Kingsoft Cloud Holdings Ltd - ADR Aktie

Good morning, ladies and gentlemen, and thank you for standing by for Kingsoft Cloud's Second Quarter 2026 Earnings Conference Call. [Operator Instructions] Please note that today's call is being recorded. I will now turn the call over to Mr. Jackie Zou, Senior Director of Capital Markets at Kingsoft Cloud. Jackie, please go ahead.

Thank you, operator. Hello, everyone, and thank you for joining us today. Kingsoft Cloud's Second Quarter 2026 earnings release was issued earlier today and is available on our IR website and DuoReswire. Joining us today are Ms. Zou Tao, Chairman and CEO; Ms. Yi Li, CFO, Mr. Liu Tao, Senior Vice President; Mr. Ken Kayan, Senior Vice President; Mr. Yu Jung, Vice President Mr. Joe Rio, Batumi Vice President; and Mr. Clark Ken, Board Secretary and Associate Vice President.

Mr. Zou will discuss our business performance and key developments followed by Ms. Li with a review of our financial results. Management will then take your questions. Consecutive interpretation will be provided for convenience and for reference only.

In the event of APs prudency, management statements in the original language will prevail. Before we begin, I would like to remind you that today's call contains forward-looking statements made under the safe harbor provisions of U.S. Private Securities Litigation Reform tax of [indiscernible]. These statements involve risks and uncertainties and actual results may differ materially from those expressed or implied by the forward-looking statements. Additional information concerning factors that could cause actual results to differ materially is included in the company's filings with the U.S. SEC. The company undertakes no obligation to update any forward-looking statements, except as required by [indiscernible] Unless otherwise stated, all financial figures discussed on today's call are denominated in renminbi.

With that, it is my pleasure to turn the call over to our Chairman and CEO, Mr. Zou. Mr. Zou, please go ahead.

[Interpreted]

Hello, everyone, and welcome to Kingsoft Cloud's Second Quarter 2026 Earnings Call. I am Zou Tao, CEO of Kingsoft Cloud. This quarter, we saw further evolution in the AI cloud market, the rapid growth of the open source model ecosystem is creating significant opportunities for neutral cloud providers. At the same time, our long-held mission of bringing AI to every industry becoming a reality through a combination of model as service, agent as service and FTE services.

Against this backdrop, Kingsoft Cloud remains committed to technology leadership and high-quality sustainable growth. We are accelerating the development of our AI cloud mass and FTE businesses with encouraging progress. First, AI continues to drive strong revenue growth. Total revenue reached a record of RMB 3.07 billion, up 31% year-over-year. AI cloud gross billings increased 82% to RMB 1.33 billion and accounted for 56% of public cloud revenue.

Mass revenue also rose strongly with Q2 revenue up more than 12x from the Q1 level. Second, profitability improved significantly. Adjusted gross margin rose to 15.4%, up 2.4 percentage points quarter-over-quarter. Operating profit turned positive for the first time with adjusted operating margin reaching a record high of 4.0%. This reflects our continued efforts to capture AI opportunities, improve revenue quality and drive greater operating efficiency.

Third, our customer mix continued to improve with stronger momentum both within and outside our ecosystem. Revenue from the [indiscernible] Kingsoft ecosystem reached RMB 810 million, up 28% year-over-year and accounting for 26% of total revenue. Revenue from our top 5 nonecosystem customers grew 51%.

Our AI cloud business now serves a broad range of sectors, including internet services, front care AI labs, embody AI, autonomous driving, AI foci, Fintech, gaming and online video to name a few. This diversified customer base supports continued growth while allowing us to allocate computing resources more flexibly and strengthen our pricing power and business resilience.

Now let me walk you through our business progress in the second quarter of 2026. In public cloud, revenue reached RMB 2.36 billion, up 45% year-over-year. First, Xiaomi continues to expand AI across its human car on ecosystem, while WPS AI continues to advance. As the only strategic platform for the Xiaomi and Kingsoft ecosystem, we see substantial AI-driven growth opportunities.

In June, our shareholders approved a further increase in the annual caps who connected transactions with Xiaomi. The combined HEPS for 2026 and 2027 now total RMB 10 billion, 39% higher than the full year adjustment. In the first half, public cloud revenue from Xiaomi and Kingsoft grew 54% year-over-year. Second, we further strengthened the mass capabilities of our [indiscernible] platform. [indiscernible] now supports 120 models with major new models launched on the platform in sync with their market release and serves more than 230 enterprise customers. Third, we defend cooperation with leading customers in emerging sectors. We delivered large-scale computing clusters to leading embody AI and autonomous driving customers, supporting rapid model iteration and expanded our cooperation with a leading AI for science customers to support the growth of this new business.

In Enterprise Cloud, revenue reached RMB 710 million. In public services, we signed an agreement with the [indiscernible] Communications Administration of the Yang Zou River to build Zanghai Cloud, a dedicated digital infrastructure platform for Young River shipping. We also followed a strategic partnership with the Wuhan municipal data bureau and Wuhan Cloud across computing resource interconnection, digital governance, intelligent computing applications and ecosystem development.

In digital health, we are leading a project under the national key R&D program on biology and information integration to develop a cloud-based virtual surgery platform, which has been deployed in more than 30 hospitals natural [indiscernible] In Enterprise Services, we defend our cooperation with [indiscernible] to jointly build and operate the Guangzhou Provincial Public Services Cloud under an integrated investment, construction and operations model.

In products and technology, we continue to upgrade our full-stack AI capabilities for intelligent computing and AI application deployment. This quarter, we further optimized the model deployment on [indiscernible] for high concurrency inference, significantly improving throughput for several core models and enabling more granular access usage and model level management.

We also launched Agent kits, providing secure sandbox knowledge and memory management and evaluation and governance tools to help enterprises build production-grade AI agents. At the same time, we are making general service cloud products such as database and storage, easier for agents to access and use. We enhanced the [indiscernible] training and inference platform with more flexible resource scheduling, sharing and allocation for training and fine-tuning workloads, improving utilization and reducing development and operating costs.

So private deployment of domestic AI infrastructure, our Galaxy Stack platform completed deep integration and full life cycle mutual management for multiple mainstream domestic AI chips.

Looking ahead, we will continue to capture opportunities both within and outside our ecosystem, improve the operating efficiency of our computing assets and strengthen our profitability and cash generation capability amid AI industry tailwinds. We remain committed to creating long-term sustainable value for customers, shareholders and society.

With that, I will hand the call over to our CFO, Li Yi, who will review our second quarter financial results.

Thank you, [indiscernible] and thank you all for joining the call today. I will now discuss the second quarter financial results [indiscernible]. Before we go through the details of the financial results for the second quarter, I would like to highlight [indiscernible]. First, our quarter revenue reached over RMB 3 billion for the first time in our company's history, up year-over-year for the last consecutive quarter. In particular, our AI cloud gross billing increased 82% year-over-year to RMB 1.33 billion, accounting for over 43% of our total revenue 31% a year ago. This reflects a continued structural shift in our business leading towards AI. Second, our profitability has improved. Our adjusted gross margins were 15.4%, up 2.4 percentage points quarter-over-quarter at 0.5% parentage points year-on-year.

Our adjusted EBITDA margin reached 36% up from 32% in the same quarter last year and 82% last quarter. Mostly when we had to break even at operating income level this quarter and recorded adjusting operating profit margin of 4%. These unconvalidate our ability to strong AIDC demand into healthy profit growth.

Third, we continue to invest to accelerate the buildout of our AI compute capacity. Capital expenditures, together with right of use assets of tenant through third party financing and the finance leases reached RMB 3.3 billion this quarter versus RMB 2.9 billion in last quarter and RMB 2.8 billion in the same quarter last year.

Now let me walk you through our financial results for the second quarter of 2026. This quarter, total revenue were RMB 3.72 million, up 31% year-over-year or 40% quarter-over-quarter. Of these revenues from [indiscernible] cloud service were 2,358 billion up 45% from 1,625 million in the same quarter last year.

Revenues from enterprise [indiscernible] reached RMB 740 million, compared with RMB 724 million in the same quarter last year, down [indiscernible] by 1% year-on-year. Total quarter revenues was [ 606 ] million, representing a 30% year-over-year or increase, mainly due to a continued investment in a cloud infrastructure.

IDC costs increased by 23% year-over-year from RMB 803 million to 1190 million this quarter. The increase was mainly due to the increase of [indiscernible]. Preamortization costs increased by 75% year-over-year from RMB 732 million in the same quarter of 2025 to RMB 964 million in quarter, largely due to the deterioration of newly acquired and listed AI infrastructure, including servers and network equipment.

Solution development and service costs increased by 4% year-over-year from RMB 564 million in the same quarter of 2025 to RMB 786 billion this quarter. The modest increase was mainly due to higher costs incurred in AI transformation in solution development and delivery. [indiscernible] costs and other costs were approximately $66 million in total this quarter [indiscernible] RMB 92 million in the same quarter last year.

Our adjusted profit for the quarter was RMB 472 million, increased by 35% year-over-year and 34% quarter-on-quarter. Adjusted gross margin was 15.4%, up from 14.9% in the same quarter last year and from 30% last quarter. The increase was driven by higher gross margin in public cloud business, thanks to strong demand tailwinds.

On the expense side, excluding share-based compensation cost expenses, our total adjusted operating expense were RMB 391 million decreased from RMB 761 million in the same quarter last year and from RMB 455 million last quarter, mainly reflecting our disciplined cost and expense control of which are adjusted risk and development expenses were [indiscernible] 84 million, up 1% year-over-year. Adjusted selling and marketing expenses were $102 million, down 7% year-over-year.

General and administrative expenses were RMB 105 million down 51% year-over-year, largely due to lower credit loss expenses. Our adjusted operating profit was RMB 124 million tolling profit from adjusted operating loss of $166 million in the same period last year. This improvement was primarily driven by the expansion of our revenue deal higher gross margin and enhanced operating efficiency.

