Hua Hongmiconductor-h Aktie

Ladies and gentlemen, thank you for standing by. Welcome to Hua Hong Grace Semiconductor Second Quarter 2026 Earnings Conference Call. Today's call is hosted by Dr. Peng Bai, Chairman and President; and Mr. Daniel Wang, Executive Vice President and Chief Financial Officer. [Operator Instructions] The earnings press release and second quarter 2023 summary slides are available to download at our company's website, www.huahong.com. Without further ado, I'd like to introduce you to Mr. Daniel Wang, Executive Vice President and Chief Financial Officer. Thank you.

Good afternoon, everyone. Thank you for joining our Q2 2026 earnings conference. Today, we will first have Dr. Peng Bai, our Chairman and President, provide an overview of our second quarter performance. I'll then take you through our financial results in detail and offer guidance for the upcoming quarter. We then open the floor for a question-and-answer session. With that, I'll turn the call over to Dr. Bai.

Thank you, Daniel. Good afternoon, everyone. Thank you for joining our earnings call. We continue to improve our operational performance in the second quarter of 2026 with profitability strengthened further. Revenue hit a record high of USD 717 million, representing a year-on-year increase of 26.8%. Gross margin stood at 16.5%, up 5.6 percentage points year-on-year. Both metrics beat guidance and achieved sequential growth. Net profit attributable to shareholders of the parent company amounted to USD 38.6 million, posting substantial growth both year-on-year and quarter-on-quarter. Hua Hong Grace maintained a high fab utilization rate in Q2, delivered growth across all process technology platforms, especially the stand-alone and embedded nonvolatile memory products.

The improved business performance came as a result of rising volumes and prices. Since the beginning of the year, the global semiconductor industry has witnessed a strong AI-driven upturn in demand, first on memory IC products then spreading to logic and analog IC products that are associated with AI applications. As a specialty technology foundry serving a broad marketplace, we have clearly seen an overall positive impact on our business by the AI wave. We have also seen divergence in intensity and strength of market demand depending on end user market segments. Amid the rapidly evolving industry landscape, our strategy of steady capacity expansion, ongoing specialty technology upgrades and continuous capacity product mix optimization will allow us to capture growth opportunity to provide substantial improvement in our business results.

Hua Hong Grace has recently obtained registration approval from the China Securities Regulatory Commission for our acquisition of Huali Microelectronics. Integration of the acquired assets into Hua Hong Grace will strengthen our technologies portfolio, increase our operational economy of scale and improve our profitability, injecting fresh momentum into our future growth. Now I would like to hand the call over to our CFO, Mr. Daniel Wang, for his comments. Daniel?

Thank you, Dr. Bai, for your very inspiring remarks. Now let me walk you through a summary of our financial performance for the second quarter and then provide our revenue and margin outlook for Q3 2026 before opening the floor for the question-and-answer session. First, let's review our financial results for the second quarter. Revenue reached an all-time high of $717.5 million, 26.8% over Q2 2025 and 8.6% above Q1 2026, primarily driven by increased wafer shipment and improved average selling price. Gross margin was 16.5%, 5.6 percentage points over Q2 2025 and 3.5 percentage points above Q1 2026, primarily driven by improved average selling price and cost reduction efforts, partially offset by increased depreciation costs.

Operating expenses were $109.1 million, 11.4% over Q2 2025 and 3.3% above Q1 2026, mainly due to increased labor expenses. Other income net was $2.2 million, 79.4% lower than Q2 2025, primarily due to increased finance costs and the decreased government subsidies, partially offset by increased share of profit of associates. The other loss net was $2.4 million, mainly due to increased share of profit of associates. Income tax expense was $7.6 million, 7.5% over Q2 2025. Profit for the period was $3.9 million compared to a loss of $32.8 million in Q2 2025 and a loss of $17.3 million in Q1 2026. Net profit attributable to shareholders of the parent company was $30.6 million, 385.9% over Q2 2025 and 84.6% above Q1 2026. Basic earnings per share was $0.022, which is $0.022 340% over Q2 2025 and 83.3% above Q1 2026. The annualized ROE was 2.4%, 2 percentage points over Q2 2025 and 1.2 percentage points above Q1 2026.

Now let's take a closer look at our Q2 2026 revenue performance. From geographical perspective, revenue from China was $563.7 million contributing 78.6% of total revenue, an increase of 20% over Q2 2025, mainly driven by increased demand for MCU, flash, general MOSFET, logic and smart card ICs. Revenue from North America was $93.8 million, an increase of 77% over Q2 2025 and mainly driven by increased demand for other power management IC and MCU products. Revenue from other Asia was $32 million, an increase of 11.6% over Q2 2025, mainly driven by increased demand for super junction and MCU products.

Revenue from Europe was $20 million, an increase of 9.1% over Q2 2025, mainly driven by increased demand for MCU and smart card ICs. With respect to technology platforms, revenue from embedded non-volatile memory was $200.1 million, an increase of 41.8% over Q2 2025 mainly driven by increased demand for MCU and smart card ICs. Revenue from stand-alone nonvolatile memory was $68.8 million increase of 149.3% over Q2 2025, mainly driven by increased demand for flash products. Revenue from power discrete was $182.3 million an increase of 9.4% over Q2 2025, mainly driven by increased demand for general MOSFET products.

Revenue from logic and RF was $83.2 million an increase of $21.3 million, 21.8% over Q2 2025, mainly driven by increased demand for logic products. Revenue from analog and power management IC was $183.1 million, an increase of 13% over Q2 2025, mainly driven by increased demand for other power management IC products.

Now turning to our cash flow statement. Net cash flows generated from operating activities was $330.1 million, 99.3% over Q2 2025 and 159.2% above Q1 2026, mainly due to increased receipts from customers, Capital expenditures were $356.6 million in Q2 2026, including $325.9 million for the 12-inch facilities, and $30.7 million for the 8-inch facilities. Other cash flow generated from investing activities was $25.4 million in Q2 2026 including a $25.4 million receipt of government grants for equipment, $8.6 million of interest income. $7.3 million dividends and $0.2 million receipts from the disposal and equipment, partially offset by a $16.1 million investment in equity instruments.

Net cash flows used in financing activities was $406 million, including $569 million of bank principal repayments. $37.6 million interest payments and $1 million lease payments, partially offset by $201.5 million proceeds from bank borrowings and $100,000 proceeds from share option exercise.

Next, moving to the balance sheet. Cash and the cash equivalents was $4.53 billion on June 30, 2026 compared to $4.8679 billion on March 31, 2026. Other current assets increased from $894.6 million on March 31, 2026 to $936.2 million on June 30 2026 mainly due to an increased value add tax credit. Property, plant and equipment was $7.2863 billion on June 30, 2026, compared to $7.8059 billion on March 31, 2026, primarily due to capacity expansion. Interest bearing bank borrowings decreased from $2.8972 billion on March 31, 2026, to $3.5675 billion on June 30, 2026, primarily due to repayments of bank borrowings.

Total assets increased from $14.9473 billion on March 31, 2026 to $15.2258 billion on June 30, 2026. Total liabilities decreased to $5.5284 billion on June 30, 2026 from $5.663 billion on March 31, 2026. Debt ratio decreased to 36.3% on June 30, 2026 and 37.9% on March 31, 2026.

Well, finally, let's discuss our outlook for the third quarter of 2026. We expect revenue to be in the range of $770 million to $780 million, with the projected gross margin of 16% to 18%.

This concludes my financial remarks. We'll now begin the Q&A session. Operator, please assist.

[Operator Instructions]

Our first question comes from the line of Leping Huang of Huatai.

Dr. Bai, so first congratulate for the very strong results. So my calculation shows that you deliver another 3% Q-on-Q ASP growth this quarter. could you impact what drives this ASP growth? And whether it's from pricing or some mid-change and how do you see this ASP trend in the second half and beyond? Also, we noticed the largest foundry in the world now also say they were reemphasizing these mature and specialty node process to serve their customers. So plus also the domestic peer also want adding capacity.

So how -- what's our view? So on this material supply-demand relation in the next few years and how Hua Hong can differentiate from peers and further improve the profitability ahead.

Thank you,. You have a number of questions in there. I try to cease them all and answer them 1 at a time. In terms of pricing, as you know, in our industry, the pricing is set by market is basically by the balance of the supply and demand. Since the beginning of the year, we have -- we started to see the demand going up and the balance is shifting towards tightness in terms of supply situation. So as a result, that has driven up price increases. Mostly in MCU and memory area, and payment area. Those are the areas that are more associated with the AI applications.

Of course, there are also yields in the consumer segment, but there are probably -- the AI demand opportunity is probably more significant, and that's why we've seen the supply/demand balance shifting towards demand and supply being tight. In fact, some of the products we clearly cannot meet the demand. The order we are receiving is anywhere between 1.5x to 2x of our capacity. So as a result, we are doing everything we can to basically optimize our capacity structure and try to produce more across the board which is somewhat difficult right now because we have been pretty much 100% loaded.

So it's really for some very hard-working innovation, trading squeeze more. Of course, we had -- we are fortunate to have a fab that was still going through the capacity expansion. So that's why we can still we can still get good capacity increase from the Fab 9Athat we have, that still going to the capacity ramp up. So in a way, if you ask me what is our advantage? We have that advantage, I think it comes on 2 ways.

One is our technology capability, clearly is from domestic standpoint and industry leading for in many, many areas, and some of them are also on par with our international competitors. So that gives us a strong foundation to basically build our capacity and serve our customers. Another thing is since last year or the last year, we -- nobody predicted this year it's going to go up. But since last year, we were also -- we were quite steady. We were quite determined to continue to increase our capacity, that decision or that strategy of steadily expanding our capacity actually, you can say -- you might say that it has a pay off somewhat as this year come in and the market is turning upward.

In terms of the future pricing prediction is really a function of the latest demand wave, how long is it going to last it's a debatable point sort of a discussion or debate amongst the industry people but short term, I think for the second half of this year and as well as 2027, that most we still believe the demand will continue to be strong. I share that view. So in that sense, I expect our price increases will continue throughout the second half of the year. Some of the pricing actions we have taken over the last quarter, it will start to manifest itself probably in the second half even next year.

So I do think that the upside this upturn in demand where we are accompanied by our continued ability to increase the price a little bit. I don't want to caution everybody that we are not like a DRAM market. There's a multiple increasing the volumes we're talking about few percentage -- in percentage term. I do think it is a nice turn for the better, and we should continue enjoy the for the foreseeable future, at least through second half of this year and perhaps through next year. Thank you. .

Okay. It's -- so the second question from me is about the memory. So the largest China-based DRAM company just listed in Asia recently, and we see very strong investor interest on China's memory industry and at the same time, looking at the global perspective that providing the logic die foundry service to memory makers has been -- become a new trend these days. So Dr. Bai, so can you share some -- your view how Hua Hong can benefit from this memory build-out in China and globally?

And do you have any view that you plan to corporate with the China global memory makers on providing similar logic die service?

Okay. The memory can be different types, like the 1 we are seeing the biggest uptick in memory in the DRAM -- the second [indiscernible] NAND, we are not directly participating in DRAM nor NAND, but we do have a substantial business in NOR Flash business, which we have seen demand increases this year, and they're probably going to continue for second half of this year and next year. So we do enjoy the demand uptick there.

In terms of how do we -- I think the fact that the memory is going up, it's really representing the overall demand for semiconductor is increasing. So in that sense, you definitely benefit everybody benefits the memory more directly because it probably would go up there faster. But it does also benefit a large foundry. Our -- I would call us as a specialty technology foundry, which is -- we have a lot of products in logic, in analog and some specialty memory like NOR Flash. So in a way, the fact that DRAM is seeing the biggest demand increase is truly -- it just to represent the fact that the AI has been driving a lot of demand increases for overall in the semiconductor.

So in that sense, it's definitely. So we do benefit from the overall semiconductor demand increase. Specifically to DRAM or even NAND because the technology direction there is such that it tends to start to have their product tends to try to tie basically [indiscernibl] is not the right word. [indiscernible] their product, they try to separate the memory elements from the peripheral logic into 2 different bands and to some kind of 3D assembly to put them together as a product.

So in that sense, if the memory houses want to spend more time or focus more on the pure memory element. They may basically -- they might let the logic -- the perfect logic die to be manufactured by the logic foundries. In essence, we do see the larger houses probably will start to explore that. Collaboration with logic foundry for us, so we can focus on truly what is their specialty, which is the memory the bid, the memory part of the overall memory product.

So -- but that is still probably in the early stages of this technical transition. And -- but overall, it is moving the direction and they might even create some new demand for logic foundries because they are logic large portion of their monolithic guys, they might get separate out the separate and give it to logic foundry manufacturing if I explained that clearly.

It's very clear.

The next question comes from Ziyuan Wang of CITIC Securities. .

Okay. This is [indiscernible] from [indiscernible]. My first question is, could you -- we see a great guidance show the solid growth in Q3 and could you break down the Q3 revenue guidance to show how much is driven by ASP increase? And how much is driven by the capacity expansion? And also regarding on the expansion, approximately how much capacity will be added in Q3 and Q4?

So let me take on the capacity expansion part. I'll let Daniel to talk about the guidance for Q3 in terms of how it breaks volume increase versus pricing as I think that's what you're asking. So the capacity increase, our Fab 9A in Wuxi, you will see -- it will ramp up to peak to the to the total capacity in Q3 next quarter, you have all the equipment in that Fab installed. And so we will start to load the fab focus 100% starting in Q3, but the output probably will start to show up in Q4 or next year. So I think the 2027, you should expect a full fab worth of output from Fab 9A.

As you know, we do have a another half that's under construction that started in March of this year. And that we have -- we start having equipment installed in Q3 as well. So we expect there we have -- we've got a complete line in Q1, and we will start to have a small volume coming out. So next year, throughout 2027. We expect we will further capacity ramp up from the next fab, which we will close at May 19.

Now I will let Dan to talk about revenue breakdown the guidance between volume and the price increases.

