Q3 2025 Earnings Call Transcript

Good day, ladies and gentlemen. Thank you for standing by. Welcome to Alibaba Group's December Quarter 2024 Results Conference call. At this time, all participants are on listen-only mode. After management's prepared remarks, there will be a Q&A session. I would now like to turn the call over to Lydia Liu, Head of Investor Relations of Alibaba Group. Please go ahead.

Good day, everyone. Welcome to Alibaba Group's December quarter 2024 earnings conference call. With us are Joe Tsai, Chairman; Eddie Wu, Chief Executive Officer; Toby Xu, Chief Financial Officer; Jiang Fan, Chief Executive Officer of Alibaba E-Commerce Business Group. This call is also being webcasted from the IR section of our corporate website.

A replay of the call will be available on our website later today. Now let me quickly cover the safe harbor. As usual, we would like to remind everyone that today's discussions may contain forward-looking statements, particularly statements about our business prospects and expected financial results that are subject to risks and uncertainties, which could cause actual results to differ materially from those contained in the forward-looking statements. Please refer to the safe harbor statements that appear in our press release and investor presentation provided today.

Please note that certain financial measures that we use on this call are expressed on a non-GAAP basis. Our GAAP results and reconciliations of GAAP to non-GAAP measures can be found in our earnings press release. Unless otherwise stated, growth rate of all stated metrics mentioned during this call refers to year-over-year growth versus the same quarter last year. With that, I will now turn the call over to Eddie.

Hello, everyone. Welcome to this quarter's earnings call. Over the past year, we have pursued our user-first AI-driven strategy and focused on our two core businesses of e-commerce and AI plus cloud after a year of transformation, our core businesses have demonstrated accelerating growth momentum. We have also largely completed the divestment of our offline assets.

Across Alibaba Group, our businesses now possess strong business fundamentals and profit-generating capabilities. This quarter, we continue to pursue our integrated AI plus cloud strategy in our cloud business, leveraging our industry-leading AI product portfolio. Overall revenue, excluding Alibaba consolidated subsidiaries, grew 11% year-over-year this quarter with AI-related product revenue maintaining triple-digit year-over-year growth for the sixth consecutive quarter. With the rapid adoption of AI technology across industries, customer demand for Alibaba cloud products has surged.

Looking ahead, revenue growth of Cloud Intelligence Group will continue to accelerate. We recently launched Qwen 2.5 Max, our flagship AI foundation model, which has achieved industry-leading performance across multiple recognized benchmarks. As of the end of January, over 90,000 Qwen-based derivative models have been developed globally, making Qwen the most popular among developers across the major model families. Over 290,000 companies and developers have accessed Qwen APIs through Alibaba Cloud's Bilian platform, and we will soon release a deep reasoning model built on Qwen 2.5 Max.

In e-commerce, Taobao and Tmall continued to invest in new user growth and comprehensive user experience enhancements. We saw strong growth in both new consumers and orders during the quarter. 88 VIP members, our core consumer group, maintained double-digit growth, reaching 49 million by the end of the quarter. Taobao and Tmall also made solid progress in monetization as planned with CMR growing 9% year-over-year. At the same time, (indiscernible) continued to strengthen merchant-friendly measures that improve our platform's business operating environment, fostering higher quality and more sustainable development.

Our international e-commerce business maintained strong growth this quarter, driven by cross-border businesses with continued improvement in operating efficiency. During the quarter, we increased investment in key markets while focusing on operating efficiency, with choices unit economics improving on a sequential basis. We expect AIDC to achieve its first quarter of profitability in the next fiscal year. Our other internet platform businesses continued to improve operating efficiency.

AMAP achieved profitability this quarter. Looking ahead, Alibaba's strategic direction and roadmap are clearer than ever. We'll continue to focus on three business categories: first, domestic and international e-commerce; second, AI plus cloud computing; and third, internet platform businesses. We are confident that our focused strategy will drive sustained, solid growth for Alibaba Group.

Today, AI technology advancements are driving profound industry transformation. In alignment with Alibaba Group's business landscape, we will scale up investments in the following three areas as part of our broader AI strategy over the next three years. First, infrastructure for AI and cloud computing. The AI era presents a clear and massive demand for infrastructure.

We will aggressively invest in AI infrastructure. Our planned investment in cloud and AI infrastructure over the next three years is set to exceed what we have spent over the past decade. Second, AI foundation models and AI native applications. AI foundation models are pivotal to transforming industry productivity.