Adjusted operating profit margin was 4% this quarter compared with minus 7.1% in the same period last year at minus 2.2% in the last quarter. Our adjusted net loss was RMB 6 million, down from RMB 300 million in the same quarter last year at RMB 237 million [indiscernible] Our non-GAAP dBA profit was RMB 1,100 million increased by 171% from RMB 406 million in the same quarter last year. Our non-GAAP EBITDA margin achieved 36% compared with 70% in the same quarter last year and 82% in the last quarter. It was mainly due to our improving growth profit as well as higher degration costs in our cost as we accelerate our AI computing capacity buildout. Ended June 30, 2026, our cash and cash equivalent totaled RMB 4,674 million compared with RMB 4,504 million as of March 31, 2026. The modest decrease was mainly due to our continued investment in AI infrastructure to support business growth.

Looking ahead, we aim to capitalize on the excellent growth in AI demand and further investing in infrastructure, expanding our product and service offerings, managing [indiscernible] equity risk and improving operating efficient. We remain commitment to our OEAI strategy and continue to deliver high-quality growth to our shareholders. Thank you all.

So this concludes our prepared remarks. We will now begin the Q&A session. If possible, please ask your questions in both Mandarin and English. So operator, please proceed.

[Operator Instructions] Your first question comes from Liping Zhao from CICC.

[Interpreted]

Let me translate by myself. So good evening, Mr. and Ms. Li. I've got 2 questions on your mass business. First, how well improvements in open store model capabilities affect the company's mass business based on your observations, what's the current usage growth trend and which use cases are driving at most? And second, given the payback period for the mass business might be shorter. Will the company allocate more resources to it?

[Interpreted]

Okay. So just quickly translate. So this answer comes from our SVP, Mr. Nota. So in relation to your first question, the development in open-source [indiscernible] models have mainly 3 impacts. Number one is that we are seeing a very big demand coming from [indiscernible] traditionally flooring model, taking the lead in this area. However, once we have seen the launch of KLM and K3, kind of high-performance models, we're seeing increasingly users from Mainland China adopting and using this made in China [indiscernible] model. And secondly, the increasing use of agentic scenario also broad change to our business. And with that, we have launched, as mentioned in the prepared remarks, the agent product to satisfy such needs. And certainly, it is worth mentioning that in terms of day-to-day routine tasks and workloads, the choice usually is the price for value kind of models, which are essentially the Chinese models. So that is why this development opens source [indiscernible] model is actually beneficial for our business. And your second question regarding the balance between mass business and the computing power. So we basically have different business models for the 2 business. For computing power business, essentially, once we sell the computing power, the utilization is financial 100%, and we usually come with long-term contracts to secure the utilization throughout a prolonged period of time, and therefore, it's relatively safe, so to speak. But for the mass business, it is subject to quality of factors, including the fluctuation of the token price, the launching of new models, which the governors might prefer to use and also the operating efficiency that we're able to achieve in doing the mass business.

So therefore, we generally balance these 2 business models and hope to have each one of them complement the other one. So we generally dynamically evaluate these 2 business and at the product side, how much resources to allocate.

We will take our next question -- next question comes from Wenting Yu from CLSA.

[Interpreted] The first question is that since joined, how has the chip procurement progressed in recent months? And what's your latest full year CapEx guidance? And the second question is about the enterprise cloud. This time the segment of revenue has decelerated in the past 2 quarters. How should we think about the full year enterprise growth and what's the AI transformation and medium-term positioning for this segment?

[Interpreted]

So allow me to quickly translate. So the answer comes from our -- so 3 points. Number one, actually, since 2023, it's been 3 years, and the market has always been hearing voices about the limited supply. So I would say this is actually a new norm. The tech supply on difficulty is actually a long-term kind of situation. But secondly, we should also be aware of the fact that despite of those constraints, financial constraints, the Chinese cloud computing AI industry development has not been restricted or largely restricted by that. And the way that we actually tackle with such situation is that we try to increase the number of business partners that we work with. We try to increase the number of suppliers we work with and we also work with the increasing the compatibility of made in China chips. You are all very well very much aware of the -- recently many of the many China chips are becoming public, and they are particularly good in use cases such as model interest.

Now number three, I would like to say that when you look at the CapEx number from a month-to-month basis, it is usually quite volatile. And I have to say that the purchasing number because of it's usually a large chunk of money in a relatively small number of projects. So the purchase number, if you look at it on a monthly basis, is actually not a linear number. So I would say that for our whole year CapEx estimate, it should still be in line with what we have been expecting and our CFO should be able to give you more details in that regard.

Our cash expenditure includes [indiscernible] asset to lease arrangement, reaching RMB 6.2 billion in the first half of 2026 accounted for over 75% of our portfolio CapEx last year, where July stature cannot fully represent third quarters of all trends. It clearly shows tangible growth acceleration. Accordingly, we maintain our full year CapEx base case unchanged at [indiscernible] Billion. Thank you, Tintin.

We will take our next question.

Sorry, we need to continue [indiscernible]

Apologies.

[Interpreted]

Okay. So this answer comes from our [indiscernible]. So generally, I don't think -- although we are seeing relatively slow growth in the enterprise cloud segment, I would say it is not the right way to look at it from the linear extrapolation perspective. I will give you 3 reasons. I think number one, just to explain why we're seeking -- looking at relative business in this regard is that the upstream supply pricing hiking hike which changed quite significantly in recent quarters, has affected the -- our prospective customers, essentially the SOE companies and also the government agencies to -- they have to frequently adjust their budgeting product, which delayed their decision-making process. So that's number one. And number two, you're all quite aware that the seasonality in enterprise cloud business is quite strong. Usually, the delivery and revenue recognition are concentrated in the second half of the year. So we have actually quite a strong pipeline to deliver in the second half of the year.

And thirdly, this is actually a result of a proactive adjustment of our business structure namely proactively from the project-based business model to an operating base business model, where operating business model from a financial reporting perspective is automatically classified into public cloud. So this is not typically as it what sees as a weakening of the enterprise cloud business. So that's the 3 points I like to offer.

We will take our next question comes from Timothy Zhao from Goldman Sachs.

[Interpreted]

My first question is regarding the mass [indiscernible] Just wondering compared to the peers in the market. How do you think about the Kingsoft Cloud competitive advantage in a master basis in terms of the application scenario, et cetera? And could you share more about the revenue recognition and the profitability profile of the mass service bid.

Second question is regarding the overall pricing trend in the AI cloud business. Just wondering if you can share -- what is the latest trend over the past couple of months? And what have you heard from the customers after you announced certain price hikes or discount reduction over the past few months and whether you are able to quantify the impact from the price hike to your overall AI cloud revenue growth.

[Interpreted]

So in relation to your question about the positioning, we do have a unique position in the mass business Mainly, we're different from some of the total providers, which they have their in-house or proprietary models, we do not have such models. And therefore, correspondingly, we do not have to sell those [indiscernible] models that our affiliate companies have to offer. And as a result, we're able to actually sell and we actually encourage our sales team to sell the models that our customers like the most, for example, GRM et cetera.

So that's number one. And secondly, it's quite important in today's market to have your proprietary or your own computing power, which is the only way that you can actually secure significant profitability in this business.

So in relation to your question about the price hike. So there are basically 2 products which are core solutions that we have employed increasing price -- number one, that is storage and number two, that is computing power. I'll talk about them separately, respectively. In terms of storage, storage is usual, the incremental amount of storage is actually comes with the intelligent computing demand. That is a relatively small portion of the intelligence computing our overall ticket size. And therefore, in the vast majority of the customers that we negotiated with they are relatively [indiscernible] such price hike. In which case, as a result, we're actually able to not only -- in some places, not only pack through the Nepris in our cost, but also increasing our profitability in that scenario.

And number two, in terms of computing power, because of our specific capabilities, including PAC capabilities as well as the operating maintenance and network capabilities, again, we are able to pass through that cost hike into our customers. In some of the cases, we also increased our profitability. And in this quarter, we have also some projects in which we are doing managed the services, which is an asset-light business model. We look forward to seeing more of that coming to [indiscernible]

We will take the next question. Your next question comes from Wes Yang from UBS.

[Interpreted]

Congrats on a solid quarter. Considering the proprietary models and user ecosystem of other cloud providers, how should we think about our long-term positioning in the cloud market and the sustainable margin level down the road?

[Interpreted]

So we believe that [indiscernible] AI cloud service providers is important to be able to offer the top models, which the customers like and also stable services to our customers. So as mentioned, as a neutral cloud player, we are able to be in good relations with all of the top model providers [indiscernible] mono labs and be able to provide the best model according to our customers' demand. And also, we're able to -- based on our technology capabilities, we're able to provide highly available and high reliable services to them. based on out of the SLAs that we signed with them. I think thirdly, in relation to profitability question you asked, it is important to work closely with the LM terms labs to, for example, to optimize, to optimize the influence of those models. And that will include the, for example, working with them based on the undisclosed weighting of the models to increase our influence model influence efficiency.

In some of the cases, we're able to get to very close level or even reach the same level of the influence efficiency coming out from the [indiscernible] companies themselves. Thank you.

We will take our final question. Your final question comes from Yang Liu from Morgan Stanley.

[Interpreted]

Let me translate my question. I would like to ask on the 2 business model, computing power leasing and Model as a service, what is the ROIC for these 2 business models? And what is the marginal change for the ROIC.

Thank you, [indiscernible] At this stage, we don't disclose separate our SD Mark and computing power services. because varies across projects, driven by payback secureness margin, fixed assets and depreciation policies. Overall, must be with much better probability than back into the core servicing at this stage. We have seen continued improvement in maintaining leverage and our business scales are fixed costs are steadily diluted and our trailing translates adjusted operating profit is trend positive, driving a gradual recovery in our overall ROIC. We adhere to demand drilling and a disciplined investment strategy with a strong focus on capital efficiency. With the continuous business structure of [indiscernible] and the material AI mediation. I think our overall LSC will keep improving steadily.

There are no further questions. Apologies. [indiscernible] the question-and-answer session. I will hand back for closing remarks.

Okay. Thank you all for joining us today. If you have any further questions, please contact our IR team. So have a good evening. You may now disconnect. Thank you.

This concludes today's conference call. Thank you for participating. You may now disconnect.

Good morning, ladies and gentlemen, and thank you for standing by for Kingsoft Cloud's First Quarter 2026 Earnings Conference Call. [Operator Instructions] As a reminder, today's conference call is being recorded. I will now turn the meeting over to your host for today's call, Mr. Wayne Wang, Investor Relations of Kingsoft Cloud. Please proceed, Wayne.