Thank you for the question. So we expect the revenue is going to be between $770 million to $780 million that's our projection for Q3. The increase is largely coming from MCUs that whole sector -- embedded memory will continue to grow strong. There will be a double-digit growth. And the stand-alone volatile memory continue to be very, very strong in Q3. And I think this trend will continue throughout the year and into 2027 as well. And our power discrete, especially the low voltage products we're talking about the MOSFET business and also the medium voltage products. They are also going strong. The IGBT super junction virtually flat, virtually flat. And then on the logic and RF side, I think there's going to be a pretty strong momentum from the RF as well. And other than that, I think we see strong momentum coming from power management IC and analog business as well.

Even though no is still a small segment, its point, they're also growing pretty strong in Q3. So overall, it is when you look at technology platforms, these are the -- what I just discussed are the major drivers and in terms of revenue increase when you look at ASP and volume, it is really a split, I think, anywhere we're looking to 7% , 8% increase on revenue. I think I would say 60% coming from ASP and also another 40% were coming from increase in volume. Thank you.

My second question is about the capacities transit or switch since our demand is strong, is it possible to flexibly switch our capacity between product, that's like can we shift the CIS capacity to memory products and which type of capacities allow this kind of conversion. And also, what impact of such conversion have onto the overall ASP in maybe next quarter or second half? .

First of all, the capacity some was fungible to certain extent they are fungible, meaning that if we build 1,000 capacity for certain technology platform, you can also -- you get some conversion rate, you can use that divested to build something else. So there is some fungibility exactly how much it depend on the technology platform we are talking about Specifically, CIS, for example, that's -- the -- is fully close to the larger flow. So if you try to -- in fungible with some technology platform that's close to logic flow, then all of them can be used.

For example, we normally group larger products, CIS and driver type of product. It won't go because they are very much mutually fungible to a large degree. If you want to use CAS fr BCD type of product -- fungibility still -- there is still some fungibility that will be decreased or if we want to use -- for memory product, there are also some fungibility but you will probably further decreased somewhat because some of the memory products have some unique tool that requires. So if you -- a lot of times, your fungibility is limited by those unique to each technology platform.

When we build a fab, we try to manage the fungibility. You obviously can't be found 100% fungible. We try to maximize the fungibility so that we can react to market demand fluctuations among different technology platforms. So right now, we are certainly exercising that fungibility to a maximum extent. This, of course, is also limited by the fact that we still want to maintain a reasonable volume of -- for each product because we are into a long-term business. We don't want to just look at next quarter or even just 1 year, there is some level of strategic decision-making that's going on to make sure that we do have a long-term view now to be 100% driven by short-term considerations. That's one -- another 1 is when we do use pricing as a tool to kind of manage demand shift in demand between the technology from some platforms so that the demand pattern matches our capacity pattern better.

So yes, overall, I think when the overall demand is tight. And in general, we managed to increase prices across the board, some more than others. Thank you.

Next question will come from [indiscernible] of [Gossan Securities].

This is [indiscernible] First is about the demand of the consumer electronics. Sso the rising memory price may weigh on the demand of the consumer parts. But still, we can see Hua Hong achieved sequential growth in consumer parts. So how do you view the growth of our consumer-related part in the second half of the year? This is the first question.

That's actually a good question. Frankly, at beginning of the year, when everybody knows is hard to know that the AI related the product will have high demand. Another thing was discussed in the industry was the fact that when the DRAM getting to pricing, you're probably going to depress the consumer demand. which is probably true in the end market and some of the end market segments, as cellphone for example. Clearly, it's going to see a decline this year.

So we were expecting actually -- maybe demand decrease on the consumer segment. As I said, a broad from supporting all different market segments. And I think we were a little bit surprised that we actually didn't see as much negative impact, some of the consumer end market demand decreased. It could be that because we're now we are foundry, so our direct customers are design houses, [indiscernible] those IC product those goes into different segment of market maybe to because we were not directly providing to the end market. So maybe our direct customers are certainly a good job of managing.

So managing between different end markets. That's 1 possibility. And that's why we don't see much of a negative impact on the consumer end market. Another possibility is -- which is also possible, but even short-term consumer market, end market might be having bit of a decrease in demand they also want to -- they don't want their stock level to -- they don't want to -- they still under some inventory level to build on inventory that for probably inevitable upturn in the future.

So those are the 2 possibilities. So -- but the NAND result is that we do see strong demand with all the AI-related products. We haven't seen a -- haven't seen that much negative impact on the consumer market from -- at the foundry level. Third possibility could be our product, our technology, as I said earlier, we are pretty strong on -- it could be, if we're in the low end the market, which we are not were mostly met and is and high-end market, mainly the lower market like seeing a more negative impact.

Thank you. And my next question is about the progress of the acquisition. So could you update the progress and also the technology road map after the accurate acquisition.

The progress we already updated in the statement that we got the final approval from the exchange to proceed. So we expect the final -- the final step of this long acquisition process is going to take place probably within a month. After that, the whole thing is complete and done. The second part of your question, I probably didn't quite good. We do expect this to be a very positive acquisition for our financial statement. Once the final step gets completed, the Huali micro results will be included in our financial results. If there is nothing -- no surprises in Q3. Q3 statement will include the Huali micro.

Did I answer your question?

Yes. And also, I have a quick follow-up. So how about the technology road map after the acquisition? So any new yes, yes.

So in terms of what product Huali Micro is doing, there's quite of a synergy with what we have in Hua Hong Grace. So we do achieve quite a bit of savings, quite a bit of synergy in terms of technology sharing. In other words, some of the technology development that we do in Hua Hong Grace or in Huali Micro previously now can be combined. So we basically for any given R&D dollar, we get a bigger manufacturing scale.

So that's good for us. So in terms of the improved efficiency. Another thing it helps us is that now we have one more fab. So our manufacturing scale for a given technology platform that effectively are virtually bigger. So we can take on more customers who have a bigger capacity needs that we previously might struggle if we just have 2 separate entities, especially for Huali Micro, they are by themselves. They are not large. So that's another benefit.

The third benefit is that now Hui Micro joins the Hua Hong Grace, the overall manufacturing system that we can optimize the capacity structure, like what kind of a technology place where so that gives us a better ability to respond to market -- changing market demand, especially right now because in short supply that we immediately can start some of the technology platform that we cannot supply in Wuxi, for example, to put it in Huali Micro. So overall, you see it really because of the bigger scale R&D savings and overall improved efficiency because of the large scale and also on the procurement also we have a bigger volume. Everything is basically positive. So we think this is going to be a very, very -- it has been a very good step for us to take.

Next question will come from the line of Qingyuan Lin of Sanford C Bernstein.

Congratulations, Dr. Bai and Daniel for a good results for earnings. My question comes from 2 angles. First one is around the future capacity expansion. Dr. Bai, what's your view on the demand sustainability for '27, '28? You mentioned it was quite clear for the second half. But I was wondering, do we expect this cycle to be -- kind of continue to be stronger even for the next few years? And you mentioned that last year, there was a good decision to continue capacity expansion. With this strong demand, do we continue to see that we might need to further accelerate the capacity expansion even for '27 and '28. So that will kind of lead to, I guess, a question for Daniel, do we have any plan to further ramp up our CapEx? That's my first question.

Yes. Let me -- in terms of capacity expansion, we definitely want to continue the capacity expansion at a steady pace so that we can manage the CapEx expenditure while still we maintain profitability that we have come a long way to establish at this point. We do -- I do think -- I said earlier, the second half of the year, everybody the short-term market demand is strong. And 2027, the consensus is also going to be strong. 2028 is where since people start to have some debate. So it's probably a little bit too early to tell in 2028, but I do think overall, the secular trend is I do see a secular growth story in the marketplace that we participate, which is the specialty technology.

So we -- so in that sense, that's the reason that give us confidence that we will continue to expand capacity. So this overall demand increase is based on new application of the semiconductor and also based on the fact that some of the industry players may start to close down some of the 8-inch fabs. As you know, we do have 3 8-inch fabs that also benefit from some of the industry capacity going offline. 12-inch is a different story, but a lot of growth is mostly on 12-inch in terms of the demand side. So I think in terms of the supply side, 8-inch, nobody is spending 8-inch capacity, but the supply side may decrease, but 12-inch, the demand side is probably going to go -- continue to go up.

So it's really based on our strategy of steadily expanding our capacity is based on our confidence that the market, even with some fluctuation, the overall direction is still going up. Another -- our confidence is also based on the second factor, which is we believe our technology capability relative to our competitors in the industry is also going to strengthen as we go because we have the scale, we have the people, we have the track record. We have the position in China as well as even worldwide now that we think our capability will increase. So we are not afraid of even the downturn comes. I think we're still going to keep on growing and keep going -- getting more capacity steadily. Thank you.

Daniel, any comment on the CapEx guidance or kind of projection for next 2 years?

I would say we are -- we start to construct the third 12-inch fab early this year. And this fab will start to ramp over the next 3 years to 55,000 wafer capacity. So it's going to be -- it's about -- overall, it's approximately $6 billion CapEx spending. So I would say roughly $2 billion a year for the next 3 years. But other than that, unless we have other new fabs that we plan to build, this is what we -- this is going to be the major CapEx spending.

Just one comment that $60 billion all the CapEx. So it's roughly less than $2 billion per year, 1.5-ish over 3 years.

Over 3 years, yes.

Got it. Very clear. And my second question is around -- in the earnings, we do call out specifically that we have about $25 million of received for the government grants for equipment. May I have kind of a bit more details behind that? And last time when we called that out was fourth quarter '25, it's about $37 billion. And this is kind of related to my question around the plan for the 9B. What's the share of local equipment? Do we plan that to go up? And what kind of level we should expect?

Well, that was actually some subsidies grant we got not in Wuxi, but it was really for Shanghai, okay? That's the grant we received in Q2 from local government here. The WuXi part most likely will be paid, I think, in Q4 -- in Q4 2026.

The second part of your question about the domestic equipment. I think the domestic equipment sector in China has been getting strong year-over-year. We do expect as a general trend, the newer fabs will have higher percentage of the domestic equipment.

Our next question comes from Bintuo Ni from Daiwa Securities.

Congrats on the great execution. Can I ask your current lead time for products across different technology platforms and which segment is expanding and which segment is decreasing?

How long it takes to get the wafer from start to finish?

Exactly, yes, to deliver to your clients.

Okay. That obviously depends on the technology platform. Some process flow longer, some are short. like power -- the discrete power devices that doesn't have too many steps. You can get it in a couple of weeks if we accelerate it. Then some of the MCU products have 50, 60 or 30, 40 mass layers that will take 2 months if we accelerate it. The speed of the wafer moving through a fab is also a function of loading. If you have a very heavily loaded fab, basically you have a longer queue time in front of the equipment. So it tends to go -- so the average speed will be slower.

But we can -- we also -- usually in the fab, the way we manage it is that we have different tiers of different tier of the speed. If some of the things like some like MTO, first time you have a new product, we try to give a high priority, they can just to the fab very fast. But for the volume production, which you know that we tend to maximize the output versus speed. So we let that -- that's still -- nothing takes more than a quarter, I would say. The faster ones can be a month, 2 months really depending on the type of products you have.

Great. Is there any changes in lead time in terms of when we receive the order until we deliver the product, is there any changes in lead time? So I'm trying to understand isn't it…

No significant changes. When the demand gets tight, when demand is high and the supply gets tight, one impact -- the effect is to tends to make the delivery time a little bit longer. But this is something we work out with our customers. We will basically -- when they place the order, we usually have a commitment to say this will come out in certain this time if the customer agrees and works to their satisfaction, then we will just proceed. So that's how that works. But if there's something they need urgently, we can also support that. Not 100% of the time, but certain percentage of the wafer can come out really fast if we need to.

Understood. That's very clear. My next question is about our investment plans. I think Dr. Bai mentioned USD 1.5 billion per year CapEx. So what kind of technology platform will be focused more in the coming 2 to 3 years?

So this gets back to our focus. Our business focus is the specialty technology. So if you look at specialty technology, they are very much application driven. So we go where the market is, so to speak. So we look at -- the reason we have those 4 or 5 large technology platform is because there's a large demand -- market demand for those like BCB for PMIC, power management and power devices for all things electric and a lot of power-related MCU microcontroller nowadays a lot of AI-related applications require microcontroller or even auto, the new EVs, which has a lot of microcontroller in them.

CIS has been there since the cell phone become a large application that drives a lot of CIS because the image sensor. But now CIS is also driven by some security needs and even the auto, the new EVs, the autonomous driving car or robots for that matter, some of the emerging applications or drive a lot of that. So I mean there's no short answer to your question. But overall, we look at all the specialty technology we participate in. We see -- we look at where the demand is high and combined with where we have our strengths like MCU, we are very strong in MCU historically. That's also a growth area. So we're going to put a lot of -- for example, we're going to put a lot of capacity there.

And BCD is another area. So in general, if you look at our financial -- the results over the last couple of quarters, the growth -- the highest growth is really in the MCU, BCD, even NOR flash, those areas we're going to grow more. on the CIS logic, it is also an area that are very much interest to us, although the growth rate hasn't been as high as the other 2 technology platform or other 3, but we also try to drive up -- try to get a bigger share there. So for us, it might become a growth problem.

So those are the areas. It's really the capacity we put it in where the current technology platforms are. And each technology platform also over time, the technology also evolve and it goes -- MCU probably going to go from 55-nanometer to 40-nanometer MCUs, and we will follow -- we will watch for where the sweet spot is in the market and try to build our technology road maps to go where the market is and also to go where we think we have a competitive advantage. So that's the complex answer I give to you. So it's a bit of a complex because by nature is complex. That's a lot of -- we spend a lot of time on those things so that we make sure we get it right so that we can have a good growth.

The next question comes from Tracy Cui of CLSA.

So my question is regarding the depreciation, Fab 9A ramp-up and also new fab coming. Wondering how much may be the depreciation in the second half and also in next year?