We will substantially increase R&D investment in AI foundation models to maintain our technological leadership and drive the development of AI native applications. Third, transforming our existing businesses with AI. AI technology presents powerful opportunities to enhance user value across our e-commerce and other internet platform businesses. We will increase investment in AI application R&D and computing power and deeply integrate AI across our businesses capturing new growth opportunities in the AI era.

Looking ahead, we're confident in our focused strategy on e-commerce and AI plus cloud and excited by the business opportunities being unlocked by this new technology cycle. Thank you.

Thank you, Eddie. The strong financial results of the past quarter show that we are making very good progress to ignite growth in our core businesses. On our Taobao and Tmall businesses, we saw a significant upswing in CMR growth, which accelerated to 9% year over year, driven by growth in online GMV and improved monetization. This outcome reflected full quarter impact of the software service fee and the increasing adoption of Quanzhantui.

Our cloud business continues to exhibit robust momentum with revenue growth accelerating to 13% and the overall revenue from businesses excluding Alibaba consolidated subsidiaries grew over 11%, fueled by even faster public cloud revenue growth. Additionally, our AI momentum remains robust with AI-related product revenue sustaining triple-digit growth for the sixth consecutive quarter. These achievements highlight our commitment to innovation and reinforce our leadership in the cloud and AI sectors. This quarter, AIDC maintained its rapid growth momentum, primarily driven by strong cross-border business performance.

AIDC increased investments during overseas shopping festivals quarter over quarter and continued to invest in selected European markets and the Gulf region to acquire users. While we made good progress in growing our core businesses, we maintained financial discipline with enhanced operating efficiency, achieving positive EBITDA growth in Taobao and Tmall Group. Meanwhile, we improved operating efficiency of other businesses with the goal of sustainable business growth and achieving profitability. This quarter, AMAP achieved profitability for the first time, while the majority of loss-making businesses will achieve breakeven and gradually begin to contribute profitability at scale within the next one to two years.

We have been actively managing our balance sheet through strategic divestment of non-core assets, share buybacks, and effectively extending our debt maturities at attractive rates. During the quarter, we entered into agreements to dispose all of our interest in Senna for up to a maximum of $1.6 billion and in time for $1 billion. These moves reflect our strategic shift to streamline operations and focus on our core businesses. Recently, we have obtained the PRC antitrust approval with respect to Senna and in-time merge control filing.

We expect the major financial impacts will be reflected in March quarter. These transactions will improve our operating efficiency and enhance our agility. For December quarter, we repurchased shares of a total of $1.3 billion or 0.6% net reduction in share count. Combined with approximately $10 billion repurchased in the first half of this fiscal year, we had achieved a 5% net reduction in share count over the last nine months.

In November 2024, we also completed a dual currency bond issuance, raising approximately $5 billion through a combination of $2.65 billion U.S. dollar denominated notes and CNH RMB 17 billion denominated notes, which was strategically structured to leverage the attractive pricing of RMB notes to significantly lower our overall financing cost. On the consolidated basis, total consolidation revenue was RMB 280.2 billion, an increase of 8%. Consolidated adjusted EBITDA increased 4% to RMB 54.9 billion, primarily attributable to revenue growth and improved operating efficiency, partly offset by the increase in investments in our e-commerce businesses.

Excluding the effect of long-term cash incentive plan, our adjusted EBITDA growth would have been an increase of 5% on a like-for-like basis compared to the same quarter last year. Our non-GAAP net income was RMB 51.1 billion, an increase of 6%. Our GAAP net income was RMB 46.4 billion, an increase of 333%, primarily due to the increase in income from operations, mark-to-mark changes from our equity investments, and increase in share of results of equity method investees. Operating cash flow this quarter was RMB 70.9 billion, an increase of 10%.

Free cash flow this quarter decreased 31% to RMB 39 billion. The decrease in free cash flow was mainly attributed to the increase in expenditure related to our investments in cloud infrastructure. As of December 31, 2024, we continue to maintain a strong net cash position of RMB 378.5 billion or $51.9 billion. The strong net cash position and healthy operating cash flow bring us the confidence and sufficient resources to increase our investments in cloud and AI infrastructure to capitalize the substantial growth potential presented by the latest AI innovations.

Now let's look at segment results starting with Taobao and Tmall Group. Revenue from Taobao and Tmall Group was RMB 136.1 billion, an increase of 5%. Custom management revenue increased by 9%, primarily driven by the growth in online GMV and improvement of take rate. This outcome reflected the full quarter impact of the software service fee and the increasing adoption of Quanzhantui.