Thank you, operator. Hello, everyone, and thank you for joining us today. Kingsoft Cloud's First Quarter 2026 earnings release was distributed earlier today and is available on our website at ir.ksyun.com as well as on PR Newswire services. On the call today from Kingsoft Cloud, we have our Chairman and CEO, Mr. Tao Zou; CFO, Ms. Yi Li; Senior Vice President, Mr. Tao Liu; Senior Vice President, Mr. Kaiyan Tian; Vice President, Mr. Wang Zhenzheng; and Board Secretary, Mr. Bo Tian.

Mr. Zou will discuss our business strategies, operations and other company highlights, followed by Ms. Li, who will discuss the financial performance. They will be available to answer your questions during the Q&A session that follows. There will be consecutive interpretations. All interpretations are for your convenience and reference purposes only. In case of any discrepancy, management's statement in the original language will prevail.

Before we begin, I would like to remind you that this conference call contains forward-looking statements within the meaning of Section 21E of the Securities Exchange Act of 1934 as amended and as defined in the U.S. Private Securities Litigation Reform Act of 1995.

These forward-looking statements are based upon management's current expectations and current market and operating conditions and relate to events that involve known and unknown risks, uncertainties and other factors, all of which are difficult to predict and many of which are beyond the company's control, which may cause the company's actual results, performance or achievements to differ materially from those in the forward-looking statements. Further information regarding this and other risks, uncertainties or factors are included in the company's filings with the U.S. SEC.

The company does not undertake any obligation to update any forward-looking statements as a result of new information, future events or otherwise, except as required under applicable law. Finally, please note that unless otherwise stated, all financial figures mentioned during this conference call are denominated in RMB.

It is now my pleasure to introduce our Chairman and CEO, Mr. Zou. Please go ahead.

[Interpreted] Good evening, everyone, and welcome to Kingsoft Cloud's First Quarter 2026 Earnings Call. I am Tao Zou, CEO of Kingsoft Cloud.

Since the beginning of 2026, the continued adoption of AI coding together with the rapid rise of AI agents have driven AI to evolve from chat-oriented to action-oriented use cases. This shift is fueling concurrent growth in both model inference and training demand, further expanding the setting of the cloud computing industry.

This quarter, Kingsoft Cloud remains firmly committed to our high-quality and sustainable development strategy. We strengthened our AI cloud infrastructure and enhanced our training inference platform capabilities. In the meantime, we also further deepened our presence in industry-specific use cases, fully embracing AI's transformative role in reshaping the world.

First, we sustained our momentum of high-quality growth. In terms of revenue, we reported a total revenue of RMB 2.7 billion this quarter, representing a year-over-year growth of 37.2%. Both public cloud and enterprise cloud services achieved year-over-year growth. Among them, public cloud revenue reached RMB 2.0 billion, a year-over-year increase of 47.5%. In terms of profitability, our adjusted gross profit reached RMB 351 million, up 8.6% year-over-year. Adjusted EBITDA was RMB 748 million, representing a year-over-year increase of 134.7% with adjusted EBITDA margin reaching 27.6%, a significant year-over-year improvement of 11.4 percentage points.

Second, AI cloud continues to drive the company's business growth. This quarter, AI cloud gross billings reached RMB 1.0 billion, a year-over-year increase of 90.1%, accounting for over half of public cloud revenues for the first time, reaching 50.1%. Notably, our token services delivered exceptionally strong growth with April 2026 revenue skyrocketing to 53x that of January.

Third, ecosystem cooperation continued to strengthen. This quarter, revenue from Xiaomi and Kingsoft Ecosystem reached RMB 838 million, a year-over-year increase of 68.9%, accounting for 31.0% of total revenue. As Xiaomi reinforces its investment in the human-car-home smart ecosystem and AI advancements, it brings for more business opportunities for us. We plan to revise the annual caps for the continuing connected transactions with Xiaomi. Following the adjustments, the revised annual caps with Xiaomi and Kingsoft for the continuing connected transactions under the 3-year framework from 2025 to 2027, we reached RMB 14.2 billion.

Now let me walk you through our business highlights for the first quarter of 2026. In terms of public cloud services, revenue reached RMB 2.0 billion this quarter, representing a year-over-year increase of 47.5%.

First, we continue to closely align with the large-scale and a highly visible cloud computing demand within the Xiaomi and Kingsoft Ecosystem. With a long-term and global strategic perspective, we are carefully planning and continuously refining our offerings to build products and solutions with sustainable competitive advantages.

Second, we earned broad recognition from customers outside the ecosystem. Leveraging our solid product and technology capabilities, extensive project experience and strong market reputation, we rapidly expanded both our customer coverage and the depth of business cooperation. This quarter, revenue from our top 5 non-ecosystem customers increased by 66% year-over-year, maintaining strong growth momentum.

We provided AI cloud services to a leading autonomous driving unicorn, enabling rapid deployment and responsive operations. This supported large-scale data processing and efficient end-to-end neural network training iterations, helping the customer capture early market opportunities.

Through our StarFlow training and inference platform with best-in-class resource scheduling, elastic scalability and model deployment capabilities, we effectively supported the token demand of many top-tier Internet companies, capturing the surge in inference-driven demand.

Third, we achieved a meaningful optimization of our customer mix. As inference applications continue to scale, our AI business now spans a wide range of industries, including Internet, AI companies, autonomous driving, logistics, fintech, gaming and video streaming, resulting in a more balanced customer mix.

This quarter, we deeply empowered a leading AI for science customer, ensuring its rapid business growth and stable platform operations. We also supported a top logistics technology company in executing large-scale co-development projects, enabling its engineering teams through flexible multi-modal utilization and significantly improving R&D efficiency and innovation capability.

This diversified customer mix and business portfolio not only drive revenue growth, but also enable us to schedule computing resources more flexibly in off-peak periods, improve resource utilization and enhance profitability.

In terms of enterprise cloud services, revenue reached RMB 710 million this quarter, representing a year-over-year increase of 14.7%. In the public services sector, we launched the state-owned cloud platform in Shenzhen, focusing on the core needs of state-owned enterprises for high security, strong compliance and strict data confidentiality and fully enabling the digital and intelligent upgrade of office, business and management use cases.

Leveraging Kingsoft Cloud's technology foundation, we adopted an integrated architecture that is provincial platform plus multi-prefecture and counting platforms to build a supply chain public information platform in Hubei, enabling resources efficiency, data interoperability and business collaboration and have now supported the sales migration of multiple municipal and country-level platforms to the cloud.

We also partnered with a leading domestic chip manufacturer to build a full stack intelligent computing service system spanning from underlying chips to upper layer applications, advancing the large-scale commercial deployment of domestically developed intelligent computing cloud solutions and meeting the demand for high security and highly controllable computing capabilities.

In the digital health care sector, we collaborated with Union Hospital affiliated with Tongji Medical College of Huazhong University of Science and Technology, one of the top-ranked hospitals in China on the data governance project. Through a systematic data management framework, we helped the hospital transition from fragmented management to standardized governance, setting a benchmark for the intelligent transformation of large medical institutions.

We also signed a contract for a large-scale medical consortium platform project based on the Data Middle Platform. This project highlights our end-to-end professional capabilities in planning, designing, construction and operation within the medical consortium space, laying the foundation for large-scale replication and rollout across health care institutions.

In the enterprise services sector, we delivered a green energy operation platform for a leading clean energy service provider, enabling effective intensive management of large heavy-duty truck fleet. We further extended into the broader green and low-carbon industrial chain, exploring digital solutions for solid waste management and driving large-scale business deployment.

In terms of products and technology, we continue to stay committed to a technology-driven approach closely aligned with AI cloud demand and have comprehensively upgraded our products and services. During the quarter, in response to increasingly diverse model requirements, our StarFlow platform significantly expanded its model ecosystem. We added new API services for speech recognition and speech synthesis, expanded image and video generation models and delivered a more refined user management experience.

To address growing demand for AI agents, we launched Agent Engine, enabling customers to efficiently develop, deploy and manage agents. We also introduced one-click agent deployment on our cloud hosts, supporting mainstream agent applications such as OpenClaw and Hermes, achieving deployment within 5 minutes and significantly lowering the barrier to adoption. For AI training and inference use cases, KS3 cache accelerator now delivers stable millisecond-level low latency, balancing performance and cost efficiency.

To meet the rising demand for private deployment of AI across industries, our Galaxy Stack platform reached a key milestone, adding StarFlow and security modules and completing a full stack closed-loop private deployment solution for AI cloud, covering cloud infrastructure, integrated training and inference token services and security guardrails.

Overall, in this wave of AI innovation from text generation to multi-modal capabilities, from training to inference, from chatting to real-world task execution and from agents to course, the pace of innovation and deepening applications continues to reinforce our conviction that AI will fundamentally reshape industries across the board. Kingsoft Cloud will continue to uphold its strategy of high-quality and sustainable development, increase investment, deepen its focus on core products and solutions and continuously enhance profitability, creating a greater long-term value for our customers, shareholders, employees and society.

Next, I will hand over to our CFO, Ms. Yi Li, who will walk you through our first quarter financial results. Thank you.

Thank you, Mr. Zou and [indiscernible], and thank you all for joining the call today. I will now discuss the first quarter financial results using RMB as currency. Before we walk through the details of the financial results for the first quarter, I would like to highlight the following aspects.

First, our revenue has been consecutively achieved year-over-year growth for 8 quarters, reaching RMB 2,704 million this quarter. For the first time, our AI business became a majority revenue driver, contributing over 50% of our public cloud services revenue and marking a pivotal structural shift in our growth mix. This quarter, our AI business revenue increased 91% year-over-year, amounting to RMB 998 million.

Second, our adjusted gross profit was RMB 351 million, increased by 7% year-over-year despite overall supply chain challenges. Our adjusted EBITDA margin was 28%, increased by 11 percentage points year-over-year, thanks to our AI revenue growth.

Third, in light of strong demand across diverse sectors, we remain steadfast in investing our infrastructure as our capital expenditures and leased assets obtained in combination grew 38% year-over-year to RMB 3 billion this quarter, and we expect to continue to invest to facilitate further business expansion throughout the year.