Good question, Tracy. So look at the second half overall, the 8-inch business will be around $55 million, okay? I understand you probably it could be useful for your model. And for the -- for our first 12-inch fab, it is going to be around $250 million depreciation expense for the second half of 2026. And for second fab, it is going to be at the roughly $210 million in the second half. These are the forecast numbers. And for the third fab, we're just starting, there's not -- virtually is not going to be any it's not going to be any. If there's anything will be minimal for this year. And then for Huali Microelectronics, okay, we're projecting about $30 million for the second half, okay? There -- for that fab, the depreciation expense is pretty much behind them. We're looking at around $50 million to $60 million a year, and it's going to start to decline even further down in the next few years.

Got it. And my next question is regarding the new business. I think last earnings call, you talked about expanding to like interconnect solutions such as silicon photonics, interposer. So wondering if any like a quick update progress in any of those business, any specific area that you see stronger growth potential?

Okay. Thank you for the question. I have to be careful what I say here. We are probably the largest specialty foundry in China. Even we are the second largest foundry in China. But in the specialty technology, as I said earlier in answering earlier questions, we will go where the market goes. So in that regard, of course, that is one factor. Another factor is we also go where we think we have strength or have advantage. So combining those 2, that's the determine where we go. The part you mentioned, there's -- we look at AI is definitely a growth driver. Therefore, anything that's related with the AI application that happens to be in the specialty technology area, we want to -- we look at very carefully and decide whether if we are not already there. We are already in all of those areas, the PMIC, MCU, I mentioned, the power devices, all those things.

And there's a few like [indiscernible] electronics silicon photonics, for example, we actually -- we are already in there that some of the MCU do go into the module that make the final silicon photonic. We like to expand the footprint there to probably get into more type of silicon-based devices. So it's going to be -- because that's where we have expertise. So yes, we are going to basically look at the application there, look at some of the chips inside the silicon photonic module to see we already participate in some of those chips in there. We want to expand a couple more in terms of the IC silicon-based IC. That's still the plan, and we are seeing the early stages of doing that.

The power devices, that's another area that is probably going through some technological market transition that up to this point, is mostly silicon-based. But now the silicon carbide, for example, these devices are also become part of power device offering. Again, there, the product is if you look at a module where it tends to contain silicon-based devices and increasingly silicon carbide. So this is something if we want to continue in the power device area, we will have to look at. We have no choice, but look at see whether we have a more complete offering. So we are definitely doing the planning and a lot of the planning. We do need to get a few things together before we say that we officially are in there. But we're not slowing down, we're going forward.

With that, I'll now take the last question from [ Hui Jin ] of [ Orren ] Securities.

This is [indiscernible] My first question is about our revenue from North America and Europe grew very strongly. So could you give us more color about these 2 regions, maybe it's from like the server PMIC or maybe MCU or this kind of products? That's my first question.

The revenue from North America, a large part of it is in the CCD PMIC area. and that a lot of the product there happens to be related to AI server boxes. That's the reason you see huge increases because that's directly going to AI. For Europe, it's really the large European company have this China for China strategy. Their product is mostly in MCU and the smart cards and some power devices as well. I think as they prosecute their China for China strategy, if we are their partner of choice, which we think we are in China, that's the reason we see growth. We expect to continue to see more growth from Europe as well.

My second question is about the equipment and material. People are always talking about the equipment supply are very tight and the materials are -- the price for materials are increasing, all these kind of things. So from our point, how do we see the supply of equipment and materials?

Good question. the equipment supply is getting tighter because overall -- all over the world, the semiconductor houses are increasing their capacity. So this is true for our overseas suppliers as well as domestic suppliers. So the one manifestation is the lead time has been increasing. But it's still manageable. It's still manageable level, and we -- since we -- for us, we started this capacity expansion last year. So a lot of the equipment we already booked last year. So we haven't seen huge increase -- huge impact, but they are definitely getting tighter.

So we are tightly managing with our suppliers, with our partners, supplier partners to get the equipment lead time to a point that doesn't affect our overall capacity growth increase schedule. And so far, I think we, by and large, can do that. For material, there's some isolated, there's a few example, because of the -- for example, because of the war in Middle East or some other reasons that we do see tightness and even price increases on some of the, for example, helium gases because of the war had a temporary spike, but now it has managed to come down.

And overall, we don't see -- there are some metals also metal prices because of general inflation or because of the supply situation. So we do see some price increases in some isolated area. But overall hasn't been a significant impact. We still managed to basically keep the price flat or down. And we -- because the overall marketplace in terms of our supplier market is still reasonably healthy for us. Thank you.

Thank you, ladies and gentlemen. That's all the time we have for questions. I'll now hand back to Mr. Daniel Wang for closing remarks.

This concludes our today's call. Once again, thank you all for joining us today and for your valuable questions and input. It has been an exciting quarter. We look forward to see you again in the next earnings call. Thank you.

Thank you.

Ladies and gentlemen, thank you for your attendance. You may all now disconnect.

Ladies and gentlemen, thank you for standing by. Welcome to the Hua Hong Semiconductor's First Quarter 2026 Earnings Conference Call. Today's call is hosted by Dr. Peng Bai, Chairman and President; and Mr. Daniel Wang, Executive Vice President and Chief Financial Officer.

[Operator Instructions] The earnings press release and first quarter 2026 summary slides are available to download at our company's website, www.huahonggrace.com. Without further ado, I would like to introduce you to Mr. Daniel Wang, Executive Vice President and Chief Financial Officer. Thank you.

Good afternoon, everyone. Thank you for joining our Q1 2026 earnings conference. Today we will first have Dr. Peng Bai, our Chairman and President, provide an overview of our first quarter performance. I will then take you through our financial results in detail and offer guidance for the upcoming quarter. We'll then open the floor for a question-and-answer session.

With that, I turn the call over to Dr. Bai. Dr. Bai?

Thank you, Daniel. Good afternoon, everyone. Thank you for joining our earnings call. Hua Hong Semiconductor generated revenue of USD 660.9 million in the first quarter of 2026, a year-on-year increase of 22.2%. Gross margin stood at 13%, a year-on-year increase of 3.8 percentage points. Both figures were in line with our guidance.

Net profit attributable to shareholders of the parent company amounted to USD 20.9 million, marking substantial year-on-year growth. Despite the rapid capacity ramp-up, the company maintained high capacity utilization rates, with strong performance across all process technology platforms. MCU, standalone flash and BCD products delivered the highest growth rates.

The company's results were supported by sustained efforts in cost reduction and efficiency enhancement, and by a positive demand signal that started at the beginning of the quarter and became stronger over the course of the quarter.

The global semiconductor industry is undergoing accelerated transformation as AI and related applications play an increasingly central role in market dynamics. The unmistakenly positive impact of AI on worldwide semiconductor market demand, persistent uncertainty in the global supply chain landscape make a more complex market picture that we're facing with.

Hua Hong Semiconductor remains steadfast in pursuit of its key objective of becoming a foundry leader in specialty process technologies, with a core strategy of continuous focusing on market needs, strengthening process technology capabilities and substantially increasing production capacity scale.

In the first quarter, the ramp-up of our 12-inch capacity progressed steadily, with its revenue contribution rising to 62.7%, while our 8-inch production line maintained sound profitability.

Meanwhile, the proposed acquisition of Huali Micro has been accepted by the Shanghai Stock Exchange, entering the substantive review phase. The acquisition is currently progressing according to the established schedule and is expected to be completed in the second half of the year.

Finally, as a veteran of the industry, I remain confident in a bright future for semiconductor industry globally and in China. I'm fully committed to make Hua Hong Semiconductor an increasingly important player in the industry and deliver sustainable value to our shareholders.

Now I would like to hand the call over to our CFO, Mr. Daniel Wang, for his comments. Daniel?

Thank you, Dr. Bai, for your inspiring comments. Now let me walk you through a summary of our financial performance for the first quarter. I will then provide our revenue and margin outlook for Q2 2026 before opening the floor for the Q&A session.

First, let us review our financial results for the first quarter. Revenue was $660.9 million, 22.2% over Q1 2025, primarily driven by increased wafer shipments and improved average selling price, and 0.2% over Q4 2025. Gross margin was 13%, 3.8 percentage points over Q1 2025, primarily driven by improved average selling price and cost reduction efforts, and flat with Q4 2025.

Operating expenses were $105.6 million, 8.8% over Q1 2025, primarily due to increased operating expenses for the new production lines in Wuxi and 18.9% lower than Q4 2025, mainly due to decreased labor costs.

Other loss net was $2.4 million, 70.5% lower than Q1 2025, primarily due to foreign exchange gains versus foreign exchange losses in Q1 2025, partially offset by decreased government subsidies, interest income and increased finance costs. It was other income net of $34.1 million in Q4 2025, mainly due to increased finance costs and decreased government subsidies.

Income tax credit was $4.7 million, primarily due to a reversal of dividend withholding tax for 2025. Net loss for the period was $17.3 million, narrowed by 66.9% compared to Q1 2025 and 7.5% compared to Q4 2025.

Net profit attributable to shareholders of the parent company was $20.9 million, 458.1% over Q1 2025 and 19.9% above Q4 2025. Basic earnings per share was $0.012. Annualized ROE was 1.2%.

Now let's take a closer look at our Q1 2026 revenue performance. From geographical perspective, revenue from China was $525.2 million, contributing 79.5% of total revenue, and an increase of 18.7% over Q1 2025, mainly driven by increased demand for MCU, other power management IC, flash and IGBT products.

Revenue from North America was $85.7 million, an increase of 51.9% over Q1 2025, mainly driven by increased demand for other power management IC and MCU products. Revenue from Other Asia was $28.2 million, an increase of 5.2% over Q1 2025, mainly driven by increased demand for MCU products, partially offset by decreased demand for super junction products.

Revenue from Europe was $21.8 million, an increase of 43.2% over Q1 2025, mainly driven by increased demand for smart card IC, IGBT and MCU products.

With respect to technology platforms, revenue from embedded non-volatile memory was $184.6 million, an increase of 41.7% over Q1 2025, mainly driven by increased demand for MCU and smart card ICs.

Revenue from standalone non-volatile memory was $57.1 million, an increase of 33.2% over Q1 2025, mainly driven by increased demand for flash products. Revenue from power discrete was $170.9 million, an increase of 5% over Q1 2025, mainly driven by increased demand for IGBT and general MOSFET products, partially offset by decreased demand for super junction products.

Revenue from logic & RF was $74.4 million, an increase of 11.4% over Q1 2025, mainly driven by increased demand for logic and CIS products. Revenue from analog & power management IC was $173.9 million, an increase of 25.8% over Q1 2025 mainly driven by increased demand for other power management IC products.

Now let's turn to our cash flow statement. Net cash flows generated from operating activities was $130.4 million, 159.9% over Q1 2025 mainly due to increased receipts from customers. It was 47% lower than Q4 2025, largely due to decreased receipts of government grants and increased payments of labor costs.

Capital expenditures were $924.9 million in Q1 2026, including $886.1 million for Hua Hong 12-inch business and $38.7 million for Hua Hong 8-inch. Other cash flow generated from investing activities was $67.9 million in Q1 2026, including $57.8 million decreased time deposits, $10 million interest income and $100,000 receipt of disposal of the equipment.

Net cash flows generated from financing activities were $638.7 million, including $649.4 million proceeds from bank borrowings and $3.3 million proceeds from share option exercises, partially offset by $12.1 million interest payments and $1.1 million lease payments and $0.8 million of bank principal repayments.

Now let's move to the balance sheet. Cash and the cash equivalents was $4.8679 billion on March 31, 2026 compared to $4.8938 billion on December 31, 2025. Other current assets increased from $787 million on December 31, 2025, to $894.6 million on March 31, 2026, mainly due to increased value-added tax credit.

Property, plant and equipment was $7.1059 billion on March 31, 2026, compared to $6.6764 billion on December 31, 2025, primarily due to capacity expansion in Hua Hong Manufacturing. Interest-bearing bank borrowings increased from $3.1908 billion on December 31, 2025, to $3.8972 billion on March 31, 2026, primarily due to increased drawdowns of bank borrowings.

Total assets increased from $14.4538 billion on December 31, 2025, to $14.9473 billion on March 31, 2026. Total liabilities increased to $5.663 billion on March 31, 2026, from $5.2895 billion on December 31, 2025. Debt ratio increased to 37.9% on March 31, 2026, from 36.6% on December 31, 2025.

Finally, let's take a look at our second quarter outlook for 2026. We expect revenue to be in the range of $690 million to $700 million with a projected gross margin of 14% to 16%. This concludes my financial remarks.

We'll now begin the Q&A session. Operator, please assist. Thank you. Operator?

[Operator Instructions] Our first question comes from the line of Leping Huang of Huatai Securities.

So congratulations for the very strong results. My first question is about the impact of the memory super cycle. So we see the further price hike of the memory in the last 3 months. So Dr. Bai, what's your view on the ripple effect of this trend on the logic foundry investment, especially on your NOR flash business? Can we expect further price hike in memory will drive the ASP and the margin expansion in your logic foundry business in the remaining of this year?

Okay. Thank you, Leping. Good to hear from you. Yes, let me take your question on the NOR flash situation. You are right. First of all, the memory is in short supply, starting with DRAM, and DRAM price, as you know, went up probably like 10x. Then that start -- then they start a spill over into NAND memory. Now it start to spill over a little bit into NOR flash as well. It's not as much spill over as I would like, but definitely, we see the demand going up for our NOR flash memory.

That will translate into a price increase, but not as significant as the DRAM, not nearly as significant as the DRAM, but we should expect to be 10%, 15% price increases on the NOR flash. So that would be the -- also with the tight supply, we won't be able to satisfy all the demand, but we do get the benefit of higher prices.

Okay. So my second question is that can you provide some update on the progress of your Wuxi fab expansion. So I remember you mentioned last time I think you start already construction in March and moving in the October. But these days, we see a very fast expansion of most of the foundry and the memory company in the world.

