the co-CEO of SB Energy, for his leadership and vision in pulling this project together in short order. I don't think Rich sleeps because I could text him any time day or night, and he would respond immediately. But we're in a very important time in human history, and we're going to win the AI race, and we're going to fuel it with first gas generation and the future of nuclear renaissance will happen.

But most importantly, I want to thank Masa-san of SoftBank for dreaming big, committing significant amounts of capital to this project and leading this historic endeavor. I'm looking forward to seeing you during the construction progress this summer and being here with some really big scissors to cut the ribbon 2 years from now on the first AI data center. Thank you very much.

Please welcome Ohio State Senator, Shane Wilkin.

And reminded me, I may want to adjust the mic down so I can -- you guys can hear me. But good afternoon. Who's excited? So look at this room, what a great day for Ohio, but specifically, what a great day for Pike County. Today, we're not just turning over dirt, we're turning a page.

For generations, the people of Southern Ohio have shown resilience, grit and pride in their communities. In fact, all of us know the cold war was won right here at this site. You built lives, raise families and never stopped believing in the future of this region. And today, that belief is being rewarded. This groundbreaking marks the beginning of a historic investment by SoftBank and its energy division bringing one of the most significant power generation projects in the world right here to Pike County to the SoftBank team to the SoftBank team, thank you. You made the right choice.

But this project represents more than infrastructure. It represents opportunity, opportunity for thousands of workers who will help build this facility, opportunities for families who will benefit from stable, long-term careers, opportunity for schools, businesses, communities that will grow alongside this investment. And perhaps most importantly, it represents confidence, confidence in Ohio, confidence in Pion. -- but most importantly, confidence in each and every one of you that will build and work at this facility. It's the confidence that shows Southern Ohio is ready to lead in the next generation of American energy. Now in the essence of time, instead of calling out everybody that had things to do or had important roles in this, I just want to thank our local partners. I want to thank the state officials that I see seated.

I want to thank our Feds and our utility partners as well. But I would be remiss if I did not recognize the leadership of President Donald Trump, whose administration prioritized American energy independence and helped foster the kind of international investment that made this project possible. Now alongside him is a guy that may know a thing or 2 about Ohio and specifically Appalachia, Vice President, J.D. Vance. Vice President Vance has been and remains a strong advocate for communities like Pike County, places that deserve investment, that deserve growth and deserve a real shot at long-term prosperity.

Now while we look -- while we celebrate today with our ceremonial groundbreaking, let us not forget that we are also taking on a great responsibility. A project of this size and scale must be done right and time is of the essence. PJM expects demand to increase by more than 51,000 megawatts by 2040. So a 9,200-megawatt power plant represents nearly 20% of that expected demand increase and will support grid reliability going forward.

Now we all want lower energy cost, and this plant will play a role, a very large role. This plant will help offset the current future supply-demand imbalance that can serve to bring down prices for consumers because we all know when supply decreases or stays flat and the demand goes up, regardless of the market we're talking about, prices go up.

Now what does all this mean for this project? It means working together to ensure environmental stewardship, supporting our local workforce with training and opportunity, making sure the benefits of this investment are felt right here at home for generations to come. Because at the end of the day, success will not be measured in how big this facility becomes. It will be measured in how strong our communities become with this facility.

In closing, let's recognize what this moment truly represents, a new beginning with much work to come and a renewed sense of pride for our entire region. and the future that is being built right here in Pipe County. Southern Ohio is prepared and ready to lead the charge of American energy independence. This starts today right here and right now. Let's show the world once again what Southern Ohio can do.

Please welcome Congressman, Dave Taylor.

How awesome is this? How long we waited for this? This is incredible. I'm so glad to be here. Many of you probably don't know me because I'm new to this, but I am very proud to be the United States Congressman for Pike County, and I'm honored to be here with you all today. This is an incredibly exciting day for me, and I know for everybody else in the room.

Just as we did at scale years ago, we are getting back into the business of producing energy in a very big way. This location was pivotal in America win in the cold war, and it's poised to be pivotal in winning the technology race globally and drive the United States into the next 250 years of global leadership.

The investment being made here is vital from the most local level providing jobs to the neediest Ohio to the most international level of securing America's place on the world stage, which benefits all freedom people around the world. It's been said already by Mr. Wilkin, but it's important that we thank President Trump for his invaluable work. Sorry, Shane, would you do to me? For President Trump's invaluable work in getting this project done.

Special thanks to Secretary Wright and Secretary Lutnick for being here today and the incredible work they've done and also, of course, the amazing people at SoftBank and SoftBank Energy and last but certainly not least, the fine Americans, the fine Ohio and the great Pike County folks that are going to get this job done.

This project will transform the lives of hard-working Ohioans invigorate local economies and reaffirm Ohio's place as a leader in energy production. I look forward to working with the centers, Marino and Husted to ensure the success of this project, and I know they'd be here with us today, but they're in Washington, D.C. doing the hard work of reopening the agencies to keep us all safe in these very dangerous times. So respect their work.

They've been an amazing team, and we wish they were here, and I know they wish they were here as well. I thank you all for having me here how excited I am for all of you. I'll be out here, keeping close tabs on this and doing all we can on the federal side to make sure this is a complete success. God bless you all, Gob bless America.

Please welcome Satoshi Miura, President, Japan External Trade Organization, New York.

Ladies and gentlemen, distinguished guests, thank you very much for inviting me to today's very special ceremony. And it's a great honor for me to be a part of this is truly historic occasion. First of all, let me express my sincere congratulations to -- on the official launch of this project. and today's groundbreaking ceremony. This moment marks not only the start of construction, but also the beginning of a new chapter in U.S. Japan Economic Corporation.

This project is especially meaningful because it represents one of the very first tangible outcome of the U.S. Japan investment agreement. Turning bilateral agreements into real on-the-ground projects is never easy. So I would like to commend all those involved for making this happen. This -- the strategic importance of this project is clear. We are living in an era where artificial intelligence and data centers are becoming an essential infrastructure of economic growth, innovation and national competitiveness.

Reliable and scalable energy supply is the foundation of this digital transformation. Without stable energy, AI cannot survive. My organization, Japan trade organization, our job is -- we are the official agency of the Japanese government and our job is to trade promotion and the investment promotion. So we are proud that Japan is the #1 investor among all the investing countries across the world. And also, we are the biggest job creator in manufacturing sector in the U.S. And here in Ohio as well, there are roughly 270 companies there.

And so in terms of the number of companies, Japan is again the #1 investor in Ohio. And -- and the story does not end here. And looking ahead, Japanese companies are showing continue to show strong interest in doing business in the United States and Jetro is committed to help those Japanese companies to come here and invest more in the United States. And this project itself is a powerful example of how U.S., Japan economic cooperation can create new business opportunities for Japanese companies and including many of those gathered here today.

So seeing this cooperation translates into real investment and real growth is deeply encouraging for us. So finally, I would like to express a sincere appreciation to the organizers, our partners and local stakeholders who have made today possible. and I wish this project every success and a bright future. Thank you very much.

Please welcome Mark Ryder, President and Chief Operating Officer of American Electric Power, Ohio.

Normally, thunder is not a good thing for my job. Well, good afternoon.

I'm going to start by thanking Secretaries Howard Lutnick and Chris Wright, SoftBank Chairman, Masayoshi-san, SB Energy Co-CEO, Rich Hosville; Congressman, Dave Taylor; Ohio Senator, Shane Wilkin and all the local Pike County community leaders and officials. Thank you. Obviously, this is an exciting day. This site has a storied history, providing critical support for both national defense and energy security.

And today, we are adding a new chapter to that story. SB Energy's investment here in Pieton is historic. And together, we usher in a new model for how to align new infrastructure costs and investments for technology campuses. This innovative partnership that includes SB Energy, the U.S. Department of Energy and AEP Ohio will not only build $4.2 billion of new electric transmission infrastructure to support this project, but the partnership agreement will also protect residents and businesses in Ohio from shouldering the cost. SB Energy has committed to paying for all of the new electric transmission infrastructure required.

And that speaks to their commitment to our state in this region. From 1920 until the early 2000s, energy executives thought in terms of megawatts. From 2010 through 2020, many changed their thinking to kilowatts as electric use was flat or in many places, declining. Now we think in terms of gigawatts. We're navigating a generational super cycle, trying to thread the needle between executing on our obligation to bring electricity to all customers who need it and ensuring the cost of that electricity are divided fairly. This new data center campus is a significant electric infrastructure investment.

When we are done, it will bring 10 gigawatts or 10,000 megawatts of power to this historic site. And for context, APO Ohio serves 1.5 million customers across 61 of Ohio's 88 counties. When we are done with this project, we will more than double our demand profile. Electric transmission infrastructure is the backbone of economic development. It powers manufacturing, research and development, small business and residential growth. This project is especially meaningful for me as a native Ohio and and a current Board member of the foundation for Appalachia, Ohio.

I'm also proud of the engineering prowess American Electric Power brings to the table. AEP pioneered 765 kV transmission and owns and operates the most 765 kV transmission in the country. Representatives from AP Transmission are here with us today, and they have started planning and developing designs for the new lines needed to support this project.

Building these lines will take hundreds of skilled workers, and it will provide opportunity for additional economic growth by creating a new energy backbone in the region. We look forward to our continued partnership with the administration, SoftBank SB Energy and our communities across Ohio as we work together to enable growth and strengthen America's energy future. Thank you.

Please welcome Masayoshi Son, Chairman and CEO of SoftBank Group Corporation.

I need to lower the microphone. I want you to remember today for all of you as a historical day. Why historical day? I want you to understand what we are building in this technology campus is the largest, most powerful intelligence in the world. Here in Portsmouth, Ohio.

I want you to understand. We are talking about 10 gigawatt power, 10 gigawatt power to do what to do -- to create the most smartest intelligence of the world, as I said. Let me tell you about the scale. This 10 gigawatt of intelligence is bigger than total artificial intelligence in the world today in one location, bigger than Google, Amazon, Microsoft, OpenAI, Anthropic, Oracle, everybody added together of all of the world, all of the company, every company, every model of ChatGPT, Gemini, Anthropic, all of this added together, bigger than all of them added together, that's the size we are building here in one campus.