Now I will walk you through our financial results for the first quarter of 2026. This quarter, total revenues were RMB 2,704 million. Of this, revenues from public cloud sources were RMB 1,996 million, up 47% from RMB 1,353 million in the same quarter last year. Revenues from enterprise premium services reached RMB 707 million, up 50% from RMB 616 million in the same quarter last year.

Total cost of revenues was RMB 2,358 million, up 43% year-over-year, which was mainly due to our investment into AI computing resources. IDC costs increased by 26% year-over-year from RMB 723 million to RMB 911 million this quarter. The increase was mainly due to increase of rack services, which served the expanding AI business.

Depreciation and amortization costs increased from RMB 379 million in the same quarter of 2025 to RMB 819 million this quarter. The increase was mainly due to the depreciation of newly acquired and leased servers and network equipment, which were mainly related to AI business.

Solution development and services costs increased by 14% year-over-year from RMB 505 million in the same quarter of 2025 to RMB 575 million this quarter. The increase was mainly due to the solution personnel expansion of Camelot. Fulfillment costs and other costs were RMB 2 million and RMB 51 million this quarter.

Our adjusted gross profit for the quarter was RMB 351 million, increased by 7% year-over-year and decreased by 25% quarter-over-quarter. Adjusted gross margin decreased from 17% last quarter to 13% this quarter. The decrease was mainly due to the higher cost of server along with the expansion of our AI business as well as upfront costs incurred for future revenue-generating activities with certain customers.

On the expense side, excluding share-based compensation expenses, our total adjusted operating expenses were RMB 455 million, remaining stable compared with same quarter last year and last quarter, of which our adjusted R&D expenses were RMB 184 million, decreased by 8% from same quarter last year. Adjusted selling and marketing expenses were RMB 112 million, increased by 4% year-over-year. Adjusted general and administrative expenses were RMB 159 million, increased by 34% year-over-year.

Our adjusted operating loss was RMB 60 million, increased by 7% from adjusted operating loss of RMB 56 million in the same period last year. The improvement was mainly due to the expansion of revenue scale. Adjusted operating loss margin decreased from 3% in the same period last year to 2% this quarter, representing a decrease of 0.6 percentage points.

Our non-GAAP EBITDA profit was RMB 748 million, increased by 135% from RMB 390 million in the same quarter last year. Our non-GAAP EBITDA margin achieved 28% compared with 16% in the same quarter last year. It was mainly due to our strong commitment to AI cloud computing development and strategic adjustment of business structure. This quarter, our capital expenditure, including those financed by third parties and right-of-use assets obtained in exchange for financially settled leases were RMB 2,985 million.

Looking ahead, we aim to capitalize on the explosive growth in demand by further investing in infrastructure, enhancing service stability, managing liquidity risk and improving operating efficiency. We remain focused on AI-driven strategy, providing customers with high-value-added cloud services.

That's all for the introduction of our operational and financial results. Thank you all.

This concludes our prepared remarks. Thank you for your attention. We are now happy to take your questions. Please ask your questions in both Mandarin and English, if possible. Operator, please go ahead.

[Operator Instructions] We will take our first question, and the question comes from Liping Zhao from CICC.

[Interpreted] Congrats on another strong quarter. So I have 2 questions. The first one is relating to your StarFlow MaaS platform. Mr. Zou mentioned that the revenue of the MaaS platform increased 53x from January to April. Could you share the current revenue scale and margin levels? And what's management's outlook on this business?

And the second question is about the AI pricing compared to the fourth quarter of 2025, have there been increases in the average pricing for the newly signed public cloud contracts in the first quarter and second quarter of this year? If so, by how much?

[Interpreted] So our token business actually started off at a relatively small base, let's say, end of last year and the beginning of this year. However, traditionally, we have already maintained a very strong customer base of very large-scale leading customers.

So at the beginning of this year, we're starting to meet huge demand coming from such customers in light of the surge of agent demand, the surge of AI coding demand in such use cases. So obviously, the demand was huge and very strong, but it's generally -- but in a way, our business was restricted by the underlying resources that is available to us.

So I would say that, obviously, we're optimistic about the growth of this business. However, we need to -- due to that uncertainty we just mentioned, we would like to, let's say, see a couple more quarters before disclosing more details to the market of how that business grows.

And secondly, in terms of margin levels, I would say that the margin levels for this token business, inference business in general is higher than traditional cloud computing business. And we do see a lot of improvement in margin from a certain perspective, for example, coming from technology advancement, coming from the optimization of algorithms, coming from the optimizing that algorithms with the relevant models and also optimization and improvement coming from operating models.

So I would say we're cautiously optimistic about the margin. But again, due to still currently in a quickly expanding phase, it's not in a static phase. So we will not, at this stage, talk more about the specific margin numbers.

[Interpreted] So in relation to your second question about the selling and purchasing price in relation to our business. So yes, as far as we recognize, the demand for our cloud computing services have been surging tremendously. So is the pricing from our upstream, which includes from components to holistic servers to other raw materials. So the current consensus of the market, including our customers, is that the price hiking, the price surge trend will actually continue, not only already happened in Q1, but also will continue in Q2 and maybe some quarters to follow.

So they would believe that the current time point to secure more computing power is actually the right moment to do so. And because of that, the pass-through of the price -- of the cost pressure coming from our upstream is actually doable and would not affect our -- it would not negatively affect our margin levels in this current market situation.

Your next question comes from the line of Wenting Yu from CLSA.

[Interpreted] The first question is regarding the gross margin in the first quarter. We noticed that the gross margin dropped a bit in first Q, and what are the main reasons? Has the positive effect of the product price increase has been factored in already?

And the second question is with ongoing high demand for computing power, large model companies are adapting their resource allocation and forming partnerships beyond public cloud vendors to GPU rental companies and telcos. How does management think of cloud competitive position and advantages in this landscape?

Great question. For the first -- the first quarter margin, we can see a 3% percentage decrease. This is mainly -- I think the first factor is the seasonal factor because there is the 30% revenue come from the public -- enterprises cloud. So that is the first reason. And the second one is the upfront cost incurred for future revenue generating activities with certain customers. For the coming quarters, we expect the gross margin will recover to normal level.

[Interpreted] So this is a very interesting question. It's a very good question. So in fact, we observed a relatively interesting change in market landscape starting in the second half of 2025. That is some of the used to be competitors are actually becoming cooperation partners in this wake of strong AI demand.

And what you mentioned in your question is also -- it also exemplifies the close cooperation between private enterprises and state-owned enterprises. And we think that this demonstrates a couple of things. Number one, this is fundamentally a complementary capabilities from -- or I would say, complementary institutions coming from different backgrounds of enterprises, which we have mentioned back in the general computing, GPU computing age, and this is actually manifesting itself again in this new AI for intelligent cloud era.

And secondly, we think that this fundamentally reflects the strong discrepancy between -- or the strong gap, the big gap between supply and demand in today's market. So everybody actually needs to come together and to overcome shortcomings that each one of us have to form holistic and overall solutions to serve the end customers.

So again, your question is more from a competition perspective. But from our understanding and experience, we're actually seeing a more kind of cooperative perspective of the story. Everybody has the shortcomings and we work together to serve the needs of the end customers.

Your next question comes from the line of Daley Li from Bank of America Securities.

[Interpreted] I have 2 questions here. First question about the demand outlook for the public cloud. The 1Q result is pretty strong. And how do we see the demand trend in Q2 and the second half this year? Regarding the demand mix, how is the trend for inferencing and model training?

And the second question is about the contract term with our clients. As the upstream costs are in a rising trend, are we taking like a shorter-term contract -- any change in terms of the contract terms?

[Interpreted] So in terms of AI demand, as we mentioned, it's seasonally strong. And looking at the second quarter, we currently actually have a very long list of our backlog, which is mainly subject to the supply chain restrictions.

Now for the sectors that drive this growth, we're currently covering quite a few sectors, which we have all seen simultaneously having explosively strong demand, starting from the Internet, from large language model AI labs, from autonomous driving and from robotics. And I would say that out of which the autonomous driving and robotics generally tend to have very strong model training requirements and demand. And particularly for robotics, they have also -- also for autonomous driving, actually, they have also a lot of data processing or data treatment requirements demand for the training of their models.

And for the inference side of the story, we have the Internet companies and large language model companies coming from their demand for AI coding and agents such use cases. So this is a general overview of the demand side of things.

Now since you also mentioned about the contract period, we had relatively standard contract periods in the past. But then now in light of this new -- the supply chain, the price surge, we currently have more flexible kind of contract period arrangements, which maximizes our profit and benefits.

Your final question comes from the line of Timothy Zhao from Goldman Sachs.

[Interpreted] Congrats on the very strong results. My first question is regarding your revenue from Xiaomi and Kingsoft Cloud -- and Kingsoft Ecosystem. I noticed that the revenue growth accelerated compared to the fourth quarter of last year. Just wondering especially related to Xiaomi after the MiMo large language model launch and especially MiMo V2.5, have you observed any specific change on the demand for Kingsoft Cloud resources and the breakdown between training and inferences? And after you announced the [indiscernible] transaction -- revenue cap with Xiaomi for this year and next, just wondering how do you think about the utilization rate versus last year?

And my second question is regarding your CapEx and also lease assets. I noticed that the total amount spent was around RMB 3 billion in the first quarter. Just wondering if you can provide us an update on how you think about this total CapEx number for this year?

[Interpreted] So on the training, the vast majority of those resource demands come from Kingsoft Cloud. We continue to see growth -- growing demand in that respect. In inference demand, especially since the launch of V2. MiMo V2, a lot of underlying resource has been reallocated to do the training for that -- to do the inference for that model.

Now in terms of the prediction for future growth of inference coming from MiMo, we are relatively optimistic. However, it's ultimately subject to the return on Xiaomi's side. So probably we don't have to personally comment on that, but we generally remain optimistic about that.

For CapEx in the future, actually, the AI era present huge opportunities for us. Fortunately, we launched our intelligent computing business back in 2023. We have well established our supply chain capabilities, and our supply network is in place now. As Mr. Tao Liu has mentioned, we have seen a certain demand from our strong customer demand. But we have to admit the supply chain capacity is the primary limiting factor for the capital spending for the 2026. And we estimate that our base estimate to say for the 2026 is around RMB 15 billion to RMB 20 billion at this moment.

This concludes today's question-and-answer session. I will now hand back for closing remarks.