Do you see any delay of the equipment delivery? So have you -- or furthermore, today with the U.S. President visiting China, which is a very important milestone for the U.S.-China relation. So do you think that this will help your equipment procurement coming in the rest of the year?

Okay. Yes, let me update you on our Wuxi side in terms of the capacity expansion. As you know, Wuxi side, we have 2 operating fabs already. The third one, as you correctly pointed out, started construction in March.

So before I go to the third fab, which we call 9B, the second fab has been on a capacity ramp-up since last year, it is Fab9A. That capacity ramp-up should complete in Q3 this year, which means that we will reach the full capacity in Q3. The output, there's obviously a delay in getting all the output with the full capacity, but we should be getting the full output starting probably by end of this year or beginning of next year.

Now the third fab, Fab9B, we kicked off the construction in March of this year. We are doing all the facility work right now because the shell is already there for the third fab. We expect the equipment start to come in, in the fourth quarter of this year.

To your question about the equipment procurement, we do not see any impact in terms of the export control, from the U.S. export control on the equipment procurement for Fab9B. That has not been an issue for us. Now President Trump is in China, as we speak. And we certainly hope this environment is going to be further relaxed going forward, but we'll see how they come out.

But up to this point, even all the press reports notwithstanding, all the press reports notwithstanding over the last couple of months, we have not been impacted in terms of getting the equipment we need to buy and also the delivery time.

Okay. It's good to see. So final question is, so we see the Hua Hong Group established an advanced packaging subsidiary. Also, the market is very closely watching the next-generation technology like silicon photonics. So my question is that how the Hua Hong Group's advanced packaging subsidiary will coordinate with the Hua Hong ListCo? And how do you -- and what's the technology you have on the silicon photonics? And what's your plan on silicon photonics?

Okay. So let me take your question one at a time. One is the advanced packaging. Yes, there is a newly established effort on advanced packaging. That is not -- it's under the Hua Hong Group, not part of the public company, not part of Hua Hong Semiconductor.

Of course, we are having a lot of coordination at the Hua Hong Group level in terms of the technology road map and as well as capacity planning. So in that regard, certainly are coordinated with the existing Hua Hong Group companies.

The advanced packaging is basically -- is definitely a high growth area in the industry. We expect that effort to substantially add to our offering and also expect to add to our revenue in the future at a group level.

Now the silicon photonics is another area that is growing pretty fast, especially with AI and related application. We are also looking at getting into silicon photonics area, and there's quite a bit of planning activity that's ongoing. And once something gets finalized, I expect that will be soon that we will update you on the details.

But suffice to say that, yes, we will get into silicon photonics since this is a growth area, adjacent to what we already do, so it should have a lot of good synergy with what we already have, and we should expect that to be started pretty soon, right?

And the next question comes from the line of Tracy Cui from CLSA Company.

So my first question is regarding your 2Q '26 revenue guidance. Would you please guide us how much growth may come from the shipment growth and how much is from the ASP increase?

It's roughly -- it's about -- we're talking about quarter-to-quarter around close to 5%, okay? And we're pretty confident with that number. It's coming in combination of revenue -- ASP growth and also volume growth.

Okay. And just following up for the wafer price. Perhaps how you will expect your wafer price trend throughout this year? And which type of chips you may see maybe a higher chance to further for the price lift?

Since we have a diverse portfolio of technology platform, we do see the price changes, price increases vary across different platforms. The ones that are in higher demand, we will see higher price increases. The ones that demand are somewhat muted, we will probably trying to stay where it is or just with a very minor increases.

The net result is that, on average, I do see -- I do think we're going to be consistent with the industry average of about 10% plus/minus some number here and there. You have to realize the price increases, it takes time to get implemented in a sense that even if you increase the prices is the order from that day forward see the new prices so that it will take some time to fully materialize all the increases to have them fully reflected in the financial results.

But during the course of this year, I think we definitely will be seeing above -- on average. Some will increase more, like some of the -- we do this mainly based on market. It's a market-driven methodology. And if you see lot of demand, we can supply, we do tend to increase the prices a little bit more.

So some platform might see up to 20%, 25% even, but some are going to be close to where they are and maybe 5%. On average, I think 10% is a reasonable assumption or 10% to 15% by end of this year, somewhere there.

Got it. And can I please also have a follow-up question on Fab9B. Would you mind remind us how much is our CapEx plan and how much is the design capacity? And any updated time line in terms of the CapEx investment and also the ramp-up and what may be the process node and the type of chips platform?

I think the Fab9B, the overall project, I think we're going to be putting down CapEx-wise about $60 billion, maybe right around there. It will be 60 -- we actually originally said $69 billion. We will try to -- $6 billion, not $60 billion. $6 billion, okay? Take off one 0 there. So originally, we think it's 6-point something, but we will try to control it to be about $6 billion overall investment.

That will be spent this year and next pretty much. And it will start to -- equipment started coming towards the end of this year. So we are talking about it will start to produce output in 2027. Probably going to take 1.5 years to 2 years to get all the capacity in place, to get -- so you are talking about in 2028, the Fab9B will be at a full capacity. So that's the rough schedule I'm giving you.

And what might be the process nodes or type of chips? Any color that you may be able to share?

It will be focused on specialty technologies. We see that it will be a range of nodes, but the bulk will be 40-nanometer. So 40-nanometer plus/minus something.

And the next question comes from the line of Ziyuan Wang from Citic.

This is Ziyuan Wang from Citic Securities. My first question is about, I'm wondering how was the demand trend in analog and power applications over the past quarters. And is there any further plan for the wafer price to increase on Q2? And have there been any price increase in the materials such as silicon wafers?

The analog area, our main product is power management ICs. We do see strong demand there. That's mostly related to AI and AI-related build-out, like the server boxes, they need a lot of power management chips. And that's some partially offset by some of the consumer weakness because of the DRAM increases.

We do see a memory price increases has a small depressing effect on consumer segment. Overall, it's a positive story for us because we obviously have products in both AI-related field as well as consumer. On average, AI-related stronger demand overwhelms -- at least more than compensate for the small weaknesses in consumer.

So in that regard, I think the analog, the PMIC area is a good example where you have one part of the market going strong, another part is a little bit muted, but overall, still a positive demand increasing story for us. So yes, we were -- you will see price increases in this particular platform, reasonable price increases here.

In terms of the supply chain, because of the -- probably mostly because of the war and some of the disruption that it causes, we do see some particular material, raw material, the prices go up. It's not across the board. Across the board, it's not a very significant factor yet, but there's a few items that prices to go up quite high.

And we're managing that and those are manageable, but we don't expect -- I do not expect that it will have a big overall impact. But yes, because of the -- mostly because of the war in the Middle East that is causing some supply chain disruption, in particular, the oil-related or gas oil-related products and some of the items, the prices spiked a little bit, but very few items. So it doesn't have much of a -- and in aggregate, it doesn't have average impact much. All right. Thank you.

Okay. Got it. And my second question is about, does our company have any plan or investments on the compound semiconductor such as the gallium arsenide or silicon carbide or even on indium phosphide. That is my second question.

Thank you for asking. The answer is yes. We are -- we have started efforts in gallium nitride. We already have TD activity going on. So we will get into gallium nitride. And as well as silicon carbide, the compound semiconductor, those 2 that complement our silicon-based power devices that we have decided to get into.

The way we'll get into might be a little bit different. Like silicon carbide, we might have a joint venture with some existing -- joint venture in terms of the capital management. But in gallium nitride, we might seek a partner and when we get into the volume manufacturing phase. So that part is still being kind of a planned or is not finalized.

But the answer to your question, yes, we will get into gallium nitride. We are getting into gallium nitride and silicon carbide effort because we have a large silicon-based power device capacity. We also have a large base of customer, and they are also asking us to get into the compound, so that -- compound semiconductor to complement our existing silicon-based power devices.

Indium phosphide -- we don't hear anybody speaking online.

Okay, that's all my questions. Very clear.

[Operator Instructions] Our next question comes from the line of [ Timothy Wong ] from [ Oriental Asset Management ].

Analyst from Oriental Asset Management. My name is Timothy Wong. My first question would be, I would like to confirm whether we are acquiring 7-nanometers or below to our portfolio since -- through the acquisition of Huali Micro as some news reports suggest? This is my first question.

Sorry, we didn't get the question.

Well, can you repeat that question again or you have not?

Yes. Okay. I would like to confirm whether we are acquiring 7-nanometer or below to our portfolio through the acquisition of Huali Micro as some reports suggested.

A couple of things. One is our Q1 results does not include Huali Micro. That's still outside the Hua Hong Semiconductor because it has now completed the acquisition. Second, Huali Micro we're acquiring -- what are we acquiring from -- for Huali Micro is basically what we call Fab5 asset and the Fab5 asset business. Those are based on 55-nanometer and 40-nanometer IC products, that's what we're acquiring.

Yes. I actually didn't believe such reports. So thank you for your clarification. And my second part of my question is would you mind giving us some color on explosion of CPO, silicon photonics optics, how would these trends will benefit us?

I think right now, in a lot of -- from a technical standpoint of view, a lot of high compute platform, the interconnect become a bottleneck. Therefore, people are looking for ways to speed up the interconnect. Silicon photonics is one way. Now there are some detail there that depending on whether you're talking about the stack to stack, like a box-to-box communication or some chip-to-chip, so you actually require a slightly different silicon photonics technology.

But that's an area that would require silicon photonics type of products. This is consistent with now AI is driving a lot of the market growth. So if you believe in thesis that AI is going to drive a lot of the growth, then you would believe that -- you would say that there is going to be more demand for silicon photonics related products because AI, obviously, at a system level is a high compute.

So that's the -- and from Hua Hong standpoint of view, we -- our technology -- although silicon photonics is a new technology for us, but it's not -- it's also adjacent technology for us. It's not too different, too far away from what we're already doing. We're already doing a lot of CIS, which require VCD -- not VCD, CIS require a lot of back-side processing of getting into advanced packaging area.

So there is quite a bit of synergy from a technology standpoint of view. So that's why we think it's a good area for us to get into because there's going to be market demand, and it also play into kind of our strength or our adjacent area of growth. Thank you.

Congratulations for all the progress.

Thank you.

Next question comes from the line of [ Jen Kuai ] from Orion Securities.

My first question is about the memory price. How do you see the memory price in the next few quarters?

Well, the price is a sensitive topic. In the end, the price will be set by supply-demand balance. As long as we continue to have short supply, then we have a little bit of opportunity to inch up the prices. We actually will stop price increases when the supply/demand balance is reached. So at this point, starting this year, we started to see tightness of the supply, and we continue to see that.

So we are going through one round of price increases. We have gone through one round of price increases. That's not fully reflected in -- it takes a little bit of time to see how it's impacting the order and how is it impacting the overall market, but we will be watching that and then decide whether the balance is reached or if it's not reached, we may inch up a little bit more and be frank and open about this one.

But if the demand is -- supply/demand is more or less at a good balance, we're probably going to pause or going to stop increasing prices. As I said, for the year, we expect, on average, as I said earlier, 10%, maybe if we're lucky, it will be 15%, 10% to 15%. And memory is one of the platform that increases, probably it will be above the average.

Okay. My second question is about the AI server opportunity. As we know, we already are making some like analog for some -- for AI server, and also we can see power discrete, the usage are also increasing in AI server. So from our point, how do we see the opportunity for us? Maybe just now you also mentioned like silicon photonics or maybe combine all this together, how do we see the opportunity?

Okay. The power devices, the overall market actually, you're correct, is actually increasing because of all the AI boxes need power devices, plus some of the new like robotics or the industrial demand is also increasing the demand for power devices. So the issue there probably is not so much our overall market demand. We do see that part as reasonably healthy.

The power devices in terms of -- the issue is probably the supply also increases quite fast, especially the compound semiconductor, like I said earlier, silicon carbide start to contribute a significant portion to power devices. So that here, you have a situation where the supply-demand balance is not off.

Therefore, for us to win business, we have to be -- it's a competitive market. We have to compete with our competitors. Now we do have a good silicon-based power device technology. We have led in that area. I think the part that we need to do more is to get the compound semiconductor as part of offering.

That's why earlier I said we are getting into silicon carbide and gallium nitride, the compound semiconductor, so that we can first provide a more complete solution to our customers and also give our customers more choices, more flexibility in terms of coming up with an optimal combination of the power devices to support increasing demand.

So that's the situation I see. Silicon photonics is a different story. That is -- we are starting from scratch. That is a nascent, a new development that is getting more and more important. I think that we don't have a silicon photonics offering right now. So for us, anything we get there will be a net addition to our business.

Next question comes from the line of Charlie Chan from Morgan Stanley.

So I actually have 2 questions. One is about your PMIC capacity expansion. I'm not sure if it's the right understanding, but could it be like a BCD process? Do you think you can expand that capacity given demand is so strong and also customers are willing to pay higher price? And if they cannot get the capacity from your fab, where they can go to get sufficient supply?

Yes. The power management product does use BCD technology, you're correct, that's the technology platform we're talking about, BCD. So that's number one. Number two, we are expanding the BCD capacity. That is already one of the highest capacity platform we already have, but we are expanding as we speak in both Wuxi as well as maybe in Huali Micro that we are about to acquire. So we will get higher capacity because we're now meeting the customer needs.

Third, you are correct that BCD area is a competitive marketplace. There's many players there, but we have been one of the biggest players in China and partly because we have a good -- we think we have a good technology advantage. And we also have some very valuable strategic customers that work with us. So that -- so we think we can increase the capacity and still be able to get good price on them.

So that's why we are doing capacity expansion. And this is also an area that we are focusing on in terms of technology development to try to get to next generation faster so that we keep that technology edge for us.

And my second part of this question is, so I also cover some Taiwanese foundry. For example, UMC, Vanguard, if I may quote, those are your industry peers. But it seems like their first quarter, second quarter wafer shipment sort of outgrow. I think 1Q they grow single digit. 2Q, like high single-digit Q-on-Q.