Let me tell you the scale, 10 gigawatts represent the 10 nuclear power plant scale of power generation in one campus. So that's the size we are talking about. This is a size bigger than any power plants, I think, in the world. At least in the U.S., for sure, this is the biggest power generation in one location, probably bigger than any place in the world. So the size that we are talking about in terms of the dollar investment -- in one campus, we are making $500 billion investment in one shop. That is a number of investment I have made a commitment to President Trump, the first day he stepped in as the President of the United States last year as a second time.

The first day of his operation in the office, we made announcement as a state $500 billion, but that was intending to do many, many campus in many projects -- but here, I say, I changed the word I said. Instead of many locations, many years, total $500 billion as a vision, now we are delivering $500 billion in one campus.

Let me tell you one more thing. We are investing chip and system on top of power plants plus the data hall and so on, all of that together is $500 billion. But every 5 years, we are replacing, upgrading the chip and system. That is additional more than $300 billion. So over 20 years, we are adding additional $1 trillion in this campus. So this is one project that has total of $1.5 trillion. And that will stimulate the whole industry of chip and system, cabling, power, all the related equipment on top of the industry to the medical industry, the educational industry, physics, science, all of the things that we live is stimulated and accelerated innovation with a super intelligence.

So this is the center of our super intelligence. So it's a great thing. It's a great excitement historical day. But people -- some people start to criticize recently, oh, AI data center will take too much power, and that will increase the electricity bill for the local community, that is not good. So people start to say, "Hey, is it a good thing? I would say, well, that negative aspect of electricity bill going up, I will commit right now right here. We will protect the electricity bill.

We would not take the electricity away from the grid. We would generate entire electricity that we use in our compute by ourselves. We here create a new generation of entire 10 gigawatt, all of the consumption that we use for our intelligence, we generate self generate. So therefore, we are not taking any electricity away from the grid. Therefore, the price is supposed not to go up, and we are investing $4.2 billion to support upgrade the network of electricity grid so that it will actually help probably reduce the cost of overall for the community and increase the efficiency.

And on top of that, we would like to donate $40 million to the local community for the children, for the school, for the medical, for the world. So I lost my father a couple of years ago because of the cancer unexpected. Our doctors and the hospital could not find early enough. I ced,ed,ide. If we had a super intelligence, if we had a better technology, my father could have lived a few more years. So why technology is needed? It's not just productivity or money or reputation. It's for helping people. I truly believe that technology is to make people smile again. Thank you very much.

Please welcome Howard Lutnick, United States Secretary of Commerce.

Well, it is amazing to be here. I want to start by making sure I thank representative Dave Taylor, your State Senator, Shane Wilkin, Satoshi Muira, who is the President of the Japan Trade Association. Thank you guys. I really appreciate it. And you're going to meet my partner, Secretary, Chris Wright, who he and I have worked together to create this historic project.

So today is the groundbreaking of the first project that's come from the Japan Invest America Fund, which was generated by the $550 billion trade deal that I negotiated with President Trump between the United States and Japan, the United States and Japan. See, you've never heard of a project like this before, right, a $33 billion project. right? The largest industrial project, obviously, in the history of Ohio, the largest -- one of the largest, if not the largest industrial project in the United States of America because it's never been done before because this is the trade policy of Donald Trump.

We negotiated with the Japanese, they wanted a lower tariff. And in exchange for a lower tariff, they committed $550 billion for America to invest in its national and economic security. And the first project we've selected is this gigantic gas-fired power plant right here in Portsmouth, Ohio. So we're operating in Trump time, okay, which means it was just a month ago that this deal was announced a month ago. And now we're here with shovels going in the ground to get this baby going, all right? These are projects that need to deliver for America. So let's take a minute and understand the thinking of Donald Trump.

See NAFTA, you remember the North America free trade agreement. That agreement signed in the early '90s, basically gutted Ohio and Michigan and the great industrial parts of our country because what it did is it opened the borders of Canada and Mexico. So our industrial companies would move the production there, break our unions, get cheaper labor and basically really good for the corporate globalists, but really bad for wonderful American citizens, right? And the key to Donald Trump's trade policy.

President Trump's trade policy is crystal clear. It's America first, America's workers first. Now when you talk about industrial policy, it sounds kind of boring, but let me explain what industrial policy is. Industrial policy is going to produce 4,700 construction jobs right here right now. So I'd like to -- I've heard Masa-san said this is like this gas-fired power plant is so big. It's the equivalent of 10 nuclear reactors. I'd like to call it 2,700 wind turbines that only work when it's windy. So you know the President, he delivered the largest tax cut in American history, the largest investment in rural health in history.

And now we're going to do the largest construction project in history. So we are focused on unleashing American energy. So we, the Japanese government in partnership with the United States Federal government is financing this project. This is not a project with debt. This is not a project that won't finish. This is a unique model in the world, and this is the first of its kind, right? This will produce reliable, affordable energy in the United States of America. It will drive down the price of local electricity. That is the model of where we come from.

We are going to assure you the price of energy is not just going to be stable, the heck with that. We are going to drive it down because you are going to have one of the greatest power generations right here in your backyard, you're going to build it, you're going to operate it, and then it's going to deliver for you lower energy prices, right, for your lifetime. And that is what is amazing about this project.

So President Trump and Vice President Vance, they didn't forget the people of Ohio. They're not going to forget real Americans. This is the example of Donald Trump's trade policy. This is the revitalization, the reshoring, the building back of American Industry here, and it starts right here in Portsmouth, Ohio. I am honored to be here. I am honored to have played the role that I did in creating this fund.

And now I'm going to turn it over to my great friend, Secretary, Chris Wright, who's my partner through all of this. Thank you all. And God bless Portsmouth, Ohio and God bless America.

Please welcome Chris Wright, United States Secretary of Energy.

Thank you, Pike County. Thank you for bringing America back. Thank you for making America great again. President Trump's agenda is in this room. It's in your hearts, it's in your souls, it's in your community.

To believe in America that's greater for our children, for our grandchildren that has better job opportunities, more affordability, more optimism, more security, more belief in our future, America has to stand up again. We have to believe in ourselves. We have to invest largely and build big things again. As you heard already, this project, this facility that's going to be built here is bigger than any facility ever built in world history. This is going to be a gigantic factory and ecosystem that will build intelligence, take electricity and turn it into innovation, turn it into opportunity, turn it into medical advancements so that cancers will become manageable conditions as opposed to death sentences.

You heard Masa's personal story about that. That intelligence is going to be critical for our national defense. It's going to be critical for our quality of lives. But this is Think of the pivot President Trump made.

The previous administration thought success was to tear down big things, take coal plants down, take natural gas plants down, shrink our industry because we were afraid of the future, tell our kids that they should be fearful of their future, that they should learn to live with higher costs and that they should understand opportunities may be shrinking, not growing. That's not America. That's not what's in our hearts. That's not the history of our nation.

President Trump knew we have to lean in. We have to believe in America. We have to make big bets. We have to grow. We have to -- we have to change if we're going to continue to lead. I'm so proud of his leadership. I'm so proud of Secretary Lutnick before me, and I'm incredibly grateful for our tremendous partners with both the government of Japan and SoftBank, just a giant technology pioneer that believes in America, that believes in Ohio and that believes in Pike County. There's a reason we're here. There's a reason all of us didn't travel out to San Francisco, California.

There's a reason -- this is a community with strong hard-working people that want to build America back, that believe in this country, that want to leave this country better, stronger, more -- with more pride and more optimism in the future than when you arrived. You can't do that everywhere. These giant projects, there's communities that are going to work and there's communities that aren't considered.

There are still states that don't want to build new things. They don't mind imposing higher costs. They're not optimistic about the future. That's not Ohio. That's not Pike County. I want to -- let me conclude with a summary of the points you've already heard, which is that this massive investment is going to create a lot of jobs, and there's going to be other industries that will come in here to be near this energy production, to be near this intelligence generation that will want to locate here.

And when a lot of businesses come in, it creates a lot of jobs. And what does that do? That raises wages. That raises wages. But at the same time, we're going to produce so much energy that we're going to raise wages and lower costs. Is anyone against that agenda? Funny enough, the previous administration was against that agenda, but that was nonsense. President Trump came back to bring back common sense, to believe in America again, to believe in you. And all of us that are here today that traveled from out of town are here because we believe in you.

The largest bet on a construction facility ever made in human history is right here, and you should be incredibly proud of that. That's a vote of confidence in you. That's a vote of confidence that we'll build a better future together and that we love our children, our grandchildren, we love our country, and we believe our futures. God blessed Pike County, God bless, Ohio, God blessed President Trump's leadership. and God bless the United States of America.

Thank you so much for joining us at the Port Technology Campus celebration. Enjoy your weekend, and please drive safely.

Welcome to the SoftBank Group Corp. Earnings Results Announcement for the 9-month period ended December 31, 2025.

Now we would like to start and also would like to introduce today's participants. From left, we have Yoshimitsu Goto, Board Director and CFO; Kazuko Kimiwada, Corporate Officer, Senior Vice President and CAO; Navneet Govil, Executive Partner and CFO, SB Investor Advisers and SB Global Advisors. Jason Child, Executive Vice President and CFO.

Today's announcement is live broadcast over the Internet. Now I would like to invite Yoshimitsu Goto to present you the earnings results and business overview.

Good afternoon, everyone. Thank you very much for your time today. This is Goto speaking. This January, I had a knee surgery. And currently, I have a bit difficulty to keep standing. But for the next earning, I believe that I can recover well, so that I'll be able to explain to you and present while I'm standing.

I had a bit bad knees and made me difficult to walk straight and try to fix that when I was younger. So that's why that I come to this. So I appreciate my parents who born me in a healthy manner, but this was my first ever surgery and being admitted in hospital in my life.

And when I am on the wheelchair, I realize many things which I never experienced, which I never imagined over 6 years in my life. So it is quite lessons that I've been learning. And I've been admitted in the hospital for 2 weeks, and alcohol-free for 2 weeks, that was also another first time in life.

And I lost 8 kilograms in weight, which is a good news also. But that's it. That's only for my things, but I would like to come into the earnings results announcement, and I will present while I remain seated, very much sorry about that, and I hope you understand.

So this time, agenda for the earnings to be the #1 ASI platform provider. So what is platform, I believe, is always a question for you and for us, too. At the General Shareholders' Annual Meetings last year, Masa presented that we would like to be ASI platform provider. And Masa was also asked and he answered, we could say a core pillar of the industry that platform is.