Thank you, operator. Thank you once again for joining us today. If you have any further questions, please feel free to contact us. Looking forward to speaking with you again next quarter. Have a nice day.

This concludes today's conference call. Thank you for participating. You may now disconnect.

[Portions of this transcript that are marked [Interpreted] were spoken by interpreters present on the live call.]

Good day, and thank you for standing by. Welcome to the Kingsoft Cloud's Fourth Quarter and Full Year 2025 Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded.

I would now like to hand the conference over to your speaker today, Nicole Shan, IRD of Kingsoft Cloud. Please go ahead.

Thank you, operator. Hello, everyone, and thank you for joining us today. Kingsoft Cloud's fourth quarter and fiscal year 2025 earnings release was distributed earlier today and is available on our IR website at ir.ksyun.com as well as on PR Newswire services.

On the call today from Kingsoft Cloud, we have our Chairman and CEO, Mr. Zou tao; CFO, Li Yi; Senior Vice President, Mr. Liu Tao; Senior Vice President, Mr. He Haijian; Vice President, Ms. [ Wang Ja ]; and Associate Vice President, Mr. Kaiyan Tian.

Mr. Zou will review our business strategies, operations and other company highlights, followed by Ms. Li, who will discuss the financial performance. We will be available to answer your questions during the Q&A session that follows.

We will be conducting an interpretation. Our interpretation are for your convenience and reference only. In case of any discrepancy, management statement in our original language will prevail.

Before we begin, I'd like to remind you that this conference call contains forward-looking statements within the meaning of Section 21E of the Securities Exchange Act of 1934 [indiscernible] U.S. Securities Litigation Reform Act of 1995. These forward-looking statements are based upon management's current expectations and current market and operating conditions, [indiscernible] known or unknown risks, uncertainties and other factors, which are difficult to predict and many of which are beyond the company's control, which may cause the company's actual results, performance or achievements to differ materially from those in the forward-looking statements.

Further information regarding these and other risks, uncertainties or factors are included in the company's filings with the U.S. SEC. The company does not undertake any obligation to update any forward-looking statements as a result of new information, future events or otherwise, [indiscernible] as required under applicable law.

Finally, please note that unless otherwise stated, all financial figures mentioned during this conference call are denominated in RMB.

It's now my pleasure to introduce our Chairman and CEO, Mr. Zou. Please go ahead.

[Interpreted] Hello, everyone, and thank you, and welcome to Kingsoft Cloud's Fourth Quarter and Fiscal Year 2025 Earnings Call. I am Zou Tao, CEO of Kingsoft Cloud.

Since the beginning of 2025, the global AI industry has reached a series of milestones from the democratization sparked by the DeepSeek moment to the active competition among multimodal models, from the leap of embodied AI into the physical world to open close loop capability of understanding and execution. AI is evolving with unstoppable momentum, linking across models, agents, computing power to industrial applications, reshaping every sector.

As a tightly integrated component of the AI 5-layer take, cloud computing is now meeting an unprecedented surge in demand for intelligent computing. This year, we stay committed to our high-quality and sustainable development strategy, embracing the opportunities in AI era, strengthening our capability through solid execution. We have delivered impressive results, achieving strong financial performance while forging lasting business trends.

First, we recorded a historical high quarterly revenue, reaching RMB 2.76 billion, representing a year-over-year growth of 22%, among which revenues from public cloud services increased by 35% to RMB 1.9 billion. Our intelligent computing services keep driving our growth. The gross billing of AI business reached RMB 926 million, representing a 95% year-over-year and contributing 49% of our public cloud services.

Second, growth in our ecosystem and external business segment is progressing hand in hand. On one hand, our ecosystem partnerships remain strong and continue to deepen. This quarter, Xiaomi and Kingsoft ecosystem revenue reached RMB 804 million, a 63% year-over-year increase, accounting for 29% of total revenue. For the full year 2025, related party transactions with Xiaomi and Kingsoft ecosystem partners reached 94% of our net annual cap, almost hitting the limit.

On the other hand, our external customers including leading enterprises across a wide range of high-growth industries also shown confidence in our products and services accounted for around 70% of total revenue. Furthermore, revenue from our top 5 nonecosystem customers grew by 44% year-over-year, sustaining strong growth momentum.

Last but not least, profitability continued to improve this quarter with adjusted gross margin increasing quarter-over-quarter was 17.1% and adjusted operating margin reaching 2.0%. We have achieved operating level profitability 2 quarters in a row, and our self-funding capability has shown sustained and significant year-over-year improvement.

Now I would like to walk you through the key business highlights for the fourth quarter of 2025. In terms of public cloud services, revenue reached RMB 1.9 billion this quarter, representing a year-over-year increase of 35%. From a customer perspective, in 2025, AI continued pushing its boundaries, driving it to fully embrace it, diversifying our customer base.

Beyond leading AI enterprises and Internet giants, we now also serve automotive manufacturing, autonomous driving, embodied AI and fintech sectors, et cetera. We've solidified our cooperation within the Xiaomi and Kingsoft ecosystem while capturing new external opportunities.

From product and services perspective, we keep pushing the limits of cluster scale, supporting large-scale training and explosive inference demand. Notably, in this quarter, we delivered a new inference cluster for a top video streaming platform, serving over 100 million users. We also secured a major fintech customer using our token-based inference service, who speak highly of our stable model and consuming power services.

On supply chain front, despite market uncertainties, our well-established and resilient supply chain built through years of experience allowed us to plan ahead strategically and stock key components dynamically to ensure sustainable business growth.

Now in terms of enterprise cloud, revenue reached RMB 859 million this quarter, a significant quarter-over-quarter increase of 18%. Driven by the AI Plus policy, industrial intelligence solutions have become a key growth driver. The demand for specialized vertical models, real-world application and strict data compliance makes cloud services more essential than ever in advancing industrial Intelligence.

The AI business of enterprise cloud is paving the way for steady long-term growth, not only representing $1 trillion market opportunity, but also playing a critical role in driving the technological leap across industries. As a B2B cloud service provider with solid technology expertise and enterprise service capabilities, we are well positioned to capture these industrial transformation opportunities.

In the area of enterprise services, we achieved key breakthroughs in high-end manufacturing industry. We provided stable and high-performance computing service to the top enterprises to support their process in intelligent manufacturing, industrial vision and AI R&D.

In health care space, we launched a data agent-based AI application in health care intelligent operation process, marking a paradigm shift from digitalization to intelligence. This analysis through natural dialogue platform enables natural language insights into DRG cost control, moving hospital management from retrospective specifics to proactive intervention. While significantly lowering the barriers to data application, we have further solidified our technical moats and differentiated competitive advantages in high-value medical AI scenarios.

In public services area, we partnered with telecom operators to provide sustainable and stable high-performance computing clusters for the public services sector, successfully entering key markets like Shanghai. We believe that by leveraging Kingsoft Cloud's deep vertical expertise and enterprise service experience, intelligent computing opportunity in the enterprise cloud segment represents a massive industrial frontier, generating synergies with our public cloud business.

In terms of products and technology, we are building a next-generation computing services system for LLM training, inference and industrial intelligence, offering full stack capabilities from computing services to Model as a Service. Our technology upgrades from basic cloud computing to an AI-first, AI-native cloud architecture contributing for digital and intelligent transformation across sectors. We focus on the technologies catering to model training and inference scenarios aiming to provide highly stable, highly efficient and ready-to-use intelligent computing services. This quarter, our StarFlow platform keeps upgrading with the launch of MCP, aka, model contact protocol, product optimization and AI search features to help enterprises develop and deploy AI agents through a unified platform building a new ecosystem centered around agent-based operations. For enterprises with private deployment demand, our Galaxy Stack provides heterogeneous GPU management, network and intelligent container scattering capabilities. We also feature full stack localization with indigenous adaptation to empower intelligent transformation across verticals.

Standing at the new starting point, looking ahead, we're truly excited by the limitless possibilities that lie before us. We will remain committed to our high-quality and sustainable development strategy. By embracing the immense opportunities presented by the AI era, developing along with the industry and refining our core technology, we will continue to capture the market opportunities both within and beyond our ecosystem, optimizing the operations of our assets to enhance profitability and thereby create value for our customers, shareholders, employees and society.

I will now pass the call to our CFO, Ms. Li Yi, to go over our financials for the fourth quarter and fiscal year 2025. Thank you.

Thank you, Mr. Zou and Clark, and thank you all for joining the call today. Before we walk through the details of financial results for the fourth quarter and fiscal year 2025. I would like to highlight the following aspects.

First, our revenue has achieved record high, RMB 2,761 million this quarter, representing a year-over-year growth rate of 24%. Within that, revenue from public cloud services was RMB 1,902 million, increased by 35% from RMB 1,410 million in the same quarter last year. Unprecedented explosive demand for our AI business grew 95% year-over-year billings growth, which totaled RMB 926 million.

Second, profitability has seen substantial improvement. Driven by shift in our revenue structure, our adjusted gross margin continued its upward trend, rising to 70% from 60% in the previous quarter. Adjusted EBITDA margin reached 28%, up 12 percentage points from 16% in the same quarter last year, though down from 33% last quarter. The year-over-year growth was fueled by a large contribution from AI-related business, while [indiscernible] represents the primary cost component.

The sequential decrease was mainly due to a loan recurring subsidiary received last quarter, which established a high baseline. Notably, we have achieved adjusted operating profit for 2 consecutive quarters, reaching RMB 55 million this quarter, which was a 2% margin. These results validate our ability to monetize intelligent cloud opportunities and our strategic focus on high-quality enterprise services.

Third, our cash and cash equivalent achieved RMB 6,018 million, raising our ability to further support the investment in the AI business.

Now I will walk you through our financial results for the fourth quarter of 2025. This quarter, total revenue were RMB 2,761 million. Of these, revenues from public cloud services were RMB 1,902 million up 35% from RMB 1,410 million in the same quarter last year. Revenues from enterprise cloud services reached RMB 859 million due to seasonally strong quarters, which was characterized by a high volume of project completion.

Total cost of revenues was RMB 2,296 million, up 27% year-over-year, which was mainly due to our investments into infrastructure to support intelligent cloud business growth. IDC costs increased by 30% year-over-year from RMB 725 million to RMB 812 million this quarter. The increase was mainly due to the increasing mix of racks, which serves the expanding AI business.