So I'm not sure after your capacity expansion, do you feel like you can win back some customers, maybe the consumer side or those non-China customers, maybe the U.S. customers or Taiwan customers, back to your fab for PMIC production?

Our PMIC business has been -- has grown in Q1. If we had more capacity, we probably could have done better. So as our new capacity come online, we do think that we can keep growing that part of the business. We already have customers from overseas, from U.S. and Europe for that matter because for European customer, they have this -- in China for China strategy that we do benefit from.

And the U.S. company, mostly because of their business, is growing very fast. So they need more supply. So in that sense, that we think BCD area or PMIC area, it is one area that is benefiting from the AI and AI-related growth. And also all the other -- like the car also need the PMIC and robotics overall.

The only -- from an end market standpoint, the only place we see some weaknesses is in the consumer segment, that we do see that. So the hope is that the area that's been growing, keep on growing like AI and robotics and the cars, auto and the consumer maybe will come back because they can't keep delaying those -- they can't keep on pushing out the new model forever. So they have to also get some of the growth or stop the decrease a little bit. So that's the overall picture I see.

Great. And my second question is some clarification on your comments. So in your prepared remarks, you sort of talk about, hopefully, some relaxation of the export control on equipment. But I think your Fab9, right, as you just described is more like 40-nanometer. If that is the case, why there's kind of export control? Or you were referring to your -- not your, but Huali's Fab8. So can you clarify your previous comment about the equipment restriction?

I was not talking about Huali. I was talking about just Hua Hong. You're correct. I also said we have not been impacted in terms of the -- by the export control. in terms of the equipment, we really need to buy. When I say if the restriction is more relaxed, in general, we will get more choices. It's always good to have more choice so that I might be able to have a little bit more optimal combination of the tools. It's not an issue if we can...

I see. So it's kind of nice...

Nice, yes, yes.

Okay. So nice to have. So even if it stays as core, it doesn't change your expansion for Fab9. Is that right way to think about this?

Yes. That's right.

And lastly, I think one of the previous callers asked about some potential new business, right? So I think for global AI semi supply chain, I think there are lots of components, for example, interposer, bridge die, silicon capacitors, there's VIC. So do you have any demand for those? I think the previous question was about CPO related, but I wanted to ask about all those kind of CoWoS or 2.5D packaging related components, no matter silicon capacitor, interposer, bridge die or VIC. Do you have any demand for those components?

Good question. We do see demand for high-density capacitor. So we are having an effort there to do the capacitors, changed or otherwise. So those high-density, high density capacitors. So that it looks like it's been there for a while now that we should start to see some revenue coming.

In terms of the other one, can you ask the question -- interposer. Interposer is also something that we are very open to it. I think that we are exploring some options there. That's still in the early stages. That's probably going to go with when we have our advanced packaging coming online. The interposer will also become a significant part of the overall packaging efforts.

Ladies and gentlemen, that's all the time we have for questions. I will now hand back to Mr. Daniel Wang for closing remarks.

Well, thank you very much for joining us today. I mean, you guys had a lot of wonderful questions. All these questions were -- has been very, very helpful, and we look forward to speaking with you again and perhaps seeing you in the next quarter. Thank you very much.

Thank you.

This concludes today's conference call. Thank you for participating. You may now all disconnect. Have a nice day.

Ladies and gentlemen, thank you for standing by. Welcome to Hua Hong Semiconductor Fourth Quarter 2025 Earnings Conference Call. Today's call is hosted by Dr. Peng Bai, Chairman and President; and Mr. Daniel Wang, Executive Vice President and Chief Financial Officer.

[Operator Instructions] The earnings press release and fourth quarter 2025 summary slides are available to download at our company's website, www.huahonggrace.com.

Without further ado, I would like to introduce you to Mr. Daniel Wang, Executive Vice President and Chief Financial Officer. Thank you. Please go ahead.

Good afternoon, everyone. Thank you for joining our Q4 2025 earnings conference. Today, we will first have Dr. Peng Bai, our Chairman and President, provide an overview of our fourth quarter and full year performance. I'll then take you through our financial results in detail and offer guidance for the upcoming quarter. We'll then open the floor for a Q&A session.

With that, I turn the call over to Dr. Bai.

Thank you, Daniel. Good afternoon, everyone. Thank you for joining our earnings call. Fourth quarter 2025 sale revenue for Hua Hong Semiconductor reached an all-time high of USD 659.9 million with a gross margin of 13% for the quarter, both in line with our guidance. For the full year of 2025, the company reported sales revenue of USD 2.4 billion and a gross margin of 11.8%, both achieving year-on-year growth and meeting management expectations.

Against the backdrop of the global semiconductor market being driven by demand for AI and related products, coupled with the recovery in consumer demand led by the domestic market, the company maintained full capacity operations throughout the year at an average capacity utilization rate of 106% which ranked among the leading levels in the foundry industry.

By optimizing product mix, reducing costs, and improving operational efficiency, we achieved strong performance across various specialty technology platforms, especially in stand-alone NVM and the power management area, effectively supporting the company's revenue growth and margin expansion.

In 2025, the company continued to advance its strategic plan for capacity expansion. The first phase of capacity construction for the second 12-inch production line in Wuxi, we call Fab9, exceeded expectations for completion. And the Shanghai 12-inch manufacturing base, Fab5 acquisition progressed as planned.

Looking ahead, the company will maintain a strong focus on developing world-class specialty technology platforms with innovation and through rapid generational iteration while deepening collaborations with strategic customers, both domestically and internationally.

We remain confident in our ability to seize growth opportunities amid changes in the global semiconductor industry and are striving to meet shareholders' long-term expectations.

Now I would like to hand the call over to our CFO, Mr. Daniel Wang, for his comments.

Thank you, Dr. Bai, for your inspiring comments. Now let me walk you through a summary of our financial performance for the fourth quarter, followed by a recap of our full year 2025 results. I then provide our revenue and margin outlook for Q1 2026 before opening the floor for the Q&A session.

Now first, let us review our financial results for the fourth quarter. Revenue reached another all-time high of $659.9 million, 22.4% over Q4 2024 and -- Q4 2025 (sic) [ Q4 2024 ] and 3.9% over Q3 2025, primarily driven by increased wafer shipments and improved average selling price.

Gross margin was 13%, 1.6 percentage points over Q4 2024, primarily driven by improved average selling price and cost reduction efforts and a 0.5 percentage point dip from Q3 2025, primarily due to increased labor costs.

Operating expenses were $130.2 million, 17.7% over Q4 2024, primarily due to increased labor costs and the depreciation expenses and 29.6% over Q3 2025, mainly due to increased labor costs.

Other income net was $34.1 million compared to other loss net of USD 40.5 million in Q4 2024, primarily due to foreign exchange gains versus foreign exchange losses in Q4 2024, decreased finance costs and increased government subsidies. It was 92.1% over Q3 2025, mainly due to decreased finance costs.

Income tax expense was $8.1 million, 22.3% higher than Q4 2024, primarily due to increased taxable income. Net loss for the period was $18.7 million, narrowed by 80.6% compared to Q4 2024 and the widening of 159.9% (sic) [ 159.5% ] in loss from Q3 2025.

Net profit attributable to shareholders of the parent company was $17.5 million compared to a loss of $25.2 million in Q4 2024, and a profit of $25.7 million in Q3 2025. Basic earnings per share was $0.01. Annualized ROE was 1.2%.

Now let's take a closer look at our Q4 2025 revenue performance. From geographical perspective, revenue from China was $539.3 million, contributing 81.8% of total revenue, and an increase of 19.6% compared to Q4 2024, mainly driven by increased demand for power management IC, MCU, flash and CIS products.

Revenue from North America was $72.8 million, an increase of 51.3% compared to Q2 (sic) [ Q4 ] 2024, mainly driven by increased demand for power management IC and MCU products. Revenue from other -- Asia was $28.4 million, an increase of 9.1% compared to Q4 2024. Revenue from Europe was $19.3 billion (sic) [ $19.3 million ], an increase of 35.6% compared to Q4 2024, mainly driven by increased demand for MCU and IGBT products.

With respect to technology platforms, revenue from embedded non-volatile memory was $180.2 million, an increase of 31.3% compared to Q4 2024 mainly driven by increased demand for MCU and the smart card ICs. Revenue from stand-alone non-volatile memory was $56.6 million, an increase of 22.9% compared to Q4 2024 mainly driven by increased demand for flash products. Revenue from power discrete was $168.9 million, an increase of 2.4% compared to Q4 2024, mainly driven by increased demand for general MOSFET products.

Revenue from logic and RF was $80.4 million, an increase of 19.2% over Q4 2024, mainly driven by increased demand for CIS products. Revenue from analog and power management IC was $173.8 million, an increase of 40.7% over Q4 2024, mainly driven by increased demand for other power management IC products.

Now turning to our cash flow statement. Net cash flows generated from operating activities was $246 million, 29.5% lower than Q4 2024 mainly due to increased payment for suppliers and the increased receipts of government subsidies, partially offset by increased receipts from customers. It was 33.6% over Q3 2025, largely driven by increased receipts of government subsidies.

Capital expenditures were $633.5 million in Q4 2025, including $559 million for Hua Hong 12-inch and $74.5 million for Hua Hong 8-inch. Other cash flow generated from investing activities was $61.7 million in Q4 2024, including $36.6 million receipts of government grants of equipment, $13.6 million interest income and $1.2 million receipts of disposal of equipment, partially offset by $3.6 million investment in the equity instrument.

Net cash flows generated from financing activities was $1.3611 billion, including $919 million proceeds from bank borrowings, $594.6 million from other financing activities, $12.1 million receipts of government grants for finance costs and $4.7 million proceeds from share option exercises, partially offset by $136.1 million of bank principal repayments, $32.8 million interest payments and $0.4 million lease payments.

Now let's move to the balance sheet. Cash and cash equivalents was $4.961 billion on December 31, 2025, compared to $3.9047 billion on September 30, 2025. Other current assets increased from $739.7 million on September 30, 2025, to $787 million on December 31, 2025, mainly due to increased value-added tax credit.

Property, plant and equipment was $6.6764 billion on December 31, 2025, compared to $6.162 billion on December 30, 2025, primarily due to capacity expansion in Hua Hong manufacturing. Equipment instruments designated at fair value through other comprehensive income increased from $381.3 million on September 30, 2025 to $478.8 million on December 30, 2025, primarily due to fair value gains recognizing equity instruments.

Interest-bearing bank borrowings increased from $2.3975 billion on September 30, 2025, to $3.1908 billion on December 30, 2025, primarily due to increased drawdowns on bank borrowings. Total assets increased from $12.5117 billion on September 30, 2025 to $14.4538 billion on December 31, 2025.

Total liabilities increased to $5.2895 billion on December 31, 2025, from $3.5026 billion on September 30, 2025. Debt ratio increased to 36.6% on December 31, 2025, from 28% on September 30, 2025.

Here is a recap of 2025. Revenue was $2.4021 billion, a growth of 19.9% over the prior year, primarily driven by increased wafer shipments. Gross margin was 11.8%, 1.6 percentage points over 2024, primarily driven by improved average selling price and cost reduction efforts, partially offset by higher depreciation costs.

Operating expenses were $425.6 million, 7.9% (sic) [ 17.9% ] over 2024, largely attributable to increased research and development expenses. Other income net was $54.2 million, 146.4% above 2024, primarily due to decreased finance costs and foreign exchange losses and increased government subsidies, partially offset by decreased interest income.

Loss for the year was $110.8 million, narrowed by 21.1% compared to 2024. Net profit attributable to shareholders of the parent company was $54 million, 5.6% dip from 2024. Basic earnings per share was $0.032. ROE was 0.9%.

Finally, let's discuss our outlook for the first quarter of 2026. We expect revenue to be in the range of $650 million to $660 million with a projected gross margin of 13% to 15%.

This concludes my financial remarks. We'll now begin the Q&A session. Operator, please assist. Thank you.

[Operator Instructions] The first question comes from the line of Leping Huang from Huatai Securities.

Dr. Bai, congratulations for the robust results and the successful acquisition of Huali. So beyond the contribution to Hua Hong's revenue and profit, could you elaborate the strategic resource Hua Hong got through this acquisition? And what's your plan to leverage this resource to accelerate Hua Hong's future growth?

Thank you. First, basically, we acquired Fab5, what we call Fab5 within the Hua Hong system, is a 12-inch fab. It has 55-nanometer, 40-nanometer based specialty technology, quite a bit of the technology platform have overlap with what we already have at HHGrace in Wuxi.

I think we look at the acquisition from the following points. We think that that's going to be favorable to our long-term growth. One is that we certainly grow the scale of our company. Through this acquisition, we added about 40,000 capacity, that's already in production with existing customers with -- the scale is one factor.

Another one is with Fab5 joining HHGrace, we can do a better job optimizing the distribution of our different specialty technologies across all the capacity, all the manufacturing capacity. This will show up in higher efficiency for our TD activity, should also show up in higher efficiency and lower cost for our entire manufacturing base.

So basically, we view this as definitely a strategic acquisition, will accelerate our growth both in terms of revenue and as well as our ability to -- profitability, ability to be more profitable. Thank you.

Okay. My second question is about the -- I want to -- about the supply-demand relation of the 8-inch and 12-inch mature fab business this year. So we noticed some foundry, including the largest one, just recently announced to exit some 8-inch business or sell their -- some 12-inch fab to the memory makers.

So what's your view on this supply demand balance of the 8-inch and the 12-inch business globally this year? And what's your impact -- what's the impact on your ASP? So I also noticed that there are some reports that you have some price adjustments in the end of last December. So what's your view of this ASP trend of Hua Hong this year?

Okay. We also noticed some of the reports talking about some of our foundry competitors might be selling some of the capacity to other people. If you look -- if they just change the ownership from one company to another without actually reducing the capacity, then it doesn't really change the supply/demand situation too much.

With respect to some of the logic capacity moving to memory because the memory certainly is in high demand nowadays. That certainly will reduce the supply in the logic side. But overall, it's a positive thing for us because we are mostly in the logic foundry business, although we do have some flash memory business as well. But overall, that would be a positive sign.