This can be also a base for the many things, and we can build those bases so that many things which we do not have can be on top of that so that we'll be able to grow together. That's something that we have as a mission for SoftBank Group.

When the company founded the business, it started with a wholesale of software. And recently, I spoke to Son -- I was talking to Masa, he said that he never thought that his original business is a wholesaler business.

Back then, software was very new and just wanted to create a system to distribute such software to people, and he started such system, which was wholesale. And then people will be able to recognize what is the software, and they will be able to learn more so that Masa started publishing business, also started the broad media and many things were put it on to the platform since then.

So that's a similar thing that we are doing. After year 2000, Internet era doomed platform or platform provider became more important, even more important. And now we are entering into AI and headed towards ASI platform provider is the one that we would like to be.

And we are steadily laying the groundwork right now. I don't play a game of goal, but we are laying the ground. In the past 3 months, this quarter, we made quite a large amount of investments in OpenAI and completed $22.5 billion follow-on investment back in December. As a result, $34.6 billion, which is about JPY 5.4 trillion invested in total for about 11% ownership.

And to advance target, we have also a strategic partnership formed. Other than OpenAI, we do have several progress. In robotics, ABB, which is well known around the world, we announced the acquisition with $5.4 billion.

For chip, we completed the acquisition of $6.5 billion completed. These are the events that happened in the third quarter.

And DigitalBridge, which is the digital infrastructure investment company, and we announced this company's acquisition with $3.1 billion in December. So each respective events are our groundwork. And I believe this make us progress in the big picture for the framework. And when you look at our portfolio, this has also been changing, and this is something that we should also highlight.

About 3 years ago, if you see by business segment, portfolio for ASI was about limited to pretty much Arm and some robots portfolio companies. So about 18% out of the total net asset value was our ASI investments.

But now Arm is growing even further. On top of that, we have decided investments in others, including OpenAI robots and so on. So now that more than 60% assets are ASI-oriented investments amongst our net asset value.

And the highlights for third quarter 2025. First was the completed follow-on investment in OpenAI, increasing ownership to about 11%. And net income first quarter to third quarter, JPY 3.17 trillion. which was increased by JPY 2.54 trillion year-over-year, so which was a very good number.

Third, net asset value. So gross less net interest-bearing debt was JPY 30.9 trillion at the end of December. So if you only look at this compared to September end, which was even 3 quarter before, this is minus JPY 2.4 trillion. But latest figure-wise, it is recovering. And as of today, JPY 33 trillion as of today.

So it's about the same level as December -- September end, excuse me. And large investments made with financial policy maintained. So OpenAI follow-on investment were also made and that actually giving us about JPY 6 trillion level of the investment so far.

But at the same time, while we make a decision on investment, we also very much focus on the financial policy of loan-to-value. And so with those investments, loan-to-value at 20.6% cash position, JPY 3.8 trillion. Loan-to-value, as of today, this morning, less than 20%, it's about 19-ish percent.

Now let me walk you through the consolidated results. First, you can see net sales gain on investment, income before income tax, all increased. So FY '25 is a good year in terms of performance. By the quarter, in terms of gross loss on investment and net income, in the third quarter, about JPY 300 billion of gain (sic) [ loss ] on investment.

You may think it goes up and down. But if you look at the last 4 quarter, we posted gain even in the range of 7 quarters, we posted a loss only the 1 quarter, but remaining 6 quarters, we posted a gain. So all in all, we are seeing growing gain. So our investment strategy is bearing fruit and represented in this way.

So if you go deeper into gain or loss on investment, in the third quarter and the second quarter, both huge increase of OpenAI's fair value contributed a lot. That means our decision on investment was successful.

Looking at the second quarter, for example, fair value of OpenAI is represented in the lighter blue, which is big, but also other investment posted about JPY 1.3 trillion of gain. In the third quarter, however, we lost other investment by JPY 346 billion or so, but it included a temporary loss of listed companies.

But all in all, we recorded a positive number. And like I said, huge contribution to our bottom line is OpenAI. If you look at the valuation for the last 12 months or so, when we started investment, valuation was about $150 billion.

In March 2025 or this fiscal year, we decided to invest as much as $30 billion. Back then, the valuation was $260 billion. On that valuation, we made investment at $30 billion. And as of October or just 3 months ago, the valuation increased to $500 billion. It's not only we are seeing this is OpenAI's valuation, but from the third evaluator's perspective, also OpenAI's valuation is around JPY 500 billion. And OpenAI's leadership in AI sector and offering -- a lot of service offerings are showing this kind of high valuation.

And talking about investment gain on OpenAI, the green bar is investment cost. At the end of the second quarter of 2025, cumulative investment gain was $19.8 billion. And in the third quarter, we added $23 billion from the prospect of investment.

So as of December end, the black portion is investment cost, which is about $34.6 billion. And on top of that, the gain in FY 2024 and recorded in second quarter and the third in 2025. Total, we are looking at investment gain of $54 billion. So since inception or since the start of investment, the investment gain for FY 2025 is $4.2 billion.

Now key indicators. Net asset value or NAV, which is the most important indicator as an investment company as of December 31, NAV was about $31 billion. But recently, it went up to JPY 33 trillion. Loan-to-value as of December end, 20%, but now or recently, 19%-ish. And cash position back then, JPY 3.8 trillion, and now we have a little bit more than JPY 3.8 trillion recently.

We keep this level of indicators while maintaining discipline in making investment. And if you look back the history of NAV, as you can see, our asset value is and has been stable and loan-to-value is shown on this slide, recently, less than 20%.

So as you can see, for the last 4, 5 years, we have been maintaining the safe LTV level, which means even though we have made huge investment in the last 9 months or so, yet, we maintained financial policy in a very disciplined manner. That's the core message I want to deliver to you.

Cash position. Our policy is to make sure that we have cash position well over the worth bond redemption for 2 years. Again, as you can see, we have been maintaining about JPY 3 trillion to JPY 4 trillion level.

So Vision Fund. So when you look at the Vision Fund gain or loss on investments, roughly speaking, SoftBank overall gain and loss investments movement is about in line with these movements. And when you look at the breakdown of the gain or loss, Vision Fund 1 for this quarter, we had a negative numbers here, mainly from the public portfolio companies.

So it's about $3,800 million due to -- primarily due to Coupang. As you may also aware that due to the data leakage, the share price has gone down. But at the same time, for this company, Coupang, we are convinced with the business fundamentals also doing pretty well in the operations. So very much expect the further recovery of the business and the share price.

On your right-hand side, Vision Fund 2 and LatAm and others, we have a large positive number here. A big contribution from OpenAI and other new investments in a couple of years before that their mark has already been growing as well. So now you see the Vision Fund 2 valuation has been growing quarter-by-quarter.

On total, we maintained our positive numbers. So I said that 7 wins over 1 lose, I mentioned earlier. And actually, the cumulative gain has been increasing. So we've been having a little bit of a difficult time.

But at the same time, now that we've been able to show the good recovery in total. And after this, we are looking at the very exciting late-stage portfolio companies, and that is something that we believe is going to be another good contributor for gain on investments.

And these are those exciting late-stage portfolio companies that I mentioned earlier. So on your left-hand side, we do have a public listing track record. So in total, 60 listings since inception. And in this third quarter, Lenskart, Meesho, those 2 made newly listed.

On your right-hand side, are the exciting late-stage portfolio or companies and so December end last year, our total fair value was at $90 billion. And you can see the logos below. There are companies, which already made filings for IPO, which includes PayPay in Japan. Also Clerk has already made a filing for IPO. On your right-hand side, there are other late-stage companies, which we can expect the good exit or the good IPOs. So that includes OpenAI, ByteDance, Fanatics, Revolut, Oyo, Yanolja.

So those are the exciting other late-stage portfolio companies. So those are -- we are closely looking at those and monitoring those companies so that we'll be able to improve our investment performance of Vision Fund. As I mentioned earlier that there are 2 listing in the third quarter, which include Lenskart. So eyewear, glasses. So from the design to manufacturing and sales is done by this company, and the performance is doing pretty well.

MOIC, 4.8x in growth and also share price since its IPO is making a good increase. Another one was Meesho. This is an e-commerce marketplace in India to democratize commerce by connecting consumers with small and micro sellers.

There are some ups down in share price, but in total, they are showing a good increase in share price since the IPO and the investment return, 2 of them are coincidentally are both Indian company, and they are doing a pretty good progress.

And I'd also like to touch on PayPay a little bit. They are making quite a smooth and great progress so far exceeding 72 million users already. And I would say PayPay is used by more than 1 in 2 people in Japan and 2 out of 3 smartphone users in Japan.

So as a payment platform that PayPay is playing in Japan, KPIs are also showing pretty good. GMV on your left-hand side, in 9 months period, JPY 14 trillion. So if you annualize, I believe that the number can be expected close to JPY 20 trillion and also 24% year-on-year increase and EBITDA, 83% increase year-on-year. So financials are also showing a great progress so far. So sometime in the future that we believe we'll be able to show you the good valuations when they go public.

Now Arm. Arm is showing steady growth. Revenue, they recorded the historic high, up 26% year-on-year. License and other revenue and royalty revenue, both has grown steadily. Non-GAAP operating income on a quarterly basis is shown here. If you compare to the last quarter and -- sorry, the last -- in the same quarter last year and the year before, they have been posting increase.

Arm requires R&D investment, that's for sure. While they are investing in R&D still deliver this kind of operating income, which is great. Arm's technology is penetrating in various areas. For example, in those 3 sectors, Edge AI, physical AI and cloud AI, starting from the left, for smartphones and IoT devices, this is edge AI area and Samsung announced Exynos 2600, and this is the chip for mobile, Armv9 and CSS-based one.

In the area of physical AI, Rivian in the U.S. Their chip for autonomous driving is ArmV9-based. In the cloud AI space, AWS announced they would use Graviton5, which is also ArmV9-based chip. So again, as you can see, ArmV9 chips is penetrating in different areas and ArmV9's royalty unit price is twice as much as V8.

And CSS-based chips royalty is even higher. So as they go into deeper in different sectors, that would further push Arm's royalty revenue. And Arm announced guidance of solid growth outlook. And those numbers, I believe, are well over analyst consensus. Revenue, $1.4 billion or 18% year-on-year. That's the outlook for the fourth quarter. Of course, the range is upper than $50 million, but $1.4 billion is a median number. And non-GAAP operating expense, $745 million, non-GAAP fully diluted earnings per share, $0.58. And I believe that Arm will definitely deliver those numbers as forecasted.