Depreciation and amortization costs increased from RMB 323 million in the same quarter of 2024 to RMB 741 million this quarter. The increase was mainly due to depreciation of newly acquired and leased servers and network equipment, which were mainly allocated to our AI business. Solution development and service costs increased by 50% year-over-year from RMB 557 million in the same quarter of 2024 to RMB 642 million this quarter. The increase was mainly due to the solution personnel expansion. Fulfillment costs and other costs were RMB 461 million this quarter.

Our adjusted gross margin for the quarter was RMB 471 million, increase to 10% year-over-year and 20% quarter-on-quarter. It was mainly due to the expansion of our revenue scale, the enlarged contribution from AI business and the cost control of IDC racks and servers. Adjusted gross margin increased from 60% last quarter to 70% in this quarter, which was mainly due to the high contribution from enterprise cloud.

On the expense side excluding share-based compensation costs, our total adjusted operating expenses were RMB 451 million, increased by 3% year-over-year and increased 9% quarter-on-quarter, of which our adjusted research and development expenses was RMB 181 million, increased by 7% from same quarter last year. Adjusted selling and marketing expenses were RMB 111 million, increased by 3% year-over-year. Adjusted general and administrative expenses were RMB 168 million, decreased [ 1% ] year-over-year.

Our adjusted operating profit was RMB 55 million, increased by 124% for adjusted operating profit of RMB 24 million in the same period last year. The improvement was mainly due to the expansion of our revenue scale and gross profit as well as the expense control. The total expense as a percentage of revenue keeps decreasing. Adjusted operating profit margin increased from 1% in the same period last year to 2% this quarter.

Our non-GAAP EBITDA margin was RMB 785 million, increased by 180% from RMB 360 million in the same quarter last year. Our non-GAAP EBITDA margin achieved 28% compared with 60% in the same quarter last year. It was mainly due to our strong commitment to AI cloud computing development, taken adjustment of business structure, strong control our costs and expenses.

This quarter, our capital expenditures, including those financed by third parties and right-of-use assets of [indiscernible] finance lease liabilities were RMB 496 million.

For the full year 2025, our total revenue achieved landed at RMB 559 million, increased by 23% from RMB 785 million in 2024, among which revenues from public cloud services were RMB 6,634 million, increased by 33% year-over-year. Revenues from enterprise cloud services was RMB 2,925 million, increased by 5% year-over-year. Adjusted gross profit was RMB 1,542 million, increased by 40% from RMB 1,358 million last year. Adjusted gross margin was 60% decreased from 70% in last year, which was mainly due to the high cost for servers and other hardware equipment. Adjusted operating cost was RMB 152 million, narrowed significantly from RMB 431 million. Adjusted operating profit margin was minus 1.6% from minus 5.5% last year. Adjusted EBITDA profit was RMB 2,336 million, increased by 366% from RMB 631 million last year. The adjusted EBITDA margin was 24%, improved by 60% from 18% last year.

Looking ahead, we aim to capitalize on the explosive growth in demand and further investing in infrastructure, enhancing service stability, managing liquidity risk and improving operating efficiency. We remain focused on AI strategy, providing customers with high-value added cloud services.

That's all for the introduction of our operational and financial results. Thank you all.

Thank you. Operator, this concludes our prepared remarks. We are now happy to take your questions. Please ask your question in both Chinese and English, if possible. Operator, please go ahead. Thank you.

[Operator Instructions] Our first question comes from the line of Liping Zhao from CICC.

[Foreign Language] Mr. Zou and Ms. Li, congrats for the very good 4Q results. I have 2 questions here. First, Xiaomi recently launched the MIMO B2 Series models, which have received positive market feedback how should we view our role and positioning within Xiaomi's AI strategy? And what strategies will be implemented around Xiaomi and Kingsoft service going forward?

And secondly, how does the management view the current pricing uptrend in the cloud service industry? Has the company already adjusted prices for AI computing services. Or are there any relative plans in place? To what extent are those price adjustments driven by demand or driven by the upstream procurement cost pass-through?

[Interpreted] Okay. The answer comes from our CEO, Mr. Zou Tao. So a little bit of background. So back in 2024, I think that was in August, we had an internal discussion around the development model for the whole Xiaomi and Kingsoft ecosystem. So the idea was that the whole Xiaomi-Kingsoft ecosystem will form a portfolio of solutions where -- a whole system where Kingsoft will stay disciplined and not really developing our own last language models, which is for Xiaomi to develop. So the MIMO model, and it's widely recognized performance, it's actually an implementation and manifestation of our overall AI strategy within the Xiaomi-Kingsoft ecosystem.

And secondly, back in 2025, so 1 year later, from the internal discussion session from a KC perspective, we formed a strategy that's called 1+N. So the 1 here actually refers to the Xiaomi MIMO model, which is the key to KC's inference strategy. So in the future, we will continue to adhere to the strategy, which essentially means that within the ecosystem, we will continue to serve the Xiaomi and Kingsoft ecosystem. And for external customers, we will also try to monetize our Model as a Service capabilities, thereby not only in the training area that we were able to make our revenue and profit, but also will make our contribution in the inference era that is approaching.

[Interpreted] So the answer comes from our SVP, Mr. Liu Tao. So a bit of background again. So in the Q3 of last year, we had anticipated the significant pricing of increase from the supply chain side. And therefore, we have dynamically and strategically stocked up some of the key components. So we did have -- so we were actually prepared for this which was unfolding today.

Now in terms of the price hike that you were asking, so we stick to two principles. Number one, for some of the customers and the business where we already have contracts in place and where we have the stocking of the underlying resources, we tend to not increase the pricing. However, for some of the new customers, new contracts, especially with significant increase of usage, there's going to be significant price hiking in these kind of scenarios.

Now also in terms of profitability, one thing is that we will actually try to pass through some of the upstream cost increases to our customers. And secondly, depending on the demand, right, we'll also try to increase some of the price to reflect and increase our profit.

Our next question comes from the line of Wenting Yu from CLSA.

[Foreign Language] The first question is that some of your cloud service partners have announced their cloud business more towards approach from the traditional server rental and also the subscription model. So will KC adopt a similar strategy? And how do you view the impact of this trend on industry competition and more for profit margins?

And the second question is regarding the impact from the Volcano Engine. It is adopting a relatively low price strategy. And how do we view the impact on the industry and potentially our business this year.

[Interpreted] Okay. So regarding your question on the shifting to Model as a Service strategy, we have noticed some of the other peer companies who have released their results earlier than I mentioned. However, my view is that this is not actually some new concept. It is actually one of the inevitable stage of the development of AI as well as large language model from the training that we do to create them to a certain state that they become applicable and workable in our day-to-day work and life.

So in relation to our own inference related work -- model as a service work, we actually launched the StarFlow platform as we mentioned in the prepared remarks last year. And because we are a neutral platform, we were able to host essentially all of the open source models, including all the models from Xiaomi for some to provide Model as a Service business. where this is essentially actually the fastest growing business in the history of the company.

Actually, so we talked about Xiaomi MIMO model earlier. The way that we're providing services for Xiaomi MIMO model is also a Model as a Service business. And also for some of the large language model customers that we use to -- and we're still providing training services to them, we also provide the Model as a Services business to them as well to cater their inference needs.

Now to your second question about the price change for Volcano Engine, I haven't really noticed that particular piece of news. However, the general market dynamics today is that, on one hand, we're seeing explosive growth on the demand side. And we're seeing particularly high price hiking from the supply chain side. So I do not personally think that under such circumstances, changing price to a lower level would actually be implemental and applicable in the real world.

Now what I have focused more is the price hiking information from for example, Ali Cloud. We have worked with them together. We have been in the industry together for many years, and this is the first time that we've seen them hiking their price. And also in addition from our, SVP, Mr. Liu Tao, there is a difference between the catalog price and the actual price that the companies that, us, as cloud players, and our customers engage into. So the change in catalog price is more of a marketing kind of purpose. And it does not necessarily mean the actual price the company to enter into business.

Our next question comes from the line of Timothy Zhao from Goldman Sachs..

[Foreign Language] My first question is on your financial outlook. Just wondering if you can share some color how we should think about the revenue, EBITDA, operating profit, growth outlook for this year? And also on the CapEx expenditure plan, what is your thoughts considering the balance sheet and also the prepayment from certain customers. Do you think it's possible to further raise your CapEx plan given the rising AI demand?

And secondly, it's is regarding the third-party revenue in the AI outlook. I'm just wondering if you can share more detailed color on what specific products or what type of customers are driving the third-party AI growth. And also, what is the breakdown and outlook between the mix of AI training versus AI inferences.

All right. I will take the CapEx first. For 2026, we expect total CapEx and [ controlled assets ] to exceed RMB 10 billion, represent expansion from 2025 level. On funding structures, we expect approximately half our CapEx is to be covered by customer prepayment arrangement. We will significantly reduce fund requirements. Additionally, we plan to access more assets through short- and long-term leases with payment structure and operating cash flows to minimize us from capital [ incompetence ].

For the funding projection and financing lease, we currently have no equity finance plan. 2026 capital expenditure at 3, 4 channels: first, proceeds from our 2025 financing; the second, customer operating receives; and the third, strategic customer prepayment; and the fourth, commitment created facilities from banks and financial lease institutions. Incremental resource requirements will be made primarily through leasing to preserve balance sheet flexibility. For the outlook for the guidance for 2026, we expect our growth rate will be accelerating, and the pace of he EBITDA rate will improve much better in 2026 as well.

[Interpreted] So if you look at the past results as discussed in the prepared remarks, so the top 5 nonecosystem customers combined revenue on a year-over-year basis revenue growth was 44%, which is really strong growth. So those would include Internet companies, autonomous riding and robotics. And then in terms of looking forward into the year of 2026, we do see extremely large demand coming from outside of the ecosystem and, to some extent, that such demand is actually higher than the demand from our ecosystem. So the final wrapping of financial results coming from that demand will actually be dependent on how much resources we're able to secure and deliver to such customers.

Now from the perspective of products and solutions, we're actually seeing more than half of the potential demand coming in for inference versus training. And then for the StarFlow platform, which we discussed earlier, it's growing really fast in that business. And we're seeing better profit margins coming from that particular business. And this is a result, of course, from the very good application, a very good application and increasing penetration for agents and applications.