I think overall, because of the AI-driven growth in the overall semiconductor market, we think -- we view that as overall positive. It might show up differently in different market segments. It might show up differently in different technology platform we have our capacity in or our product in. But overall, we view that as a positive development for us.

In that context, if the supply gets tighter, it does give us more opportunity to increase prices. We have been doing that over the course of last year. Surgically, it's not across board increase by any means. But surgically for some certain area where we think that we can really meet the supply, so we take the opportunity to move up the prices a little bit, that also show up in -- some of them already show up in 2025 results.

And we expect that in 2026, we might still have some room to go (sic) [ grow ], especially on the 12-inch side. 8-inch, the supply/demand is more in balance compared to the 12-inch. So even if we try to -- we would like to increase prices as well in 8-inch, but our room probably is going to be limited. But overall, we do -- we are cautiously optimistic that we might be able to do something in that area as well. Thank you.

Our next question comes from the line of Ziyuan Wang of Citic Securities.

Firstly, I would like to wish you all a Happy Chinese New Year. [Foreign Language] I have 2 questions. And the first one is, as we can see this quarter, the capacity utilization rate declined slightly. And what are the reasons for that? Are there any uneven or unbalance on the different platforms? And can our capacity be reallocated between different platforms quickly? That's my first question.

The change is fairly small. It's probably almost a calculating error. But I think the main reason is Fab9 will rapidly bring the capacity online. There's always a little bit of lag between how fast they get the equipment installed and get the capacity online versus when you have the loadings and the order for that capacity.

So there's always -- as you -- that's a typical case in a ramping fab that especially when you ramp very fast, there is a bit of a lag between the capacity. And because the loading is based on what's the capacity brought online, so that's the reason there's like a couple of percent decrease.

Got it. That's very clear. And my second question is about our future performance drivers on the demand side. How much of driver will the AI-related product be for the company's future revenue growth? And also, as we see the localization trend, do you think any -- which kind of product categories will be the most significant boost by the localization? And could you provide a ranking or list -- priority list for these products?

It's actually -- this is a complex question. Let me try to kind of see whether I can answer very clearly. I think if you look at from end market standpoint of view, clearly, AI-related products are increasing fast. What does it mean for us is that AI-related product actually cut across quite a few of our technology platform.

For example, the AI-related products in power management area is growing fast, is increasing. MCU, not so much, but there's also a little bit of impact. And power -- discrete power devices also have some impact. But the power management is one area we clearly see strong growth related to AI.

So in that regard, if you stay at this end market dimension, is now AI is one growth area. Other areas like autonomous driving, automotive, the car related, all the new robots is also growing, all the green energy-related end market also show growth. All those end markets, the growth area do cut across -- in somewhat a complex manner, cut across different technology platform.

So if I look at our technology platform in that different -- in that dimension, if you just look at our 2025 results, you already see that the two biggest growth area is power management and MCUs. So we -- and those 2 areas and plus in addition to the discrete power devices, constitute the 3 largest technology platform that we have from the revenue standpoint of view.

So going forward, I expect the power management area, the BCD platform we have, will continue strong growth. MCU, where Hua Hong HHGrace has a great advantage, has a great competitiveness, very competitive in this area, is also going to be an area that it's going to grow fast.

And there, I will just take this opportunity to do some marketing. We also have new technology problem in 55-nanometer and 45-nanometer all coming on strong. So that should be also a strong growth there.

In the end, I think if you look at our distribution, our revenue distribution, I will still continue to see MCU, probably one of the biggest power management segment, discrete power devices probably going to -- we'll be more stable. Growth is not as fast, but it will remain #3.

The other 2 in terms of logic and RF, we like that to grow a little bit faster. And stand-alone, we actually -- stand-alone memory will also I think will grow reasonably fast.

Some of the memory shortages mostly in DRAM, it's probably going to have some spill over into -- we probably already spilled over into the NAND memory area, but I think it's going to fill a little bit over to the NOR flash area, so that we should also benefit. So that's how I view the market going forward.

Okay. Can I add a little question on that? How -- Dr. Bai, how do you view the sustainability of this current memory cycle? And is there -- what's the impact on Hua Hong and what kind of measures will we take?

That's a good question. If you look at historical pattern, memory tends to go through boom and bust cycle. Although this time around, a lot of people think because AI is a different beast, that maybe this cyclical nature of the memory market will be a little bit different.

I don't have crystal ball. I do believe that eventually, it will be going to a cycle. Maybe this time, the boom cycle will last longer. It probably will heavily depend on how the AI -- is mostly driven by AI, this latest cycle, AI, how long this cycle is going to last.

But I think in the near future, certainly for 2026, there's no sign that's going to slow down. Maybe in year 2 or 3 that if you go by historical pattern, it should start to come down somewhat.

I should add that right now because of the AI-related area driven up DRAM prices so much, it does have a little bit of a depressing effect on the consumer market because a lot of the consumer product probably can't afford this high DRAM prices. Therefore, they might push out their product refreshment cycle a little bit. So that might -- will come across as a negative for some of our product as well.

But overall, I think the growth area still outweigh the area that's going to be somewhat impacted -- negatively impacted by this super memory cycle that we seem to be in the middle.

[Operator Instructions] Our next question comes from [indiscernible] from [ Guosen ] Securities.

[Foreign Language] I have 2 questions. The first question is about the price. So considering the rising cost of the raw material, we also can see some products such as power device raise the price, but the demand just now, Dr. Bai mentioned, is structural. So how do you see the sustainability of the price hike? So this is the first question.

Okay. In terms of raw material, by the time they get to us in the fab, we call that direct material or indirect material, we do see a few areas where the raw material prices start to show up in the semiconductor materials that we use.

I'm trying to think like a copper is probably one area that will add a little bit of a cost to the copper cable, should we happens to be building a fab, which is where we are. We do see that. And we also see some of the other -- some other raw material increases affecting a little bit of our -- the material we buy.

But I would say, by and large, I do not see this as a significant factor for our cost structure. There's going to be some places that -- there's going to be increases, and there's also going to be some decreases. And overall, I do not think it's going to be a significant increase.

Another factor is that we -- over time, we use more and more domestically produced materials. And in general, their costs are better. So overall, I don't think we're going to be looking at the situation, the material will be a cost increase for us going forward.

Very clear. And my another question is about the utilization. Since we are in good position, but some 12-inch foundries are not yet at full utilization. So how do you see the cycle? So can you give us a little bit of your perspective? So where are we today? And is that possible maybe there is some potential order shift of our customers maybe after we increase the price?

Yes. So the fab utilization are affected by a few factors. There are 2, probably one is how competitive is your technology offering. That includes how -- whether you have a complete offering of the solutions to the customer for what a customer needs. That's one factor.

Another one, of course, is pricing. If you price too high your utilization, you will lose customer on one hand. On the other hand, you can -- there's always -- if you use -- if the prices, you are willing to go down the prices, it does tends to increase your loading.

So for those 2 factors, for example, on technology front, HHGrace is well positioned. We are a premier foundry in China. And lot of our technology platform, I would say we are probably #1 domestically and very competitive even internationally, not all of them, but some of them so clearly. So especially in this specialty technology area, which Hua Hong has -- HHGrace has been working on for the last 3 decades almost. That's one factor.

Another factor is that some of our international customers, especially the European ones, and now we start to see American company as well that have this China for China strategy, that they try to move some of their product that originally were manufacturing overseas to be manufactured inside China. So that's another factor that will help our loading.

When those companies looking for a partner in China, they clearly want to have somebody who is technology-wise is in a good position as well as they want a more stable company, the bigger company. So we are usually being viewed as a first choice many times -- many, many -- in many cases, we're the first choice for their Chinese -- as their Chinese partner.

So we do benefit from that factor as well. I think this trend will probably going to continue giving the whole, the world, this geopolitical situation and the world semiconductor market is evolving. Thank you.

Looking forward to a better performance in the coming year and Happy Chinese New Year.

Thank you.

Next comes from [ Scarlett Ku ] from BNPP. Okay. Otherwise, we will move on to our next questions. One moment, please. We have follow-up questions from Leping Huang from Huatai Securities.

Dr. Bai, I have a follow-up question. What's the current status of Fab9? Is it fully completed? And I noticed the CapEx this year -- last year is $1.8 billion, which is down slightly versus 2024. So how we should model the CapEx for 2026? And when you plan to initiate the next phase of the expansion? And what's your plan on this? What should I say, the Phase 2 of the Fab9 or the new fab?

Look, let me address the question. Basically, the total capital expenditures for this project Fab9A is at $6.7 billion, okay? So by end of last year, we spent about slightly over $5 billion. So we spent another $1.3 billion. Basically, these are the POs, I mean we have basically issued not completely spend from cash flow perspective, okay?

So I would say, basically, there's another about -- to get to $6.7 billion, there's probably another $1.2 billion to $1.3 billion on cash flow -- from the cash flow perspective, okay? So most of POs have been issued for that project. So I would expect the cash will be spent mostly this year and some probably remaining in 2027.

Mostly this year because this year, we're going to reach the peak capacity for Fab8, the first half of Fab9. Your second part of the question is on Fab9B, which is our next project to fill up the remaining, the other half, the empty half of Fab9. That project, we got all the approval, all the necessary paperwork.

We plan to start the actual engineering construction work after the Chinese New Year basically in March. So we should get -- we should be able to start getting the equipment in by end of this year, probably October time frame. The spending obviously is going to be mostly in 2027. So in 2027, we start another capacity ramp on the Fab9B, and we hopefully can complete that ramp even faster, in a velocity that's even faster than Fab9A.

Fab9A we ramped very fast. In 2 years, we pretty much get to the peak. And output will take a little bit longer because, as I said, there's always a little bit of lag between the capacity in place -- being in place versus when you get the wafers out and turn that into revenue.

But in terms of the capacity -- construction capacity in place, using that as a milestone, Fab9B will start in 2027, and we should get that done in less than 2 years as well.

So this year, 2026, the CapEx will be slightly down and 2027 will be up significantly? Is my understanding correct?

Correct. That's correct.

Our next question comes from Ziyuan Wang of Citic Securities.

Okay. I want to have a follow-up question on that. And in terms of our equipment localization rate, will Fab9B have a higher rate than Fab9A?

The fab, yes, you're talking about the fab utilization rate, they are already a little bit above 100. So it's not going to be significantly higher.

I mean on the equipment localization ratio.

Yes. So the answer is yes. The general direction, as the domestic equipment industry become more and more capable every year, that we -- every new project we have, we tend to have a higher procurement of domestic equipment. We obviously still going to be making our procurement decision based on what is the best, both technically as well as commercially for the company. That's the decision criteria.

But the reality is that the domestic produced equipment are becoming more and more capable. And commercially, they tends to be, not across the board, more attractive, a lot of players dependent on individual equipment, they can be more attractive. Then end result we expect is that the Fab9B project, we will end up with a higher domestic equipment content.

Ladies and gentlemen, that's all the time we have for questions. I'll now hand back to Mr. Daniel Wang for closing remarks.

Well, once again, thank you all for the -- for joining us today and for your wonderful valuable questions. The year of horse is right around the corner. We would like to take this opportunity to thank you all for all the support and trust you have given to us and wishing you and your family a very joyful holiday season and a healthy and prosperous New Year. [Foreign Language] We look forward to catching up with you very, very soon. Thank you.

[Foreign Language] Happy New Year.

Thank you. Ladies and gentlemen, that does conclude the conference call. Thank you for your attendance. You may now disconnect.

Ladies and gentlemen, thank you for standing by. Welcome to Hua Hong Semiconductor's Third Quarter 2025 Earnings Conference Call. Today's call is hosted by Dr. Peng Bai, Chairman and President; and Mr. Daniel Wang, Executive Vice President and Chief Financial Officer. [Operator Instructions] The earnings press release and third quarter 2025 summary slides are available to download at our company's website, www.huahonggrace.com.

Without further ado, I would like to introduce you to Mr. Daniel Wang, Executive Vice President and Chief Financial Officer. Thank you. Please go ahead.

Good afternoon, everyone. Thank you for joining our Q3 2025 earnings call. Today, we will first have Dr. Peng Bai, our Chairman and President, share some remarks on our third quarter performance. I'll then take you through our financial results in detail and offer guidance for the upcoming quarter. We'll then open the floor for a Q&A session.

With that, I turn the call over to Dr. Bai.

All right. Thank you, Daniel. Good afternoon, everyone. Thank you for joining our earnings call. Third quarter 2025 sales revenue for Hua Hong Semiconductor reached a record high of USD 635.2 million, in line with guidance, while gross margin stood at 13.5%, above guidance. Driven by the recovery in global semiconductor demand and the company's lean management practices, our capacity utilization remained high.

Both sales revenue and gross margin showed year-on-year and quarter-on-quarter growth. The enhancements in our core competencies, including process technology, R&D, market development and operation, along with the results of our cost reduction and efficiency improvement initiatives are gradually becoming evident. Our overall profitability is improving, laying a solid foundation for long-term sustainable development.

Hua Hong Semiconductor possesses extensive expertise and exceptional management experience in specialty technologies, facing the rapidly evolving global semiconductor landscape. The company must continuously advance across multiple core dimensions such as technology capability and capacity expansion. The acquisition, which is currently progressing smoothly, will further increase our production capacity and diversify our process platform portfolio while creating synergies with our 12-inch production line in Wuxi to strengthen our profitability.

Furthermore, the company is actively engaged in strategic capacity planning, focusing on technological breakthrough and ecosystem development to continuously enhance our core competitiveness amidst the global industry transformation.

Now I would like to hand the call back to our CFO, Mr. Daniel Wang, for his comments.

Daniel?

Thank you, Dr. Bai, for your exciting remarks. Now let me walk you through a summary of our financial performance for the third quarter, followed by our revenue and margin outlook for Q4 2025 before opening the floor for the Q&A session.