Now I would like to put a little bit more color on OpenAI. So here, our investment in OpenAI historically. cumulatively, we have invested $34.6 billion so far. Latest, at the third quarter, we made $22.5 billion follow-on investment completed in December last year. As a result, ownership in OpenAI Group PBC became 11%. There are other major investors like OpenAI Foundation, which is NP -- nonprofit organization, Microsoft and others.

And their service is steadily very well performing. So it's about 3 years since the product launch, but weekly active user is now exceeding 800 million and keep growing.

Annual recurring revenue in December 2025, it exceeded $20 billion. Actually, this is 10x in 2 years. So I believe this is quite obvious that they are making a great progress here. And app downloads global and also cumulative since the service launch, about 1.52 billion downloads are made so far, far bigger than their competitors.

And the total number of ChatGPT for work has already exceeded 7 million accounts. And they are expanding enterprise business as well. They are also releasing new products one after another. For example, ChatGPT for healthcare or Prism.

This is the specialized for scientific writing and collaboration. Codex, this is the agent coding model, optimizing -- optimized for real-world software development. On your far right, Frontier, this is an enterprise AI platform that enables organization to build, deploy, operate and improve AI agents. And we are working on crystal intelligence together with OpenAI.

And this position as a foundational platform to deploy to enterprises in Japan. So with these launches, we believe that they will be able to accelerate their business growth even further.

They also started advertising in ChatGPT. So small letters here, so you may need to pay attention. On your left-hand side, this is smartphone and asking about the dinner, Mexican dish, for party and asking that question to ChatGPT. ChatGPT answers you with several examples of good party dishes for Mexican.

So one is the steak dish and the other one is chicken. Then in the same screen in the very bottom of the ChatGPT answer, you will see the advertisement to offer you the good source for which matches to these dishes. So if you try to cook a bit -- if you prefer a bit hotter dish, it may be good use of such an advertisement.

So this is just one example, but they will be able to show several advertisements. So whenever you ask something to ChatGPT, ChatGPT will give you an answer, but at the same time, give an offer, which matches to the answers.

So in e-commerce area, I believe this can be good use, and very attractive for many of the users. So we believe advertisement model itself can be a good also growth for the revenue.

So this week, they are starting this testing advertising in a limited and experimental manner for free and low-price subscription plan in the United States.

Also I'd like to touch on the Stargate. OpenAI, SoftBank Group and SB Energy, which is our group companies, signed a strategic partnership to advance Stargate back in January 9 this year.

So OpenAI, SoftBank, each providing investment into SB Energy, so that this can accelerate the progress of SB Energy business. Data center in Milam County, Texas is something that we've been working on, 1.2 gigawatt data center.

And SB Energy actually selected to build and operate. And also 15-year data center lease agreement signed with OpenAI. As you can see, the photos, which is the actual local site. So they are working on the civil works right now. And once the lands are ready, there will be building and start operations.

Next, DigitalBridge. On December 29, while you are probably on New Year holiday period, we announced acquisition. And that was agreement to acquire all outstanding common shares for $3.1 billion and closing is expected in the second half of 2026. DigitalBridge has been making investment in digital infrastructure, specifically. The AUM is $108 billion or JPY 17 trillion and the portfolio companies are around 45 and well over 100 digital infrastructure investment professionals. And it will be welcome to get this kind of company in our group.

Key portfolio companies include those on shown screen. Vantage, for example, deploying data centers across countries and Switch also is a very well-known EPC company to deploy data centers. And towers, JTOWER, that's a Japanese company. And I believe almost 100% is owned by DigitalBridge.

Last but not least, let me remind you our financial strategy. First, financial policy remains unchanged. We maintain loan-to-value below 25%, and we maintain at least 2 years' worth of bond redemption in cash.

As I said earlier, we are and will be making a lot of investments actively, but this investment should be supported by strong financial policy.

Looking at investment amounts, in the last 9 months, we have made a huge investment -- in the first quarter, $8.3 billion; second quarter, $3.2 billion. In the third quarter, $32.3 billion, over $40 billion of investment we have made in the last 9 months alone, which is around JPY 6 trillion.

Still, however, we keep less than 20% of loan-to-value. And also, we have JPY 3.8 trillion of cash position. Our financial activities are done always with financial policy in mind.

And the question is, how do we finance? How do we raise capital? We are an investment company, and we relatively use our own balance sheet. And also, we are a listed company. I think in that sense, we are in a very unique position globally.

What we do is first utilize assets held, obviously. We have about JPY 40 trillion of equity value holdings. With that, we raised capital, for example, we do margin loan, we use color transaction, and it's important to recapitalize portfolio.

In that sense, sometimes we sell some assets. And in fact, listed shares, 51% of total holdings are sold. And also even unlisted shares, we use a source of capital raise. At the same time, by leveraging appropriately, we can communicate well with equity investors as well as credit investors and domestic and foreign bonds and loans, we use those markets and also retail bond, corporate bond for institutional investors and also, we loan from Japanese and overseas banks.

And hybrid bonds and loans are also utilized for our capital raise because we can maintain 50% as equity capital. Cash position through those financial activities, we have, at this moment, JPY 3.8 trillion. Utilizing debt market, what does it mean for investment company like us?

Investment company in terms of general financial activities, we recoup investment and make new investment. That cycle is being run. But recoupment of investment and making investment, there is a time lag to support that gap, leverage should be used to make sure that we won't miss investment timing, and we won't miss recoupment of the investment. So the leverage means a lot in terms of filling the gap between recouping previous investment and making new investment.

So again, while leveraging, we can maintain a safe investment and even expand investment. In terms of utilizing assets held more in detail, what we have done is on the left-hand side, you can see we sold T-Mobile shares, Deutsche Telekom shares and NVIDIA shares.

And that equipment is used for new investment. So we sold $21.3 billion in total. And also, we did asset-backed finance of $15.9 billion in total using Arm shares, using SoftBank shares. We did asset-backed finance. And also, we utilized SoftBank Vision Fund value to be able to raise $1.9 billion. We are discussing to figure out the best possible option we should take.

In terms of utilizing that in the last 9 months, that, of course, has to be returned. Just to remind you, otherwise, you will end up being borrowing forever. You have to return the debt.

Sometimes we utilize a bridge loan last year, for example, for the first investment in OpenAI or investment in Ampere, we structured a huge bridge loan. We repaid a little bit. And we return the loan and while keep investing. And we issued senior bonds amount about JPY 1.7 trillion.

Of that, JPY 800 billion will be redeemed. In 1 year, about JPY 800 billion to JPY 1 trillion is redeemed. And at the time of redemption, for investors, we would say thank you very much for supporting us for the last 5 years. And we have plenty of cash to get back to them. And investors know we have enough cash position to be redeemed -- to be used for redemption. And at the time of redemption, we issue -- we prepare issuing new bonds and offer new bonds to investors.

At that time, not only investors who are going to see the redemption timing pretty soon, but also offer an opportunity to new investors. Hybrid bonds compared to senior or corporate bonds, the volume was small, but still, we issued JPY 0.6 trillion of hybrid bonds.

So debt raising capital and redemption and repayment. So in terms of debt market, we communicate well with investors to be able to keep investing and returning back to shareholders. By utilizing that well, our share price should go up, then that would satisfy or that would make equity shareholders -- equity investors very happy.

To summarize, this is the executive summary of Q3 of FY 2025, as I mentioned at the beginning of the presentation. That's all for the third quarter results from myself. Thank you very much for your kind attention.

Now we would like to take questions. First, we would like to take questions from the floor. [Operator Instructions] Now we would like to take questions from the floor. The very front person.

MJ from Bloomberg. I have 2 questions, please. First, is Vision Fund 2 -- the shares has been transferred to Vision Fund 2. Why it was that? And follow-on investment has been speculated. And what is your stance as of today?

Thank you. You said shares. Are you talking about -- referring to OpenAI shares?

Yes, yes.

Yes. We have transferred OpenAI shares to Vision Fund 2 as Vision Fund 2 are the investment vehicle in OpenAI. About 2 years ago, we started these investment activities. Whether to use Vision Fund as an investment vehicle or should we invest from our own balance sheet. There are several discussion points internally.

And if this company has a road map for investment return, for example, IPO possibly, or any exit or divestment in the future? So do we have a clear view on the return? That's 1 of the discussion point.

And another one, is it majority investment or minority investment? That's another discussion point as well. From that sense, OpenAI, there are several other investors.

So as a way of return -- making a return that I believe 1 of the option can be IPO. And also, although the amount is large, but our holding ownership ratio is just 11%. So this is through the minority investment.

As a result, we have decided Vision Fund 2 as an investment vehicle in OpenAI. And for your second question, follow-on investment. I understand there are financing plans by OpenAI and so on newspapers, but there is nothing has been decided or no fact has been decided.

One more thing about market, I would like to ask you. So investment size, not only OpenAI, but the size of the ticket size is becoming larger and larger. And JPY 100 billion Vision Fund size is becoming more and more relevant to this ticket size, I believe. But as an AI investor, what is the investment pace that you would like to see going forward? For example, JPY 30 billion or JPY 50 billion every year? Or are you going to plan such investments?

I believe our financial policy manages that or rules that. So loan-to-value, how much the loan-to-value will be. So we do have a financial policy of 25%, managing less than 25%. And the cash position also needs to be maintained to cover 2-year equivalent of the bond redemption. So can we keep that or not?

So financial policy is not the covenant. So it doesn't mean that we cannot exceed $0.01 or anything like that. But as a large investment projects, can we maintain our healthiness and soundness? But even we step a little bit above, but still, are we able to return back to normal case relatively soon.

So those kind of preparation is necessary. If you only think about financing, there are many options for the finance. So I believe if you only look at the financing point of view, you'll be able to invest, but it doesn't mean that you can invest anything because you can finance or raise money.

I believe that we do need to focus, can we maintain our soundness of our financials as a company, that's the kind of important consideration we need to have, which will grow our amount of the investment.

Any other question?

[ Yagi ] from Nihon Keizai Newspaper. First, OpenAI's weekly active users exceeds 8 trillion you said. But in 2025. OpenAI expected over 1 billion active users and OpenAI issued code red because of the threat from competitors, we heard. For SBG as an investment company, what you find attractive? And what are they talking about to you as their attractiveness and requests for more investment, for example?