Due to time constraint, this concludes our Q&A session. Thank you once again for joining us today. If you have any other questions, please feel free to message us. We look forward to speaking with you again next quarter. have a nice day.Thank you all.

This concludes today's conference call. Thank you for participating. You may now disconnect.

[Portions of this transcript that are marked [Interpreted] were spoken by an interpreter present on the live call.]

Good day, and thank you for standing by. Welcome to Kingsoft Cloud's Third Quarter 2025 Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to your speaker today, Nicole Shan, IR Director of Kingsoft Cloud. Please go ahead.

Thank you, operator. Hello, everyone, and thank you for joining us today. Kingsoft Cloud's Third quarter 2025 earnings release was distributed earlier today and is available on our IR website at ir.ksyun.com as well as on the PRNewswire Services.

On the call today from Kingsoft Cloud, we have our Vice Chairman and CEO, Mr. Zou Tao; and CFO, Ms. Li Yi. Mr. Zou will review our business strategies, operations and other company highlights followed by Ms. Li, who will discuss the financial performance. They will be available to answer your questions during the Q&A session that follows.

There will be consecutive interpretation. Our interpretations are for your convenience and reference purposes only. In case of any discrepancy, management statement in original language will prevail.

Before we begin, I'd like to remind you that this conference call contains forward-looking statements within the meaning of Section 21E of the Securities Exchange Act of 1934 as amended and as defined in the U.S. Private Security Litigation Reform Act of 1995. These forward-looking statements are based upon management's current expectations and current market and operating conditions and relate to events that involve known or unknown risks, uncertainties and other factors, all of which are difficult to predict and many of which are beyond on the company's control, which may cause the company's actual results, performance or achievements to differ materially from those in the forward-looking statements.

Further information regarding these and other risks, uncertainties or factors are included in the company's filings with the U.S. SEC. The company does not undertake any obligation to update any forward-looking statements as a result of new information, future events or otherwise, except as required under applicable law. Finally, please note that unless otherwise stated, all financial figures mentioned during this conference call are denominated in RMB.

It's now my pleasure to introduce our Vice Chairman and CEO, Mr. Zou. Please go ahead.

[Interpreted] Hello, everyone. Thank you, and welcome to Kingsoft Third Quarter 2025 Earnings Call. I am Zou Tao, CEO of Kingsoft Cloud. In the era that artificial intelligence is implemented across various industry verticals, and reshaping the technological landscape, Kingsoft Cloud has firmly established its strategic positioning and define its development orientation. On the premise of steadily meeting the demand of model training, we have made adequate technical and resource reserves for the explosive growth of inference. In the face of the dual trends of rapid model iteration and increasing adoption of artificial intelligence, we have provided our clients with stable and efficient integrated training and inference intelligent cloud computing services and have laid out model API business to turn inference scenarios into new growth entrants. The substantial high growth in revenue and a stable profit margin level validate the steady execution of our strategic measures, achieving high quality and sustainable development.

First, our revenue in the third quarter reached RMB 2.48 billion with year-over-year growth rate accelerating from 24% in the previous quarter to 31% this quarter. Both public cloud and enterprise cloud achieved year-over-year and sequential growth, among which public cloud revenue increased significantly by 49% year-over-year, reaching RMB 1.75 billion.

Second, intelligent computing cloud business remains on fast development track. This quarter, gross billings of intelligent computing reached RMB 782 million, with a year-over-year growth of around 122%. It accounted for 45% of the public card revenue, realizing a significant increase from 31% in the same period last year. Generative artificial intelligence and cloud are symbiotically integrated in many aspects including technology, products and customer cross-sells. The demand for artificial intelligence not only drives the rapid development of intelligent computing cloud, but also leads to the growth in technological innovation of basic public cloud and accelerates the iterative process of cloud computing technologies. From 20 clusters native technology, our computing power services model API services, storage services and data services have all been upgraded.

Third, the Xiaomi and Kingsoft ecosystem continued to offer solid foundation. This quarter, revenue from the Xiaomi and Kingsoft ecosystem reached RMB 691 million, increasing by 84% year-over-year, and its proportion in the total revenue further rose to 28%. From January to September 2025, the total revenue from the Xiaomi and Kingsoft ecosystem reached RMB 1.82 billion. We anticipate adequately fulfilling the business cooperation under the continuing connected transaction annual quota this year and are optimistic in the further increase of the quota next year.

Finally, our adjusted gross profit for this quarter reached RMB 393 million, representing a year-over-year increase of 28%. The adjusted operating profit turned from loss to profit, reaching RMB 15.33 -- reaching RMB 15.36 million, and the adjusted operating profit margin was 0.6%. The adjusted net profit recorded a historical positive profit of RMB 28.73 million for the first time. The company is aiming at both revenue growth and profitability improvement as the economies of scale is becoming increasingly prominent. While accelerating the construction of intelligent computing infrastructure and technological capabilities, we are also strengthening the control of costs and expenses.

Now I would like to walk you through the key business highlights for the third quarter of 2025. In terms of public cloud services, revenue reached RMB 1.75 billion in this quarter, making a year-over-year increase of 49%. The intelligent computing cloud business has maintained strong growth. We have successfully supported the large-scale training and inference demand of various top Internet customers providing high-quality, high-performance, high stable and highly efficient cost of [ inference ] services. Especially from many artificial intelligence and Internet enterprises facing the simultaneous demand for model training and inference, we have provided customers with stable and integrated intelligent computing services for different scenarios.

Meanwhile, we actively expanded customer coverage and the cross-selling of intelligent computing cloud and basic cloud. In terms of ecosystem partners, we continued to provide high-quality services to Xiaomi and Kingsoft and continue to prepare underlying resources for ecosystem customers to enhance the rapid expansion capability for intelligent computing demand. In terms of enterprise cloud services, revenue in the quarter was RMB 730 million. We firmly believe that in today's rapidly evolving generative artificial intelligence landscape, intelligence will evolve from model capabilities to industry solutions, empowering and reshaping diverse sectors of the economy. As the indispensable carrier for intelligent computing, cloud services enjoys tremendous potential for such digitalization and intelligentization.

In this $1 trillion sustainable expanding market, we have deeply explored our inherent DNA of 2D enterprise services, targeted advantageous selected verticals and geographical regions and build core competitiveness for the future. As a result, it has received widespread recognition from our customers in the broader markets. For example, in the public services sector, we aim to become the preferred cloud partner for intelligent computing in the public services agencies and enterprises for their inference demand. Taking [ Xinyang ] City in [ Gansu ] province as an example as one of the 8 major nodes of the national project is [ Data West ] computing and a central area for intelligent computing business. We will be responsible for building the public services cloud platform in Xinyang to fully empower the local public services affairs with intelligence and digitalization.

In the field of health care, we have achieved a milestone breakthrough in a project of integrating artificial intelligence with traditional Chinese medicine clinical scenario, whereby not only have we achieved a deep integration of traditional Chinese medicine theory and artificial intelligence, [ seeing ] the commodity position in chronic disease management technology, but we have also verified the practical value of artificial intelligence in improving patients' quality of life and disease control rate at the clinical level.

In enterprise services sector, following the successful implementation of a landmark project for intelligent generation of bank credit reports, we continue to advance the intelligentization transformation across the entire credit approval process. This evolution extends from the single function of credit report initiation to a comprehensive intelligence system, including customer onboarding, credit report generation, loan disbursement, monitoring and early warning and post-loan reporting. We firmly believe that this proven accumulated successful experiences, market reputation and replicable core of solutions will enable us to seize a pioneering position in the emerging industry way, build a solid core competitiveness and achieve high-quality and sustainable shareholder returns.

In terms of product and technology. In public cloud space, we continue to enhance the technology of intelligent computing cloud this quarter, strengthening the capability of the [ Staff ] platform and made significant progress in the following 3 aspects. First, we have launched our model API service, delivering highly available and easily integrated capabilities for model invocation and management, laying a solid foundation for the subsequent provision of diverse model service paradigm.

Second, we upgraded our online model services, integrating multiple open source foundation models and equipped with automatic scaling capabilities, offering a highly available inference platform. Third, we launched our data annotation and data set marketplace aiming to provide customers with end-to-end support for data flow and help them efficiently advance the model training process.

In enterprise cloud space, in order to meet the demand for private deployment scenarios we have built a computing power to [indiscernible] platform, a lightweight mass platform and generative artificial intelligence knowledge base, and we have closely collaborated with [ WPS ] AI to build a trusted intelligent product architecture for public services use cases. Meanwhile, through the organizational development of the dual R&D centers in Beijing and Wuhan, we attract talent from various regions, build a talent pipeline and maintain sustained investment intensity in the intelligent computing field.

As of the end of Q3, the number of employees in Wuhan is 2.8x the head count back in 2022 when we first launched our Wuhan strategy. Overall, we will firmly seize the historic opportunities presented by the Xiaomi and Kingsoft ecosystem, continue to invest in infrastructure, focus on refining core products and solutions and to create long-term value for our customers, shareholders, employees and other stakeholders.

I will now pass the call over to Ms. Li Yi, our CFO, to go over our financials for the third quarter of 2025. Thank you.

Good evening and good morning, everyone, and thank you all for joining the call today. Before we walk through the details of financial results for the third quarter, I would like to highlight the following aspects. .

First, our revenue has achieved year-on-year growth for 6 quarters, reaching RMB 2,478 million this quarter. This represents an excellent year-over-year growth rate of 31%, up from 24% in the previous quarter. Revenue from [indiscernible] service stood at RMB 1,752.3 million, a significant increase of 49% from RMB 1,165.5 million in the same quarter last year. Meanwhile, robust demand from our intelligent cloud, which also called as AI cloud business drove around 120% year-over-year billing growth which totaled [ RMB 282.4 million ].

Second, profitability has seen substantial improvement. Our adjusted gross margin rose to 16%, up from 50% in the previous quarter. And adjusted EBITDA margin improved to 33% compared with 17% last quarter. Notably, returned quarterly adjusted operational and adjusted net loss into profit simultaneously for the first time. These gains validate our strong execution and pursue high-quality, sustainable development as well as our ability to monetize opportunities in the intelligent cloud space.