First, let's review our financial results for the second -- for the third quarter. Revenue reached an all-time high of $635.2 million, 20.7%, over Q3 2024 and 12.2% over Q2 2025, primarily driven by increased wafer shipments and improved average selling price. Gross margin was 13.5%, 1.3 percentage points over Q3 2024, primarily driven by improved capacity utilization and average selling price, partially offset by increased depreciation costs and 2.6 percentage points above Q2 2025, primarily driven by improved average selling price.

Operating expenses were $100.4 million, 23.3%, over Q3 2024, primarily due to increased engineered wafer costs and depreciation expenses and 2.6%, over Q2 2025. Other income net was $70.8 million, 65.7% lower than Q3 2024, primarily due to decreased foreign exchange gains and interest income, partially offset by decreased finance costs and 67.4%, over Q2 2024, primarily driven by foreign exchange gains versus foreign exchange losses in Q2 2025.

Income tax expense was $10.4 million, 9.6% lower than Q3 2024, primarily due to decreased taxable income. Net loss for the period was $7.2 million compared to a profit of $22.9 million in Q3 2024 and a loss of $32.8 million in Q2 2025.

Net profit attributable to shareholders of the parent company was $25.7 million, 42.6% lower than Q3 2024 and 223.5%, above Q2 2025. Basic earnings per share was $0.015, 42.3% lower than Q3 2024, and 200%, over Q2 2025. Annualized ROE was 1.6%, 1.2 percentage points lower than Q3 2024, and 1.2 percentage points above Q2 2025.

Now let's take a closer look at our Q3 2025 revenue performance. From a geographical perspective, revenue from China was $522.6 million, contributing 82.3% of total revenue and an increase of 20.3% compared to Q3 2024, mainly driven by increased demand for flash, other power management IC and MCU products. Revenue from North America was $63.8 million, an increase of 36.7% compared to Q3 2024, mainly driven by increased demand for other power management IC and MCU products. Revenue from other Asia was $30.3 million, an increase of 5.6% compared to Q3 2024. Revenue from Europe was $18.4 million, an increase of 12.6% compared to Q3 2024, mainly driven by increased demand for IGBT and smart card ICs.

With respect to technology platforms, revenue from embedded non-volatile memory was $159.7 million, an increase of 20.4% compared to Q3 2024, mainly driven by increased demand for MCU products. Revenue from stand-alone non-volatile memory was $60.6 million, an increase of 106.6% compared to Q3 2024, mainly driven by increased demand for flash products.

Revenue from power discrete was $169 million, an increase of 3.5% compared to Q3 2024, mainly driven by increased demand for super junction products. Revenue from logic and RF was $81.1 million, an increase of 5.3% over Q3 2024, mainly driven by increased demand for logic products. Revenue from analog and power management IC was $164.8 million, an increase of 32.8% over Q3 2024, mainly driven by increased demand for other power management IC products.

Now turning to our cash flow statement. Net cash flows generated from operating activities was $184.2 million compared to net cash flow used in operating activities of $26.8 million, primarily due to increased receipts from customers. Capital expenditures were $261.9 million in Q3 2025, including $230.7 million for Hua Hong Semiconductor Manufacturing, $19.3 million for Hua Hong 8-inch business, and $11.9 million for Hua Hong Wuxi.

Other cash flow generated from investing activities was $8.6 million in Q3 2025, mainly including $15.6 million interest income and $7 million receipts of government grants of equipment, partially offset by $14 million investment in equity instrument. Net cash flows used in financing activities was $104.2 million, including $99.9 million proceeds from bank borrowings and $14.4 million proceeds from share option exercises, partially offset by $5 million interest payments, $3.2 million of bank principal repayments and $1.9 million lease payments.

Now let's have a look at the balance sheet. Cash and cash equivalents were $3.9 billion on September 30, 2025, compared to $3.85 billion on June 30, 2025. Other current assets increased from $688.5 million on June 30, 2025, to $739.7 million on September 30, 2025, mainly due to increased value-add tax credit. Property, plant and equipment was $6.2 billion on September 30, 2025, compared to $6.1 billion on June 30, 2025. Equity instruments designated at fair value through other comprehensive income increased from $290.5 million on June 30, 2025, to $381.3 million on September 30, 2025, mainly due to an increased fair value of the equity instruments.

Total bank borrowings increased from $2.3 billion on June 30, 2025, to $2.4 billion on September 30, 2025, mainly due to withdrawal of bank loans. Total assets increased from $12.2 billion on June 30, 2025, to $12.5 billion on September 30, 2025. Total liabilities increased to $3.5 billion on September 30, 2025, from $3.4 billion on June 30, 2025. Debt ratio increased to 28% on September 30, 2025, from 27.5% on June 30, 2025.

Finally, let's discuss our outlook for the fourth quarter of 2025. We expect revenue to be in the range of $650 million to $660 million with a projected gross margin of 12% to 14%.

This concludes my financial remarks. We now begin our Q&A session.

Operator, please assist. Thank you.

[Operator Instructions] The first question is from the line of Leping Huang of Huatai.

I have two questions to Daniel and one question to Dr. Bai. So the first question, I noticed there's a very strong beat on the gross margin this quarter. Also, I think ASP up 5.2% Q-on-Q this quarter. So Daniel, can you explain -- break down the reason of this strong margin and ASP beat between the product mix improvement and the price adjustment of the existing products? Also, you gave the guidance for next quarter. So what's the -- your outlook for your ASP in the fourth quarter?

Thank you, Leping. First of all, we had extremely high utilization rate, okay? So the three 8-inch fabs were consistently above 110% utilization rate. And the first -- our first 12-inch fab with 95,000 wafer capacity, the loading was consistently above 100,000 wafers, okay? And then the other -- the fab that is currently ramping, it has about 40-plus thousand wafer capacity, but the loading is above 35,000. And pretty soon it's going to get to about 40,000 wafers in loading, okay? And in terms of that itself helps the margin.

And also, I think the most critical thing is the ASP improvement. We start to raise ASP in the second quarter and start to take effect in the third quarter. So you said, absolutely, is correct, overall, gross margin is about -- the price ASP is quarter-to-quarter or even compared to last year, it was about 5.2%. Basically, the ASP improvement was coming from all technology platforms -- all technology platforms, including embedded non-volatile memory, okay, power discrete and logic and RF, analog and power management IC, okay? So it basically came from all technology platforms.

I would say, if you want to really look at between product mix and ASP improvement, I would say 80% came from ASP improvement. And the other 20% is largely due to product mix. As far as going forward, I think we will continue to improve ASP. I mean we're looking at every order that comes in. We make sure that we -- if there is any opportunity, we can adjust price, okay? So I'm extremely positive about our Q4 in terms of ASP and the gross margin as well.

Okay. The second one is also for Daniel. So you also mentioned the utilization rate of your fab is very high. It's already 109%. So it seems to be -- do you think that -- so first, what are the actions you take to improve your factory utilization rate? And also, what you expect the utilization rate looking forward? It seems to be you are adding more wafer into your Wuxi fab. So do we -- should we expect the utilization rate will further improve in the coming quarters?

You know what? Excellent question. I'll let Dr. Bai answer the question.

Okay. Let me try. There's a few factors playing together. It's kind of -- there's some interplay of a few factors. First of all, utilization number, as you know, is based on standard IE calculation, meaning you're supposed to set up certain capacities and based on that number, if you do better, your utilization can be a little bit above 100%. But clearly, you can be significantly above 100%, otherwise, the number would be incorrect.

In our case, Fab 9A capacity is -- continues to come online. So we can -- there's two benefits. One is Fab 9A itself started to contribute to revenue. It actually adds more pressure to gross margin because all the depreciation also starts to come online. But it does provide some avenue to make our existing capacity a little bit more flexible in the sense that now you have a bigger scale when the product mix shifts that you can kind of use each other's capacity in each factory. That's how we can get the capacity utilization a little bit better above 100% or 105%.

It's -- let me also add a comment to the gross margin. Gross margin also compressed because I always said there's always a balance between how much more depreciation coming online, which is inevitable as the new capacity start to contribute to revenue on one hand. On the other hand, you have -- whether you can manage to increase prices. Daniel already talked about, we did manage to increase our prices by, you already calculated, around 5%-ish in Q3.

Another factor was our general cost reduction effort to make our cost structure a little bit better. That effectively balanced out some of the pressure we get from increased depreciation coming online. In the end, the Q3 story was a very good one. It was -- it also exceeded our expectation. But I do see that momentum in cost reduction as well as the price stabilization, if not increase, also start to take a hold. So that's a good trend for us.

One more add to this is that when the demand is a little bit higher than our supply, which relates to why our utilization is so high, it also gives us a little bit of leeway, a little bit of flexibility in optimizing our product mix, namely, we can choose to focus or give priority -- or give it capacity priority, a little bit more for the product that has a higher margin than the ones that have a lower margin. That also help our price increases.

If I -- I know I said a lot of things, but all those factors are there, and it's really the end result, and net is an interplay of all those factors that gives us the overall Q3 results. Thank you.

Okay. So the final question I want to ask is that I noticed that in Daniel's statement that the flash business is one major driver for your -- especially your China business. And in the investor community, we are talking about the memory super cycle. So Dr. Bai, so what's your view on how Hua Hong can benefit from the coming memory cycle and why the -- is it initial sign we see this quarter that your memory business or your flash business is going very strong? Is it initial sign of this memory super cycle?

Okay. Thank you for the question. First, I want to clarify the memory business that Hua Hong is engaged in is in the NOR Flash, which is one segment of overall flash business.

And that NOR Flash business, there's two parts to it. One is the stand-alone, just the stand-alone NOR Flash product. Another one is MCU that's basically integrated with the logic circuit. We are participating in both, MCU as well as flash memory. You are correct. We see a strong business in Q3 in both, MCU as well as stand-alone flash. I would say overall NOR Flash market has a steady growth. It's probably a little bit different than the overall memory business. The other memory business, like the correlation with the other memory business is not that strong. For example, the NAND might be doing -- has its own dynamics. DRAM, MCM, HBM, DRAM-based HBM, all those related to AI applications, those has its own dynamics.

Our part of the business, we do see steady growth in the NOR Flash business, in both MCU and stand-alone. Our, if you will, Q3 growth rate clearly is faster than the overall market growth. That probably has more to do with our own situation where our 55-nanometer NOR Flash started coming into the mass production phase in the last couple of quarters, that started to pick up volume. And also our 55-nanometer MCU business also is going into the mass production. And in the next year or 2, we're going to have 40-nanometer. In next year, 40-nanometer NOR flash business stand-alone as well as the MCU will come online. That will give us another push.

So in general, we do see that our flash business will have a strong growth over the next few quarters, even next couple of years, mainly based on our new technology -- new technology transitions. I would say the other memory business, their dynamics might be a little bit different. Some of them are also growing strongly. It's probably not correlated with our situation.

Our next question comes from Ziyuan Wang from Citic Securities.

[Technical Difficulty].

Operator...

Can you help?

We couldn't hear the question clearly. So can you please ask him to repeat?

Ziyuan, would you be able to dial again or change to another better connection to repeat your question, please?

Sorry, can you hear me now?

I'm afraid the line is bad. Would you like to dial back, and we will take your question.

[Operator Instructions] The next question is from Jian Hu from Guosen Securities.

[Foreign Language] So I have -- first, I have a quick follow-up. Just now, Daniel also mentioned the growth drivers for the next few quarters, some factors like capacity expansion and also the price increases. So how about the product structure adjustment for the future growth? So could you give us more details? And this is the first question.

Sure. All right. In terms of capacity expansion, we basically -- you will see a continued increase from our Fab 9A because we are still in the capacity expansion phase. Earlier, Daniel mentioned that the Fab 9A reached about 3,000 to 4,000 wafer -- 30,000 to 40,000 wafer per month right now. It's been climbing over the last 3, 4 quarters. That expansion will continue all the way towards till middle of next year. That will reach the peak of, I would say, 60,000 to 65,000. So you -- and those capacity will come online, will continue to give us -- contribute to the revenue growth. So that's on the capacity expansion front.

In terms of the product mix, optimizing the product mix, that's really come down to the technology evolution, how our technology platform will evolve over the next few years. The key technology that we see that it will become better and more competitive. One, starting with the flash related, I talked about earlier, flash is a factor -- a growth sector for us. I think we -- with our 55-nanometer products online and next year, 40-nanometer products online, that will give even stronger position in this sector. So hopefully, that will bring the added value of the prices up with it.

Another significant technology platform is the BCD platform for power management. Now we see a strong growth, and we see -- we are also purposely expanding capacity for BCD and basically skewing our product mix -- capacity mix towards supporting more BCD and BCD technology. BCD platform happens to be one that has a better margin among the technology platforms that we offer. So that will also give us a better product mix in essence.

So then we continue to add -- continue to strengthen our overall technology development. This is one of the areas that is definitely a focus for the company. When I talk about how can we further improve our core competence, it's really talking about our technology capability and associated marketing capability.

So all the key specialty technology platform, we basically will continue to invest heavily. And some of the platforms, we're already the best. We're already #1 in China, also very competitive worldwide. Some of them, we still have a little bit of distance to travel to become world-class, to become the best. In general, we're best in China in most of the technologies we participate in that we were -- then in some of the areas, we still need to improve a little bit to become really truly world-class.

Here, I can also mention that some of the partnerships we have with our mainly European companies to -- in the context of China, their China for China strategy is also a way for us to increase our competitiveness. So I think I will stop right here in terms of answering your question. I don't know if that clarifies for you.

Very clear. My next question is a relatively big one. So driven by the boost from AI, we can see the global semiconductor sales have grown for like 8 quarters. So compared to the previous cycle, it's a relative long growth period. So how do you see the growth momentum in following quarters?

It is a very big question, a broad question. I think the -- I'll give you my personal take on this. AI is still at its infancy. I think the AI will continue to grow. How it manifests the growth -- how does the growth manifest in the different segment of semiconductor, that's a little bit complex. The direct benefit, obviously, is for the advanced technology, advanced node, which Hua Hong Semiconductor is not directly participating.