Well, as an investor, investment in OpenAI, which is private entity. Of course, we need to hear from them, what they are doing, based on which we will make a decision. That kind of information includes confidential one, between OpenAI and us.

The numbers that we present today are all public information and confirm with OpenAI that it can be disclosed. You mentioned code red. We have seen on media. Maybe because of competitors' move. And they maybe sent message internally taken into account of competitive landscape. I think that's the matter. Of course, it's a fair thing to do.

In our financial results and announcement of our financial results, the data set that we can share with you as information based on which we make investment. That's a part of information based on which we make an investment. OpenAI's technical capability, OpenAI's vision. That's something which is communicated between the senior management.

And that kind of information or communication is not shown in the numbers like we presented to you. In fact, that kind of confidential data or information or communication might be playing a more important goal for us to make an investment. But at this moment, I just appreciate if you could keep watching how OpenAI grows going forward.

One more question about the Stargate. In 2026, how much cash out is expected as Stargate?

Well, cash out as a SoftBank Group, well, if that's the case, of course, it's likely to change. So I cannot tell you exact number. But as we keep saying, target project is including something that SoftBank Group works on and Oracle works on. What I mean by saying work on is we do equity investment, but project itself for that is financed as project finance. So in terms of project finance, our size is limited. So the amount of investment should not be too huge. That's how much I can say.

Any other questions? The second row from the front, in the middle, please.

Suzuki from TV Tokyo. As you announced, the business is doing quite well, but last year, December -- November, share price is declined about 30% or 40%. How do you analyze the current situation of share price? That's my first question. Second is about the investment in OpenAI and the concentration of OpenAI investment may be concerned by the people. How do you look at that?

First of all, the share price itself -- actually I am thinking this is an increasing trend. When you look at each respective portfolio may affect a little bit, but the equity market itself and the environment itself is very steadily growing. So over 1 year, Japanese equity market as well as U.S. equity market does not have any immediate concern, which may ruin the current situations, which is my personal view.

If there is any recession in the United States, and the possibility becomes more clear, then we -- there may be some revisiting. But at this moment, I don't see any such negative events to coming -- happening soon.

And because we invest in wide variety and unless we have any market crush and otherwise, we believe it doesn't mean that we see all of our portfolio companies goes down. So we do have a kind of many -- wide range of the investments. So we don't concern too much about that.

And for your second question regarding concentration in OpenAI, which may be repeating myself, but there are still as a private company, many numbers or many trends are not disclosed or discussed publicly at this moment, which as an investment -- investor, we do have a dialogue and also have some information. And there are many communications between the top management and also their directions or their stories for the future.

And also their role and missions alongside with the AI technology evolvements. So something similar to other many hyperscalers, we assume OpenAI will be able to lead this industry in this era. And we are quite convinced so that we are making an investment in this company. Thank you.

Any other question?

Samura from Nikkei Business. About OpenAI, private entity -- because of that, there is an information that is not disclosed yet. But from people like us, we could be concerned that you depend too much on OpenAI. But you still have kind of hidden critical information that makes you very confident.

While hidden or not, in the private equity world, whether big or small, you need to make an investment with confidence even though there is undisclosed information.

The question is, what do you want to look at when you make a decision on the investment? Venture capitals, what are they looking at when they make an investment?

Let's say, SoftBank -- when SoftBank made investment in Alibaba, what did we see? I asked Masa. And Masa said he decided in 5 minutes after seeing Jack Ma's eyes. And there are a lot of private companies that still have founders and the founder's leadership and founders' capability could be an important source of information based on which venture capitals would make an investment.

Through our long-term experience of investment, we made a decision to make an investment in OpenAI, and we want to provide them a full support. A question about financing. The equity is going up globally. And you raised capital in the debt market as well. So since the stock price is so high, I don't know if there are enough investors who intend to buy your bonds.

Looking at groups of investors, some equity investors explore ups and downs and other investors may want to have confidence of return, fixed return. These kind of investments are in the credit market. I think in a sense, they are on the different end of the spectrum, and they are both there. There are always credit investors, and there are always equity investors.

From SoftBank's perspective, if we do equity finance, for example, of course, there are equity investors who want to buy our shares. But when you look at investment cost, that could be a huge -- sorry, financing costs will be very high. So that's why we want to utilize both markets. And also, we want to raise capital based on our balance sheet as well.

Any other questions? Second row from the front.

Shimada from TV Asahi. So Akazawa -- Minister of Finance, Mr. Akazawa is visiting the United States and talking about the first project. And there are discussion that the SoftBank Group are also regarded as a candidate for the first project from the gas turbine power generation plant construction. It doesn't have to be first project. Second project is also good. But what is the size? What is the progress of the project? How do you look at it? Is there anything that you'll be able to comment?

I understand Minister Akazawa is visiting United States, but this is the discussion between 2 governments between Japan and United States. So there are nothing I can make as a comment from my position. But this will be the project of the collaboration between Japan and United States. And if we can be picked up as one of the projects, we will be very honored. So we would like to patiently looking at the situation.

Thank you. Any other question from the floor?

Makino from NHK. Data center in the U.S., Switch is an operator of data center in the U.S., and there was a media report that you consider acquiring Switch. But later, you announced the acquisition of DigitalBridge officially.

So about potential acquisition of Switch, do you have any comment on that? And also, looking at data center business in the U.S., what kind of growth potential you see? And any other possibility that you make more investment in U.S. data center sector?

About Switch, the speculation of acquisition of Switch, that kind of report I saw here and there. But the fact is there is nothing that we can make an announcement about Switch. And about DigitalBridge through the fund, DigitalBridge owns 50% of Switch. And Switch, we know is a great company. And-- on business side, we have -- we think that we have an opportunity to communicate with the Switch about data center business in the U.S. There is media coverage or discussion like data center business is bubble or part of AI bubble. Of course, data centers are built here and there.

And whether it's overheated or not, the question is who will be users of those data centers? Hyperscalers are in great need of data centers that can report, you can often see in the media. In order for AI to grow further, computing power is needed and the data center capacity at the moment is not enough to support growing AI's computing. And that kind of computing is needed immediately. So from myself, from SoftBank's perspective, data center business is not a part of AI bubble narrative because shortage of supply is real.

So as business sector, investing in certain sector is valued and fair. From a risk perspective, I don't know risk is the right word or not, but technical innovation can be a risk. By that, what I mean is what kind of impact the technical evolution could have on data centers. Maybe some technology may come to provide capacity without using data centers and also electricity with some new technology coming up, maybe drastic paradigm shift may happen. So investment in physical assets has such kind of risk posed by further technical evolution.

Any other questions from the floor? So that concludes Q&A session. Thank you very much. This concludes the SoftBank Group Corp Earnings Results Announcement for 9 months period ended December 31, 2025. The video footage of this meeting will be distributed on demand from our corporate website. Thank you very much again for joining the SoftBank Group Corp Earnings Results Announcement for 9-month period ended December 31, 2025.

[Statements in English on this transcript were spoken by an interpreter present on the live call.]

[Interpreted] Thank you very much for waiting, everyone. Now we would like to start the SoftBank Group Corp. earnings results announcement for 6-month period ended September 30, 2025. First of all, I would like to introduce today's participant. From left, we have Yoshimitsu Goto, Board Director and CFO; Kazuko Kimiwada, Corporate Officer, Senior Vice President and CAO; Navneet Govil, Executive Partner and CFO; SB Investment Advisers and SB Global Advisers; and Jason Child, Executive Vice President and CFO, Arm. Today's announcement is live broadcast over the Internet.

Now I would like to invite Yoshimitsu Goto to present you the earnings results and business overview.

[Interpreted] Good afternoon, everyone. Thank you very much for joining us -- with us today. So we would like to have this earnings results announcement for the second quarter. First of all, I have to talk about this. As you -- many of you may know, I am also CEO of SoftBank Hawks, the professional baseball team in Japan. And thanks to all your support that we are able to have this Japan championship title first in 5 years. Fukuoka SoftBank Hawks, the company itself, has been growing very steadily, which I don't have much chance to share with you, but maybe some time that I will be able to speak about it as well.

Set aside, Masa, our Chairman and CEO, with the visit of President Trump from United States, he was on media quite often these days, and he's been going back and forth between Japan and United States quite often to achieve our vision. And also based on that vision, he's back calculating what do we need to do now. So based on that, he is working very hard to welcome the efforts that what we need to do now for achieving our vision and goal.

Especially the keyword that we have right now is physical AI. What that mean is that robot, for example, autonomous drivings. So AI and mobilities or robots will be combined together to provide you more complicated techniques or services. And that is coming very close to our life. And to be prepared for such time to come, we need to be prepared and we need to work on, so that we do also need to complement the area that we are not very good at.

2010, when we created SoftBank's Next 30-Year Vision, message never changed, but actually, this is a slide that we used back in 2010. Our fingertip is actually carrying heart and sharing is robot's fingertip. So brain computer and the muscles actually connecting, then robot or robotics will be mingled with the society so that the technology or robotics may be able to solve the unsolved issues or problems that the society is facing or societies will be facing. And SoftBank Group wanted to achieve the coexisting with the robotics or robots that has hurt or feelings.

Recently, we announced agreements on the acquisition of ABB Robotics, and we believe this can be a very big step forward to achieve such goal and vision. To realize ASI, what do we need to do? What we should be doing? That's something that we try to be more specified. So the Arm, which is the semiconductor business player, can be expanding the business domains for the chip area together with those companies and powers and data centers. And this time, adding the ABB Robotics has become a very big part of our AI model. And with this timing, having such a great company to be our family member is going to be something great, which we'll be able to show you the good result out of.

And here on that I would like to go into the more numbers or the financials. There are four key highlights. First is the fair value of investment in OpenAI, which has been growing dramatically since our investment. So it's been increased by $14.6 billion, maybe JPY 2.2 trillion or so. And with that, to number two, JPY 2.9 trillion record high net income for this first half. I've been seeing some articles showing the net loss, but actually, this time, this is a large net income increase. And net asset value, which is the most important KPI for the company, and this was reaching JPY 32.3 trillion at the end of September end. Since then, actually, this net asset value keep growing. And fourth, KPI to measure our soundness of our finance -- financials, which is loan-to-value and 16.5%, which is more than safe and at the same time, JPY 4.2 trillion of cash position. So safety -- soundness also has no issues with our business. So those four points are our highlights for the first half.