Third, we would like to express our gratitude to shareholders for the support during our recent equity financing in September. We successfully raised HKD 2.8 billion and 8% of the fund will be allocated to further investments in air infrastructure and 12% to general operational leads. This funding will fully underpin the growth of our intelligent cloud business and enable us to create long-term value for all stakeholders.

Now [indiscernible] grew on our financial results for the third quarter of 2025 and I use RMB as currency. Total revenues were RMB 2,478 million. Of these, revenues from public cloud services were RMB 1,652.3 million up 49% from RMB 1,475.9 million in the same quarter last year. Revenues from enterprise cloud services reached RMB 725.7 million compared with RMB 710 million in the same quarter last year. Total cost of revenues were RMB 2,097.1 million, up 33% year-over-year, which was mainly due to our investment into infrastructure to surpass the intelligent cloud business growth. IDC costs increased by 15% year-over-year from RMB 673.8 million to RMB 7255.7 million this quarter.

The increase was mainly due to the purchase of REC, which serves as expanding [ intention ] cloud business as well as the base computing and storage cloud demand, brought by AI development. Depreciation and amortization costs increased from [indiscernible] million in the same quarter of 2024 to RMB 649.7 million this quarter. The increase was mainly due to the depreciation of newly acquired and listed servers and later work equipment, which were mainly allocated to intelligent cloud business.

Solutions development and services cost increased by 19% year-over-year from RMB 499 million in the same quarter of 2024 to [indiscernible] this quarter. The increase was mainly due to the solution personnel expansion. Fulfillment costs and other costs were RMB 502 million and RMB 70.6 million this quarter. Our adjusted gross margin for the quarter was RMB 392.6 million, increased by 28% year-over-year and 12% quarter-over-quarter. It was mainly due to the expansion of our revenue scale, the [indiscernible] contribution from intelligent cloud and the cost control of IDC RECs and servers. Adjusted gross margin increased from 50% last quarter to 60% in this quarter.

On the expense side, excluding share [ concession ] costs, our total adjusted operating expenses were RMB 420.9 million, decreased by 70% over year-over-year and 25% quarter-over-quarter. Of which, our adjusted R&D expenses were RMB 188.4 million, decreased by 90% from same quarter last year. The decrease was mainly due to the decrease of personnel costs resulting in our strategic adjustments for research team as well as the expense serving from Beijing, Wuhan research and strategy.

Adjusted selling and marketing expenses were RMB 127.6 million increased by 15% year-over-year. Adjusted general and administrative expenses were RMB 104.9 million, decreased by 29% year-over-year due to the reversal of credit loss. The impairment of long lead assets was new this quarter compared with RMB 190.7 million in the same quarter last year. Our adjusted operating profit was RMB 15.4 million. Total profit from adjusted operating loss of RMB 140.2 million in the same period last year. The improvement was mainly due to the expansion of revenue scale and the gross profit, the expense control as well as the reverse of credit loss. Adjusted operating profit margin increased from minus 7% in the same period last year to 7.6% this quarter, representing an increase of 8 percentage points.

Our non-GAAP EBITDA profit was RMB 826.6 million, increased by 3.5x of RMB 185.4 million in the same quarter last year. Our non-GAAP EBITDA margin achieved 33% compared with 10% in the same quarter last year, mainly due to our strong commitment to intelligent cloud development, strategic adjustment of business structure, stricter control of our costs and expenses as well as the loan recover impact of subsidy in other income.

As of September 30, 2025, our cash and cash equivalents totaled RMB 3,954.5 million, decreased from RMB 5,464.1 million as of June 30, 2025. The decrease was mainly due to our infrastructure investment for intelligent cloud. This quarter, our [indiscernible], including those finance by third quarter and right of use assets of [indiscernible] finance lease liabilities was RMB 2,787.8 million.

Looking forward, AI technology drives the revolution of cloud computing. We do more than just fulfill the computing demand of model trailing inference. We also involve enterprises to evoke API and AI probabilities to the business. Stepping into the face of rapid development in AI applications and excessive growth in demand, we will further invest into infrastructure, strengthening technology, enhanced service stability and provide customers with high value-added cloud services.

That's all for the introduction of our operational and financial results. Thank you all.

Thank you. Operator, we are now due to start the Q&A session. Please answer your question in both Mandarin and Chinese and English, if possible. Operator, please go ahead. Thank you. .

[Operator Instructions] Our first question comes from the line of Xiaodan Zhang from CICC.

[Interpreted] First of all, what are the key drivers of AI revenue growth in Q3? And has there been any structural change in the demand of your ecosystem and external clients for the past quarter? And secondly, how does management see the margin trend in the coming quarters? And what's the expected mix of different computing resources acquisition models?

Please stand by while the speakers reconnect. Please stand by. Thank you. Speakers, you are now reconnected. Please go ahead.

Sorry, we didn't get your question. Could you repeat that again? Thank you.

Yes, no problem. So my first question is regarding the AI revenue. So could management break down the key drivers for AI revenue in Q3? And has there been any structural change in demand of your ecosystem and external clients for the past quarter? And secondly, how does management see the margin trend in the coming quarters? And what's the expected mix of different computing resources acquisition models going onwards?

[Interpreted] So basically, the core of the reason behind the AI revenue growth in Q3 is that we had some quarters that partially delivered in the previous quarters for example, like the second quarter of 2025. And these clusters and these services have only been partially accounted for revenues in -- from a full quarter basis. But now in Q3, they are starting to be recognized as full quarter revenues. And also, there's the factor of partially delayed revenue as well, some of the revenue which we had in Q2, but was not accounted for. And then this revenue are delayed into the third quarter.

So regarding the second part of your first question, which is about the structure of internal and external customers. I think I used to say that from a large trend -- general trend perspective, we're currently in the phase of transitioning from large and top customers training demand to general and wider spread customers' inference demand. Most -- at the current stage, we still see a majority of our demand coming from the larger customers in their training demand. However, especially in the latest quarter, we're increasingly seeing the trend of customers adopting artificial intelligence models in their diverse industries.

So in face of this general trend, we have also, as we mentioned in the prepared remarks, we have launched our [indiscernible] platform to meet the demand of such general trend. And this also goes back to the margin question that you also asked about. We generally think that the -- in the future, the inference demand will tend to exhibit higher margin profile than the current stage of training. And therefore, we think that when that wave of demand comes, we expect to have higher margins.

As EBITDA level, as the proportion of the AI business continues to rise and its cost structure is mainly dominated by depreciation, we expect this year's EBITDA margin will remain about 20%. But I have to mention that the significant quarter-on-quarter improvement this quarter was mainly driven by a onetime other income which will return to the normal level next quarter.

Our next question comes from the line of Wenting Yu from CLSA.

[Interpreted] The first question is, could management share the outlook and guidance on the revenue outlook for the next year and beyond the Internet [ COVID ] post model training and embody intelligence scenarios that are already underway this year, which other industry application scenarios are expected to have strong computing power demand that could drive the revenue growth next year?

And the second question is with multiple cloud providers in both China and U.S. increasing the proportion of server leasing in their computing resource mix? And how does management view the current market dynamics for procurement versus leasing and from a cost-effectiveness and profit margin perspective, how would the company allocate the resources between these 2 approaches?

Thank you for your question. The company's budget process is currently underway and expected to be completed around the beginning of the next year. We will show the specific details with you once it was finalized. However, regarding the demand for our AI business, we are fully confident in the subsequent demand growth.

And for your second question regarding the procurement methods, we primarily like our capital channels with actual customer needs including collector scale, delivery time and supply inventory level. There is no rigid total allocation target from the cost-effective perspective, both approaches have their own product and cost. The leasing model expands our supply chain channels and provide certainty with ability in resource allocation with the flexibility to also offer through short-term and long-term contracts. Sales procurements on the other hand, give us great alternatively in control delivery timelines and managing clusters. It also reduced the profit sharing with [indiscernible] thereby awaiting our pressure on profit margin.

Yes. As you mentioned that the robot companies in China is a growth environment fastly. So as you -- this year, we have covered most of the robot companies in China, and we can see the revenue is increasing very rapidly. So in the next year, we believe the increase of the robotic companies will also be passed. Meanwhile, as more and more Internet companies in China using token services which is the API services. We are seeing the increasement of the business very quickly. So we believe in the next few years, this will be a very important factors to drive the revenue to increase.

[Interpreted] so this is the [indiscernible]. He added that understand your question -- your second question is really about the choice between the leasing model and the CapEx model. So we talked about that before. So generally, there's a general rule of thumb. So when we're looking at the larger customers, especially the customers that have solid profile, test all fundamentals and the trustworthy, premium customers, for example, like Xiaomi, we will tend to choose the CapEx model. While in other growth stage companies, medium and small-sized to small- and medium-sized companies, we generally tend to adopt the leasing model, which is also a way -- a meaningful way to reduce our own risk.

So as we rightly mentioned, there's no kind of a top-down target split between these 2 different methods. And we also talked about in the last quarter as well that the impact of these 2 different methods have different impacts to gross margins. However, we have seen the financial results for the past 3 quarters, which we have adopted various combinations of these 2 different models. And especially when you compare the gross margin for the third quarter versus the second quarter, it actually also improved sequentially. So I would say that at the current stage, we do not expect material changes to the current status. But generally speaking, in the future, we do expect the margin to improve.

Our next question comes from the line of Timothy Zhao from Goldman Sachs.

[Interpreted] My question is regarding the differences between air training versus inferences. Could management share what is the pricing methodology between these 2 kinds of demand? And what has been the pricing trend over the past few months or year-to-date. And in terms of the overall utilization rate of the chips of GPUs, pricing and profitability, can you share more color on the gap between training and inference.

Look, let me answer the questions. When I'm talking about the price strategy for inference and training there's not too much difference between 2 things. So the price is based on the qualities, how many [ solvers ] use, which is the most important factors. And also comparing the margin rate, there are 2 kind of inference services. One is customers buy resource and use our platform foe inference. So that margin ratio is very similar to the training margin ratios. But compared -- another one is customers directly by our API talking services that we think that will have a better margin ratio. But this business just in the beginning. So we have -- we need time to see what is the big difference between the 2 things.

Due to time constraints, this concludes our question-and-answer session. So I'll hand the call back to Nicole for closing remarks.

Thank you. Thank you all once again for joining us today. If you have any questions, feel free to contact our team. We look forward to speaking with you again next quarter. Have a nice day. Bye-bye.