But there's a lot of supporting technology associated with the Al products. We are a big part of those segments, like power management, because when you have -- you make AI systems, you need a lot of power management, either for training or now the industry seems to switch towards more deduction type of applications from training. So I think we all -- we definitely benefit from overall AI growth through their increased demand for power management, for MCU, for all kinds of -- power discrete products, they all need those. They need those in order to make the AI system work. So we are definitely part of that ecosystem.

So we were -- in fact, some of the power management demand increase -- strong demand increase over the last year and continue -- that seems to continue into next year or 2 is primarily related to AI and plus, some of the new application on the horizon like cars and robots, those type of things. AI definitely is a big factor.

So that's how I see it. I think AI will continue to grow. You might -- the product mix there might go through its own evolution. But overall, I think that bring along the whole -- all the chips that's needed in AI system, that's where we get the direct benefit, is all those associated chips in AI system that we do directly participate.

Yes. So I have a follow-up. So how about the power semiconductors. So compared to last few years, power semiconductors have shown some recovery, but still besides the AI demand, the rest of the part, the demand is relatively flat. So how we increase the pricing following quarters? So how do you see the pricing in this part?

All right. You are a very astute observer. I agree with you, the power discrete platform amongst our technology platform that we participate in probably has the biggest pressure in terms of growth. I think there are a couple of factors. One is, there are increased competition and increased capacity in the power discrete. Because the power discrete area, the barrier to entry is relatively small. So there are -- there has been over the last few years, large capacity coming online. So that's one factor that put some pressure on us.

The second factor is some -- second factor is technological because right now, the compound Semiconductor like silicon carbide, mostly silicon carbide, but gallium nitride also start to become a significant factor. Those compound semiconductor-based devices become a significant factor in the overall power devices market that inevitably take away some of the silicon-based devices, especially, for example, the super junction, that used to be a Hua Hong -- still is a Hua Hong strength. But that is directly -- there is a direct competition for super junction-based product, silicon super junction based product from silicon carbide. And silicon carbide looks like over there, people are willing to cut price very, very drastically. So they start to have some competition with our super junction.

And so this is one of the topic we've been -- inside the company, when we talk about our technology road map and also talk about our market perspective, is one of the focus area that we will come out with some strategy. We already started gallium nitride development. So we will definitely -- we have been a big player in the power discrete, we definitely will not give up this market segment. We will continue to be bigger and stronger by adding all the -- whatever the customer needs.

So there's a few new initiatives in the power area that we will try to meet the challenge. The challenge is mostly a little bit long term. Short term, I don't see a huge problem, but longer term, over the next 3 or 5 years, that, we do need to do something there to make sure that we continue to keep our very strong position historically in this area. Thank you.

Thank you, Dr. Bai. I also agree with you, and we also think gallium nitride on silicon is a good direction for the new power semiconductors. So that's all my questions and looking forward to a better performance in next quarters.

Once again, we will take the question from Ziyuan Wang from Citic Securities.

Sorry for the connection previously. My first question is about the international customer adoption. We can see that this quarter, the proportion of customers in U.S. and Europe increased. And we also know that STMicro previously announced that they plan to produce 40-nanometer MCU in Hua Hong by the end of 2025. So how is it going? And are there any other new developments or new customers that you can share?

You're correct that we have a partnership with ST on MCU, 40-nanometer MCU. That project has been going very smoothly. In fact, it's a little bit ahead of schedule. We already started with production in this quarter. So that's a little bit ahead of schedule. That will continue -- it will start to contribute to our revenue in the next quarter.

In terms of ramp rate, it will take a while to get up to a fairly high volume, but it's definitely a steady and very robust addition to our product mix. So that one is going well. Actually, this is one of the first collaboration projects we have with ST as well as other European companies for their China for China strategy. I think since now that we have worked together for quite a while and with this track record, everybody's confidence level has increased significantly. That's probably going to play a very positive role in terms of expanding our collaboration in the number of the products and also the area of the -- where we can collaborate with each other. So I would say that definitely is a hugely positive start, and that will start -- next year, you will see a multiply of those collaboration projects come to fruition in the next year.

In terms of the international portion of our business, we always like to increase our international business because this is one of our strategy. Ever since I started here, we set a strategy to see how can we increase our international business, in terms of the percentage of international business. We are right now probably 15% to 20% range. I haven't looked at the number exactly, but that's the range we are in. I think Europe and North America, both regions, I still do see strong growth going forward. Asia is a little bit more challenging, but our two biggest international region, North America and Asia, will continue to grow strongly over the next few quarters.

Okay. And these kind of customers also benefit from the AI, especially the AI power, right?

Correct. If you look at our business from North America, a big part of it -- part of it is the power management chips. Indeed, those are the ones that got used in the AI systems.

Okay. Very clear. And my second question is about the CapEx. Is there any outlook for CapEx for next year as we are continuing to expand our capacity in Fab 9? Will there be any increase compared to this year? Or will it be stable?

Thank you, Ziyuan. Let me just give you an update on that. Basically, for the -- for 2025, the three 8-inch fabs, roughly, it's about $120 million overall on a cash flow basis. That is the CapEx spending for the three 8-inch fabs. And then we -- the expected CapEx for Fab 9A is about $2 billion for this year. So we spent about $3-plus billion up to the end of last year. So we're going to be spending about $2 billion this year. So that gets us to about $5-plus billion. The overall project -- the total investment for the project is $6.7 billion. So it will be about $1.3 billion to $1.5 billion for next year for Fab 9, okay? So that is it for basically the CapEx for this year. So it would be $120 million for the [ 8-inch ], plus $2 billion. Next year, it would be just around $1.3 billion to $1.5 billion for the remaining of the CapEx spending that we have to spend for this Fab 9.

And of course, in the future, if we do have -- we want to continue to grow, we want to continue to expand. We have plan to basically build another fab, but that will be a different story, okay? So when that happens, we'll let you know what would be the total capital spending for that new project.

Our next question comes from the line of Tony Shen of SPDBI.

[Foreign Language] Dear management, this is Tony from SPDB International. I've also got two questions here. The first one, can we have some color into the semi cycle, especially for Hua Hong into next year, 2026. In our current stage, it's very good. The cycle is trading up. We have a little bit of tight supply with high demand, and the gross margin is also trading up into third quarter and also into the fourth quarter. Do we still see the tight supply will continue into the next year? And can we continue to raise prices for most of our products into next year? This is my first question.

From the market standpoint, we do see the momentum will go into next year. We think next year should be better than 2025. There's uncertainties and but just from the pure market unless something big happened like some of the geopolitical or otherwise, we do see that the growth will continue into next year. That will give us some opportunity to raise prices or at least keep the prices stable. I think we -- I want to be a little bit cautious in terms of raising too high expectation of how much prices we can raise because we are still in a very competitive industry. And there's many, many factors involved.

But I do see overall, if the demand -- if demand goes up, if nothing else, you give us -- will give us a way to optimize our product mix, I can choose to make more higher-margin products than lower ones. So basically, I have at least that possibility -- that flexibility. And also with our improved technology offerings, we basically add value to our customers' products. We tend to be able to -- in that scenario, we should be able to also share the benefits that come out from those improvements with our customers, have some kind of win-win situation. So in general, I'm cautiously optimistic to use the cliche that 2026 will be better than 2025.

Okay. Perfect. That is very clear. And my second question is still related to AI servers. Can I have a basic sense of how much revenue may come from directly or either indirectly from AI servers? And how do we see the growth potential, especially into next year? Yes. This is my second question.

For AI server, there's -- power management is the obvious product that goes into that. If you look at our power management business, I think it's about 10% to 15% there. And not all of them are going to the AI server, but the growth -- the bigger growth part is related to power server. So look at the numbers here that -- so I would say the power management, we put it into analog and power management category, is about 25% of overall revenue. Probably right now, more than half -- about half of it -- about half of it is related to AI servers. So it's about 10% to 12%. That portion, we believe will continue to grow strongly. Thank you.

[Operator Instructions] Our next question comes from the line of [ Scarlett Ker ] from BNP Paribas.

[Foreign Language] First of all, congratulations on the strong performance. My question is on the -- one of the numbers. So the operating cost is around USD 100 million. Could you share a bit of a breakdown of the wafer engineering cost and the depreciation? And could you also elaborate a bit on what drives the increase of the wafer engineering cost? And going forward, what you expect the trend to be for both, engineering cost and the depreciation?

So out of that $100 million, the depreciation costs related to the R&D. It is about $18 million for this quarter, Q3, okay? And we continue to invest in the R&D. We have a lot of new products, tape-outs. So this number, we expect in the future will continue to be stable and will probably continue to grow as well, okay? So basically, you have to realize the more we invest, that number will go a little bit higher. Compared to a year ago, that number is much higher now. So that is basically the breakdown. But the other, I would say, $80 million is all related to mostly labor, IT and some other stuff.

Our next question comes from Qingyuan Lin from Bernstein Research.

Congrats on the good results. My first question is around the -- one segment around the industrial and auto. How much is auto versus industrial? And do we see stronger growth on the auto segment because we are indeed seeing the auto demand in China are still quite strong?

So the industrial and auto, that part is about -- overall for that segment was about -- it's going to be nearly about -- for Q3, it was about 22%, okay? But going forward, I expect that segment will continue to grow. We have -- we expect this segment will grow -- have a pretty big growth percentage for Q4 quarter-over-quarter, okay? So out of that 22%, about 26% is related to industrial, about 6% is related to automotive. In fact, industrial has been recovering throughout this year compared to a year ago.

Just to add a little bit to Daniel's answer is that the category you call industrial auto, it actually cut across all our product lines. Some of them are in power discrete, some of them are in MCU, some of them are in power management because those end product -- end user products like auto, industrial, they use all kinds of products. And so it's not just about, say, power management or MCU or power discrete. So it's basically a different -- but the number that Daniel gave you is the correct number.

Got it. Got it. It's very clear. And the second question is around the power discrete actually. It looks like the percent of revenue from power discrete coming down a little bit. Is it capacity constraint? Or we're actually pushing out a bit of a demand because we don't see sort of enough demand there?

We talked about this in one of the earlier questions, too, that power management -- the power discrete as a percentage of our revenue is going down a little bit because the other -- it has not grown as fast as the other segments. So that's how the number play out. But in absolute numbers, the power discrete business is still grow a little bit, but as a percentage, because it hasn't grown as fast with others. So relatively speaking, it will have a smaller percentage of our overall business -- our overall revenue. So that's number one.

The second point is that the reason we discussed earlier that this is one segment that we do see more competitive pressure, mostly on pricing. We still have fully loading, and demand is still strong, but the pricing is -- we won't be able to -- we haven't been able to raise price on this area too much. And going forward, it's probably going to be continued, a bit of pressure, just because of the entry barrier to this market segment is relatively low. And plus, we talk about early silicon carbide and start to have a bigger -- to a larger extent and gallium nitride to a smaller extent, start to have added pressure for this market segment.

Very clear. Maybe last question quickly on any updates on Fab 5 consolidation, time line impact, synergy, et cetera?

Well, it has been moving along according to schedule, okay? So we had our first announcement. We basically already announced the price for the deal. We're negotiating almost close to the completion. Pretty soon, we're going to have our second announcement, second Board meeting. And we expect that will happen very, very quickly. So we expect -- we -- literally, we're going to start to take over the operation start beginning of next year, expect the transaction will be closed by August next year, okay? And all of that will happen -- even the shareholder meeting will probably happen in December. So we're moving according to the schedule.

We're working with the shareholders, the other side, very closely to make sure in the end, we're going to have a very fair deal, a fair deal, a fair transaction. And whatever we're going to pay for the value is going to be -- from a company's perspective, we're looking out to the interest of all the independent shareholders, okay? So we want to protect the interest of all shareholders.

Just to add a comment there that the acquisition deal, obviously, is moving along at a pace that's commensurate with the regulatory requirements. We are following all the requirements of both exchanges because we're listed in both Hong Kong and in Shanghai. So we definitely follow all the regulatory requirements and taking all the steps.

Our goal is to have a good deal for all the shareholders. as well as the seller as well as the buyer side of the shareholders. So this obviously require a lot of work to negotiate to kind of get the assessment right. Daniel and Daniel's team has done a great work so far. We're getting close to the second milestone of announcing something very quickly that will set the deal price, I think then we will follow the regulatory requirement of having all the appropriate approvals that we still need to get -- go through.

We hope that in this process, all the people who support our company, support Hua Hong Semiconductor, please do your part to make this thing go through smoothly. Acquisitions are always not an easy thing to do, especially when listed in both places. But we are pretty confident that we will have a very good outcome for everybody, for all the stakeholders as well as all the people who have been cheering for Hua Hong. Thank you.

Right. I mean just one last thing. It's going to be a good acquisition. It's going to basically give us, as I mentioned before, to many investors, $600 million, $700 million revenue addition. The company is profitable. Most of depreciation is behind us. So it's going to be good for Hua Hong Semiconductor. And then long term, consolidation is the way.

Yes. I think Daniel makes a very important point that this acquisition, strategically, is definitely a very, very good deal for the company because it has a lot of synergy. It can -- our growth model is both organic, which we have been doing very aggressively over the last few years as well as inorganic through acquisition. If we think -- you ask if the total 1 plus 1 is going to be larger than 2, we will do that. And this is a good example of having a target -- having an asset that can significantly add to our growth as well as increase our synergy that we're -- it should help our long-term growth and the profitability picture. Thank you.

And also with the additional -- the specialty technology platform under Dr. Bai's leadership, I think it's going to be more profitable.

Thank you. Ladies and gentlemen, that's all the time we have for questions. I'll now hand back to Mr. Daniel Wang for closing remarks.

So this concludes our today's call. Once again, thank you all for joining us today. It's been very exciting. Thank you for all your thoughtful questions. We appreciate your continued support and look forward to speaking and seeing you again soon, next quarter, okay? Thank you.

Ladies and gentlemen, thank you for your attendance. You may all now disconnect.