Now let me go into the more specific of the consolidated results. As you can see, sales, gain and loss on investment, income before income tax. And actually, those three are actually hitting the record high for the first half, JPY 3.6 trillion for the income before income tax and JPY 2.9 trillion for net income. So if you look back the history of net income, important thing is the kind of trend. JPY 2.9 trillion, this number is also very important and we appreciate very much. But at the same time, as an investment company, market shares has volatility goes up and downs, largely ups and largely down as well. But the investment company should not be too happy or too sad about those daily volatility.

What is important is we should have a momentum and trend of keep increasing in net income. And as long as we will be able to keep that trend, then that I think that we are doing the right thing. So you can see here, we do have ups and down, of course, but trend-wise, we are keeping increasing trend. And here is the gain or loss on investments and net income. The same things applies. There are ups and downs. But again, with the increase in fair value of OpenAI became the big driver for the gains for fair value at this time.

If you have a breakdown of this gain on investments, of course, large contribution is coming from OpenAI. It's about JPY 2.2 trillion or $14.6 billion. The reason we were able to have this result is because of September last year, that was the first time we invested in OpenAI. And OpenAI held several times of the financial round and the latest October round was JPY 500 billion valuation. As a private company, I believe this is one of the largest valuation or the fair value.

Our investment in OpenAI is here to simplify a little bit. So $10.8 billion is a cumulative investment cost. And this, along with the increase in fair value of equity interest, we were able to achieve this number this time. On top of that, as you may know, in total, we are investing up to $30 billion for this year, which contracted at the April this year. And the final investment payment will be happening this year -- by the end of this year. And that's happening from now, but this has already been contracted so that the share price for this payment or investment has already fixed as well.

So this is very much similar to forward contract. And the fair value of forward contract has also been finalized this time and recognized on an accounting point of view. So that is why you see this green bar increase in fair value of forward contract, which is $8 billion. In total, that leads to $26.5 billion. And you see the investment gain is the gap between $26.5 billion and $10.8 billion, which gives you $14.6 billion as an investment gain for second quarter.

And the key indicators. So these are the three KPIs are the most important indicators for the company and no change from the past. Net asset value at the end of September, JPY 33.3 trillion; loan-to-value 16.5%, it's a very safe and low level; cash position, JPY 4.2 trillion. Last year -- excuse me, last quarter compared to the June end, all these indicators are improved. Net asset value, latest number as of Friday last week, for your reference, this is a pro forma basis and JPY 36.2 trillion. So again, this is the record high number.

Loan-to-value, it remained at very low level. During the COVID, like 2021, we were exceeding a little bit above the 20%. Of course, the lower is better. But if you go too low, then that means we're not making any investment activities. So we need kind of a good position in between there. So '23 to '24, it's been kept very low to 7% or 8%. And from there, we came back to investment mode. And with all those investment activities, that gives us a 16.5%, which I believe is very good level of loan-to-value. Cash position, JPY 4.2 trillion, which is also very stable. If it's too much or too high, that's not efficient. So I think we also need to find a good level there, too.

Next, let me talk about SoftBank Vision Fund. Looking at SVF segment for the second quarter, we recorded $19 billion, driven by fair value gain on OpenAI. Of $19 billion, $12 billion came from OpenAI, plus others $6 billion. $19 billion is amazing. Excluding OpenAI, $6 billion shows strong growth backed by other portfolio companies. OpenAI, that kind of home run is needed when you manage a fund, but also you need regular hit like other investments.

On the right-hand side of this chart shows SoftBank Vision Fund 2. Vision Fund 2 is the vehicle to invest in OpenAI. So gain or loss from OpenAI is reflected on the right-hand side. Obviously, OpenAI takes a majority of the part. Left-hand side shows SoftBank Vision Fund 1. It showed $5.7 billion, which is also a big number. If you look at the breakdown, public portfolio companies, which have been growing since listing grew and also private investments grew as well. So both SVF1 and SVF2 grew, which is very meaningful in the second quarter. As a result, since inception, the cumulative investment sometimes performed good, other times performed bad due to poor market conditions. And there were some challenging times, especially in the COVID, 2020 or 2021. And we did appropriate accounting treatments like write-down, and we are beginning to see a strong turnaround.

It shows public listings and pipeline. Left-hand side shows public listings. Since inception, we listed at 58. In the second quarter, finally, Klarna was listed. We expect a further growth and bigger valuation. On the right-hand side shows pipeline. Total fair value of late-stage portfolio from our perspective is $56 billion. And PayPay and Lenskart already filed. In fact, Lenskart just recently, actually yesterday, listed after filing. And also Meesho and other companies already filed for listing.

And also left-hand side shows select late-stage portfolio companies, including OpenAI, ByteDance and Fanatics. Those big names together with OYO, Revolut, Yanolja, these are among the companies that is in our pipeline. And we expect them to grow their value further. And I hope that you will have a high expectation as ours too. Now PayPay. Timing of IPO, we cannot make a comment, as you know, but we can talk about business of PayPay. Number of registered users exceeded 71 million, as shown on the screen. It means one in two people in Japan used PayPay, and two out of three smartphone users use PayPay in Japan, which is significant. Two of their key indicators are GMV and EBITDA. GMV grew 25% year-on-year to JPY 9.2 trillion, whereas EBITDA doubled year-on-year to JPY 48 billion. They are ready for IPO, and they are performing pretty good.

Now let me talk about Arm. To put it in a one word, steadily growing. As you can see, they hit record high revenue, license and royalty revenues in the first half of the year. And non-GAAP operating income was up by 14% year-on-year. You can see how Arm is performing pretty good. Of course, they want to invest in R&D, which is growing as the business grows. But still, bigger income offset those R&D investment in even more. Let me pick up two highlights. One is a compute subsystem, which is one of the pillars to support Arm's business. It enables even higher royalty rates than Armv9 by utilizing multiple Arm technologies. In the second quarter, they got 3 more deals totaling 19 companies. Well, 3, 19 may sound very small. However, those include hyper players or big companies like Microsoft and Samsung. So those big companies have been adopters of Arm's latest technology.

Also, they announced Lumex CSS for mobile device, mobile devices like smartphone and laptop PCs. This CSS integrates the most advanced CPU, GPU and system IP to accelerate AI performance on edge devices. As you can see, compared to previous generations, you can see up to 5x AI performance, 2.8x faster audio generation and 3x better energy performance.

Now guidance. As you can see, revenue-wise, $1.225 billion or 25% year-on-year. As revenue grows, as you can see, investment in R&D is growing, but still -- we expect they can still post revenue and profit. But in fact, cost should be used as early as possible, which I think is true not only to Arm, but also SoftBank's business model. So properly and in a big fashion, we want to spend the necessary expenses.

Next is OpenAI. Since the investment in OpenAI, we spent about an year or so. I think their focus area was probably whether they will be successfully making a restructuring. If it was a former structure, they may be difficult to go into market and so on. But now that they have completed all the dialogues with the concerned party and completed the recapitalization. As you see, OpenAI Group PBC, the profit benefit corporation, which will be able to access to the market. As a shareholder, OpenAI Foundation, nonprofit organization will be holding 26% and Microsoft 27%; Vision Fund or SoftBank Group, about 11%, and other employees and other investors. So this is the current ownership ratios amongst the concerned party.

So the investment so far and the investment for the future that I would like to summarize once again. So April, we agreed investment commitments total $40 billion, of which $10 billion was syndicated out to the other investors, $30 billion is investment from the SoftBank Group. First closing completed back in this April, $2.5 billion co-investors syndication has also completed, and $7.5 billion has been also paid out. Second closing totaled $30 billion is paid by the end of December. Syndication has been already fully subscribed. $22.5 billion will be proceeding as scheduled by the end of December.

And I also like to check the business status of OpenAI. If you look at the weekly active users, continuously growing. As of October this year, they have already exceeded 800 million users. So that is actually making an even bigger gap between other competitors who provides a similar type of service. So as a comparison, a number of the app downloads can be one good example for you to refer to. ChatGPT, 871 million times of the downloads, which is also outpaces competitors.

Annual recurring revenue, which is also called as ARR. So as of September, which is the black bar, if you multiply it by 12, that's the number. And this is also making a significant increase. We believe that the subscribing numbers are increasing well. So that's going to be very positive for the financial numbers as well. ChatGPT is not the only service. Actually, as you see, they are providing several new release of the services one after another. These are also the good reference point for you to understand what kind of services OpenAI is providing.

And follow-on investment, the number-wise, it's not too large. But because they have a financial round, $6.6 billion of the offering has made, of which we have also invested or participated $1.45 billion. As you see on your right-hand side, other than us, there are other very well-noted investors are participating such as Thrive, Dragoneer or T. Rowe Price.

Stargate project has been announced back in January together with OpenAI and Oracle. And in September, we made a joint press release, and we have been seeing a clear path to securing full $500 billion of a 10 gigawatt commitment. We will be making a specific site structure to the project. And I think that we are coming closer to be able to introduce you each respective project. Another announcement we made about was the Crystal Intelligence. This is mainly played by SoftBank Corp., our domestic telecom company, together with OpenAI. And a few days ago, November 5, SB OAI Japan, who will be exclusively offering in Crystal Intelligence was launched.

We, as a holding company, SoftBank Group and SoftBank Corp, the mobile telecom operator, is jointly investing in the company called C Holdings as a holding company. And these C Holdings and OpenAI is investing 50-50 as a joint venture, which is called as SB OAI Japan. 2026, SB OAI Japan is aiming to start launching services. And with Mr. Miyakawa's lead, who is the CEO of SoftBank Corp., AI agent development, internal tests have been conducted so that they will be able to be fully prepared as an enterprise services for the enterprise customers.

Now let me dive into physical AI. This October, we agreed to acquire ABB's Robotics business in October 2025. So we would like to realize ASI and robotics fusions to pioneer humanity's future. And Marc and Morten, ABB CEO and ABB Robotics Division President, shared the time with Masa, our CEO, as you can tell from his smile. And this is a quick video to share with you.

[Presentation]

[Interpreted] Towards ASI. These four elements or areas are the efforts that we are making right now. How are we going to make it happen? We believe that these are the four key pieces that we believe is important. And this time, another piece is added to realize ASI, which is ABB Robotics. ABB Robotics business that I would like to touch on. They are second largest in the industry and located in Zurich, Switzerland. And they have about 7,000 employees, which includes many engineers. Revenue, $2.2 billion.

And this is a comparison to share with you the competitive landscape. Revenue is a second in the industry, and cumulative robot shipments actually also the second in the industry after FANUC. So they are having a position in very well place. So $5.3 billion, which about JPY 800 billion, which is, of course, a large investment. It was fully discussed at our Board of Directors meeting. FANUC, if we try to buy FANUC, their market cap is like JPY 5 trillion. We'll never be able to do. Of course, FANUC is a great company we understand because that's why their market cap is large. But this time, we are able to add ABB Robotics with JPY 800 billion for a family member as additional key piece, which is a great migrations for the company overall.

And their strengths that they have many R&D to manufacturing, sales to service. So they do have full platform capabilities and global presence in 44 -- over 44 countries with 7,000 employees, and 500,000 robots are shipped. So manufacturing plant head office located in Sweden, U.S. and China.

Physical AI's another agenda is autonomous driving. We've been investing in Wayve in U.K. They have unique characteristics of the autonomous driving technology. And actually, their world model enables autonomous driving without HD map. So not only U.K. where the head office is located, but they are developed into U.S. market as well. And in April, they also launched in Japan. It was a little bit of a struggle in the beginning, but actually, they are very steadily providing the autonomous driving in Japan right now. So another video to share with you. This is quite interesting.

[Presentation]

[Interpreted] I don't really understand the technology and autonomous driving, I don't know how that can happen, how that can make it happen. But I think that not only Masa, but also interview was also on the driving. But -- and I was so afraid that if the car hit the people, but it did made us very safe drive, and I'm very much expecting the good future of this company.

And the next one is the -- also on the investment portfolio on Nuro. Just announced a next-generation autonomous robotaxi program with Lucid and Uber in July this year. Lucid EV vehicle, Nuro autonomous driving system and the global ride-hailing platform, Uber, that with this framework around the second half of 2026, they are scheduling to launch in major U.S. cities and also planning to deploy worldwide over the next 6 years. So this is another interesting challenge that the autonomous driving technologies are facing and addressing.

So we've been making diversified robotics investments. And with the ABB Robotics, we hope that we'll be able to pursue further of commercial opportunities and synergies. You just saw the video of autonomous driving. But other than that, we have robotics areas like distribution systems or the auto factories. And so there are several portfolio companies that we have invested so far. So with that, organically that we would like to create good synergies among those members, so that we'll be able to create another level of the value.

Now well, Masa and President Trump were often covered by media. And as you know, Japan and U.S. signed in September 2025, the memorandum of understanding on strategic investments which outlined Japan's commitment to investment up to $50 billion in the U.S. Many Japanese and American companies, including SoftBank, formed up to $332 billion funding framework. As you can see, well, Masa, for some reason, always sits in the middle. And SoftBank Group is one of the members to support critical power infrastructure in the U.S. This was the photo taken when the framework agreement was signed with President Trump; Commerce Secretary, Lutnick; and Ambassador, Glass. Masa wore red tie.

Let me touch upon financial strategy now. Financial policy that we communicate with you for a long time. It remains unchanged. We maintain LTV below 25% in normal times. Sorry, I keep saying again and again, 25% itself is pretty safe. But upper threshold of 35% even in times of emergency, it may happen in the future, whether 25% or higher or lower, but we want to maintain the level. Also, we want to maintain at least 2 years' worth of bond redemption in cash. It shows investment amounts. Since the beginning of 2025 in the first and second quarter, OpenAI's first half total was $8.6 billion, including $1.1 billion in second quarter.

Looking at Q3, next round of investment in OpenAI is planned to be executed at $22.5 billion. And also all process should be completed by end of December to invest in Ampere. So all in all, in the third quarter, we estimate about $30.5 billion of investment, and we are ready for the funding. With completion of the estimated investment, investment in OpenAI in total would be $34.7 billion. I said that we are ready for funding, and how we do? I'm not going to go into details in numbers per deal or per transaction, but we have three pathays. Cash position, obviously, we have access to. And other than that, we can utilize assets held. NAV is net asset value. Here, we are looking at asset value of holdings, which is around JPY 40 trillion. Of that assets, 70% is listed shares. So by utilizing different tools, we can utilize assets held, for example, for margin loan or derivatives or collar transactions or forward transactions.

We have a variety of options that we could utilize, and that's financing with asset-backed finance. Also, we can utilize debt market. Domestic and foreign bonds market are the markets that we can utilize to issue bonds. And not only Japanese mega banks, but other Japanese banks and foreign banks, including JPMorgan, give us a strong support always. Not only just simple loans, we could do hybrid funding with 50% traded as equity capital. So such funding can lead to strengthening capital. So we want to provide a lot of investment opportunities for investors while we can still maintain financial strength. So through those options and tools, we make sure that we are ready for funding in a very safe manner.

Talking about asset monetization. We sold T-Mobile shares, for example, and we sold NVIDIA shares, which we held for a long time. And also, we sold Deutsche Telekom shares. With Arm shares, we upsized margin loan to $20 billion, and $11.5 billion remains undrawn.

To summarize our financial highlights in the first half of this fiscal year is shown on the slide. Our share price recently have been going up and down dynamically. But since the beginning of the year, we grew 3.8x. We want to provide as many investment opportunity as possible. And also, we need to pay attention to TSE's guidance. We recently decided to do share split ratio 1:4. Based on closing price as of November 10, 2025, minimum investment amount would go from JPY 2.2 million to JPY 0.5 million after the share split. I think it's needed to be -- go cheaper for retail investors to utilize for their individual savings account investment.

Last but not the least, our initiatives in physical AI, which is to combine brain and body. We want to figure out how much impact this could have on our daily life, and we want to really create happiness for human. And in this sector, we try to be a leader. Hopefully, at AGM this year, Masa claimed that we want to be #1 ASI platform provider after 10 years. And we hope that you get a flavor of that vast idea in my presentation.

Thank you very much for your attendance. Thank you.

[Interpreted] Thank you. Now we would like to take questions. First, we would like to take questions from the floor. [Operator Instructions] We would like to take up to two questions per person so that we can take questions from as many people as possible. So now we would like to take questions from the floor. On your right, third row from the front with the jacket.

[Interpreted] Toyoshima Hironao, TV Tokyo. Goto-san mentioned regarding the dynamic movement of share price. Recently, especially including your share price, AI-related share price volatility is very wide. Therefore, stock market is questioning that where are we in terms of AI investments. So overheating in expectations for the AI business, is it just a beginning? Or is it overvalued, so that we need to be very worried? Which is the -- your stance? That's my first question.

And related to that, my second question, the share price of SoftBank Group has been driving at Tokyo Stock Exchange share price largely as well. So AI portfolio companies, including OpenAI, are they appropriate? It's like a heating like a so-called AI bubble, that's also the risk for some people say. Do you feel any risk for that?

[Interpreted] Thank you for your question. So two questions, but I believe some of them are overlapping. So right now, many service and products will be developed in AI era and how you value that is very important, but also difficult question. And the market tried to obtain that in advance to the others. And sometimes that you are too much in advance, so that you try to withdraw a little bit. So that kind of movement has always happened in the history as well.

For example, back in year 2000, a so-called Internet bubble. Was that really the bubble or not, was actually valued after the bubble finished. So 25 years from now, and when we look at -- look back, was that really the big volatility? Not really. So that is because based on the autonomous movements in the market, sometimes movements made in advance or sometimes it takes a little bit withdrawing. Right now, that the size of the market is larger, so that, that movement is becoming bigger as well. And as an investment company, we look at this market and our portfolio company, how do they face and materialize their AI technology or service into their value. And that's we are trying to support with the way of the investment.

And I don't think I can answer whether this is bubble or not. We'll be able to tell later, but this is because of the technology just the beginning, and there are lots of movement, which is I believe is a good thing. And as a finance activity for the investment company, how do we capture the movement while protecting our financials without losing any opportunity, that will be very important for the company.

Are we going too far? That's a difficult question to answer in one word, but that's the kind of a big picture I have. So do you believe that the risk is bigger if you don't invest? Is that SoftBank Group's point of view? I think it depends on the people, but our company believes in that way, yes.

[Interpreted] Next question from the floor.

[Interpreted] Mirai, Diamond. I have two questions. First, about the Stargate project. Progress of the Stargate project, could you share with us? According to the announcement as of September '24, it sounds like five sites or data center sites in the U.S. and $400 billion for the next 3 years, those are finalized. Of five sites, two, Rosetown and Texas are the centers where SoftBank and OpenAI are involved. So I wonder, should we focus on those two sites as you get involved with OpenAI? And of $400 billion for the next 3 years, how much would be borne by SoftBank Group?

[Interpreted] Thank you for your question. The announcement in September was done, and two sites, Rosetown and Texas were referred. Since then, there have been some changes. As for the Texas deal, we are on track. And the scale and the time line, once a project is finalized, we would love to share with you. The another site in Ohio, we made a slight change against the original plan in order to meet demand. So the site will be more for manufacturing equipment for AI. That's a slight change of plan. But not only those two sites, but also there are a lot of many candidate sites. And we will figure out how we will proceed with the project. And once each and every project is finalized, we can share with you how much will be borne by SoftBank Group.

[Interpreted] The next question may be related to the first question or not. When Mr. Trump came to Japan at the dinner, $550 billion of investment scheme was announced there. As you shared with us the photo, SoftBank Group made announcement that they wanted to involve up to $25 billion, especially in powered infrastructure. I wonder specifically, what kind of project you have in your mind, $25 billion that is. SoftBank Group wants to do power infrastructure for things like Stargate project. Is my understanding correct?

[Interpreted] On that, I think same as other companies that showed their interest. At the right time, we would like to communicate with you more properly. At this moment, it's early to tell you more in detail. I appreciate your understanding.

[Interpreted] Thank you. Any other questions? The middle row, third from the front, gray jacket. Gentlemen, please.

[Interpreted] My name is Miyajima from monthly magazine Facta. Congratulations on the Hawks Championship, and share price also doing great. So this fall, it's a golden autumn for you, and for the company isn't the best season. Is it right? That's my first question. And in addition to that, robot, I believe that Masa has been working for about 20 years, and Pepper didn't do well. And now you are making a bet on ABB Robotics. He is now 68 years old and how passionate he is and he tried to address robotics. And FANUC is a great company like you referred to. And I believe ABB Robotics on about 1/3. And I believe there are many efforts needs to be taken place. And with Masa's passion, he needs to work on, but I think this is going to be his revenge on the robotics. And what kind of discussions are taking place internally in terms of robotics now?

[Interpreted] So first, for me, together with the great numbers, financials this quarter, but this is very relieving autumn for me with the great results. And Masa, on the other hand, Masa is even more passionate, energetic. So we need to run along with him. As for robotics, as you just said, for Masa, robotics is something that he has a deep in mind. And that has been for his important factor. We started it with Pepper. Pepper was a very cute robot. And because we have done Pepper business, so that is why that we are able to grow our robotic segments or robotic thinking.

And now with the great collaboration with ABB Robotics and how we're going to develop in the global market, I believe Masa really can't stop thinking about it. He was so excited that he cannot stop thinking about how we're going to create and draw the future. So he is 68 years old, but he doesn't look like that at all. He is still very energetic. He's passionate and energetic even that we're going to be burned from that with the heat, but we are trying to run along with him.

[Interpreted] Thank you. Any other questions on the floor?

[Interpreted] [indiscernible] from Nikkei Newspaper. First question, additional investment in OpenAI. You mentioned that there was a round in October and SBG, the portion was JPY 1.4 billion. Can you tell us when are you going to pay? And OpenAI has the biggest shareholder, Microsoft. And do you have a target equity ratio of OpenAI? How much percentage of OpenAI you want to have?

[Interpreted] October, we completed a round. The amount-wise, it was not big. So the payment was already done. About the target percentage of stake in OpenAI, well, we believe that we have the strongest expectations in OpenAI than any other company. So the higher percentage is better, I would say. But now we have 11% after the recapitalization, that's the current percentage at this moment.

[Interpreted] Second question is about NAV. JPY 3.3 trillion as of end of September. Calculating from that, NAV per share will be about JPY 23,300 or something. You announced a share split recently. So I wonder what's your view on the discount and whether it's included in a premium or not. What's your view?

[Interpreted] It's not appropriate to make an comment about stock price. So in a different way, if I may, I would say there are a lot of ingredients with which we can grow. As you can see in our history, we always have a lot of things to do, but not all of them are appreciated. Sometimes most of them were not appreciated, and that resulted in a huge discount. And trend-wise, that could continue for some time. So again, please keep watching us how we grow.

[Interpreted] Any other questions? On your left, second from the top.

[Interpreted] I'm MJ from Bloomberg. I have two questions. First, regarding NVIDIA. You mentioned as a part of the monetization you sold and it was October. Why did you -- did you sell NVIDIA in October time frame? Can you give us more color? Is it because to -- was that the proceed for the OpenAI investment or because of the level of the share price of NVIDIA?

[Interpreted] As an investment company, we're always repeating the new investment activities and divestment activities. And out of our portfolio companies, what kind of asset should be replaced with what kind of assets, that's something that we always keep in mind. This year, the OpenAI investment is large. Even more than $30 billion investment needs to be made. So for that, we do need to divest our existing portfolio so that, that can be utilized for our financing.

We don't have any specific meaning in October or it's nothing to do with NVIDIA itself. I don't think that I should make any comments about that. And the objective was -- as you mentioned, that's our objective.

[Interpreted] My second question, you mentioned about Stargate earlier. Announcement in September, OpenAI, Oracle, 4.5 gigawatt, $300 billion plan was announced. And the SoftBank name was not really referred to this project. Are you involved or are you not involved too much about this project?

[Interpreted] Stargate project is shared with the OpenAI, Oracle and SoftBank Group to build and create a mega data center for AI. Oracle has its own data center business that traditionally, they've been running the data center business. And they partner with OpenAI, sometimes that we partner with OpenAI or three of us address the same project. So Stargate project has quite a wide range of interpretation there.

[Interpreted] Any other questions on the venue?

[Interpreted] Makino from NHK. Two questions. First, you mentioned that OpenAI contributed a lot to your growth performance. And Goto-san, how do you see OpenAI's growth potential?

[Interpreted] I think easiest way to understand OpenAI's growth potential is their products roll out. GenAI products like ChatGPT, how their products are positioned compared to other competitors' products. Obviously, they are leading others and competitors are trying hard and OpenAI are trying even harder. And the position of OpenAI, which is in a strong leading position is harder to catch up. And OpenAI keeps enhancing their products going forward. So long as OpenAI keeps trying harder for some time, I think compared to competitors, OpenAI is in the strongest position. That's how we see OpenAI, and we are confident.

[Interpreted] Next question. You made announcement to acquire ABB Robotics. And data center and power, you have been making a lot of investments in those sectors. But going forward, what would be our focus areas of investment?

[Interpreted] Very difficult question. I mentioned four sectors, and all of them are important for us. Not in these sectors, autonomous driving or other logistics and manufacturing. As AI will transform lifestyle, there are a lot of areas that would contribute to such growth. And we try to invest in great companies. Of course, we cannot do everything. So we want to be selective, and there are a lot of more things we can do.

[Interpreted] Thank you. Any others? Yes, a person who raised hand.

[Interpreted] Asamura from Nikkei Business. Not only SoftBank Group, but these days at the stock market, there are people who -- the companies are doing the capital recycling, just buy and sell, not keeping it long. So buy and sell is, I believe, something similar to SoftBank Group and many people respect the way that you do. And you are buying in good time and selling in good time. What is the secret sauce?

[Interpreted] I don't think we are managing well too much. When you sell or when you buy, you try to buy and bottom, you'd never be able to do that. If you sell, of course, you want to buy, sell in ceiling, but you cannot do that. But the capital recycling that you just mentioned, many companies do. They have their own business as well. They have their own operation. But we are quite unique because we have our balance sheet investments and the fund investments, both we do. So buying and selling itself is our main business. Then new investment and divestment, and it does not happen at the same time. So we need a bridge, which is the leveraged finance. And combining all that and to make the company grow, that's something we try to do. And that's how we are right now, I think.

[Interpreted] Any other questions from the venue? If not, that's all from questions on the floor. Next, we'd like to take questions from Zoom participants. [Operator Instructions] First, Masuno-san from Nomura Securities.

[Interpreted] Jason-san, I have questions to Jason-san. First, in the earnings call, license revenue for the SoftBank Group in this quarter from April to June to next quarter from July to September, which was JPY 178 million. So my question is, what kind of license between you and SoftBank Group when Arm develops a new product and you make it license to be provided to SoftBank Group, I assume. So how you design license between SoftBank Group and Arm? And second, Arm is investing in R&D more and more. And from tradition business, you expand to chiplet and other business. So my question is, at what time -- how much scale of the respective new areas you anticipate or you expect?

Can you hear me? Okay. So on the first question, there's actually three deals that Arm and SoftBank have entered into over the last year. And those are each for different successive generations of projects where we've sold a license to SoftBank to -- for Arm IP. And then we're providing design services for SoftBank and SoftBank portfolio companies to work on projects that could be used certainly in part for Stargate as well as for others. We have not announced exactly what those projects will be. However, you should expect us to probably provide more detail when those projects are closer to realization sometime probably in the next 12 months or so.

For the second part of the question, the increase in R&D or engineering expense has stepped up pretty significantly. Part of it is the design services that SoftBank is paying us for. So that's covering part of those costs. And also some of it relates to other investments and exploration into basically full SoC projects. And that is also something that we have not announced exactly what those projects will be. But as we get closer to that realization, and there's three specific milestones that we need to see. We need to tape out a full SoC, then have samples back, which typically happens 4 to 6 months later. And then lastly, have to have noncancelable POs from a customer.

And so as soon as we realize all three of those milestones, that's when we will announce exactly what these full SoC solutions could be. And until then, we continue to explore and not yet ready to announce. But again, you should probably hear something about those efforts and what we will actually be delivering sometime within the next 12 months.

[Interpreted] Just a follow up. So tape-out for SoC is different from your project with the SoftBank Group. Is my understanding correct?

It could be. I didn't say exactly for SoftBank, that project, we haven't laid out what the milestones are. However, you should assume it'd be somewhat similar, and that is once there is a project that has been fully designed, which is what the tape-out is. And there are samples and that there is a customer that's going to provide noncancelable orders, that's probably when SoftBank would announce that project. But SoftBank would announce that in this case because we're designing something for SoftBank. So they'd be the one announcing that.

[Interpreted] The next is from Mr. Yasui, UBS Securities.

[Interpreted] I have questions regarding Stargate. I have about three detailed questions. First, it's mainly in United States, but some outman is traveling in many places around the world talking about Star project -- Stargate projects other than United States. Any possibility SoftBank is going to be involved in other area? My second question is, I believe that you're taking some time to the site selection, but also power shortage is also becoming an issue. So if you can -- when you decide the site, do you see any problems or bottleneck?

And the third question is about project finance, I believe it is going to be taken place for this project. So I believe that the major banks in Japan, are they the one -- the lender for the project finance rates, any ideas?

[Interpreted] Yes. So for your first question, yes, in many places, Stargate name is picked up on media. And in the future, we may involved in many ways for the future. However, right now, our key focus is in the United States. For your second question, do we see any bottleneck? Not really the bottleneck. But first, we need to firmly select the site suitable for the project. That's our process. In the future, we may face potential bottleneck, but that will be the challenge ahead. But actual things to do is creating a big box and also providing the power there. So there will be no -- if there is any bottleneck, maybe the delivery time for the materials, parts, those can be a potential challenge or bottleneck in the future.

And third, cost for financing. I think it's going to be quite a large project finance in each one after another. So three major Japanese banks, they have a good expertise in the global project finance as well as other global financial institutions. JPMorgan and others are also expected to participate. But when it comes to cost, of course, we need to see the market -- standard market level. But because of the variety of the reason, there may be some increase or decrease of the cost. For example, is it higher possibilities of payout or not? That's something that we cannot really tell at this moment. So please -- hope that you understand.

[Interpreted] This concludes the SoftBank Group Corp. earnings results announcement for 6-month period ended September 30, 2025. The video footage of this meeting will be distributed on demand from our corporate website. Thank you very much once again for joining the SoftBank Group Corp. earnings results announcement for 6-month period ended September 30, 2025.

[Portions of this transcript that are marked [Interpreted] were spoken by an interpreter present on the live call.]