We increased the efforts to grow our user basis and continue to invest in strategic initiatives to enhance user experience. These efforts led to strong growth in new consumers and strong order growth. During this quarter, the number of 88VIP members continued to grow rapidly, reaching 49 million. With solid profitability and increasing up on a cohort basis, we will continue to balance the scale and profitability of this program.

Taobao and Tmall Group adjusted EBITDA increased by 2% to RMB 61.1 billion, primarily due to the increase in revenue from customer management service, partly offset by the increase in investment in user experience. Revenue from AIDC grew 32% to RMB 37.8 billion this quarter, primarily driven by strong performance of cross-border businesses. Revenue from international commerce retail business increased by 36% to RMB 31.6 billion, primarily driven by the increase in revenue contributed by AliExpress and Trendio. Revenue from our international commerce wholesale business increased by 18% to RMB 6.2 billion, primarily due to an increase in revenue generated by cross-border related value-added services.

AIDC's adjusted EBITDA was a loss of RMB 5 billion compared to a loss of RMB 3.1 billion in the same quarter of last year. AIDC increased investment during the overseas shopping festivals quarter-over-quarter, and continued investments in select European markets and the Gulf region to acquire users. However, the UE of the AE Choice business improved our sequential basis. Moving forward, we will continue to enhance operating efficiencies within each business and drive high-quality growth by strategic investments in select markets.

Revenue from Cloud Intelligence Group grew 13% and overall revenue excluding Alibaba consolidated subsidiaries increased by 11%, mainly driven by the double-digit revenue growth of public cloud products, including AI-related products. Cloud's adjusted EBITDA increased by 33%, primarily due to a shift in product mix toward higher margin public cloud products and improving operating efficiency, partly offset by the increasing investments in customer growth and technology. We will continue to invest in anticipation of customer growth and technology innovation, particularly in AI infrastructure to capture growth opportunities in the AI era.

Revenue from Cainiao decreased by 1% and its adjusted EBITDA decreased by 76%. There is an ongoing restructuring with our e-commerce businesses taking on certain logistic platform roles. Cainiao will continue to focus on building its global smart logistics network and making end-to-end logistic capabilities available to our own e-commerce businesses, as well as third parties. Revenue from local service group grew by 12% to RMB 17 billion, driven by the order growth of both AMAP and Ele.me, as well as revenue growth from marketing services.

While its adjusted EBITDA loss narrowed significantly as unit economics improved due to operating efficiency and as scale increased. Revenue from Digital Media and Entertainment Group grew 8% to RMB 5.4 billion, while its adjusted EBITDA loss continued to narrow. Revenue from all other segments increased by 13% to RMB 53.1 billion, mainly due to the increase in revenue from retail businesses, including fresh apple in Alibaba Health, while adjusted EBITDA was a loss of RMB 3.2 billion. In closing, during this quarter, we are making significant strides in enhancing the competitiveness of our e-commerce and cloud businesses.

Additionally, we are focusing on improving the efficiency of our loss-making segments to establish a clear path to profitability. In addition, this quarter, we continue to optimize our balance sheet and shareholder return with significant non-core asset sales, share buybacks and effectively extending our debt maturities at attractive rates. As Eddie mentioned, we will continue to execute our strategy and make significant investments to seize opportunities presented by the AI era. Looking ahead, our unique business positioning coupled with our strong financial position give us full confidence to grow.

Thank you. That's the end of our prepared remarks. We can open up for Q&A.

Thanks, Toby. Hi, everyone. For today's call, you are welcome to ask questions in Chinese or English. A third-party translator will provide consecutive interpretation for the Q&A session.

Please note that the translation is for convenience purposes only. In the case of any discrepancy, our management statement in the original language will prevail. If you are unable to hear the Chinese translation, bilingual transcripts of this call will be available on our website within one week after the end of the meeting. [Foreign Language] Operator, please connect speaker and conference lines now. Please start Q&A session.

Thank you.

Thank you [Operator Instructions] To give more people, the opportunity to ask questions, please keep yourself to not more than one question at a time. Your first question comes from Alicia Yap with Citigroup. Please go ahead.

Thank you. Good evening, management. Thanks for taking my question. Congratulations on the solid results.

My question is related to cloud AI and also the CapEx. So with Baba's leading cloud infrastructure service and strong foundation AI models, Baba is well-positioned to capture the transformation era of these AI adoptions in the coming months in China and also possible outside of China. So maybe can management share some insight as how it will translate into the financial upside in terms of the cloud revenue growth and also the cloud margins trend in the next few quarters? We know that CapEx spend this quarter almost doubled from last quarter to $31 billion and also management noted on the prepared remark that your expected investment in CapEx in the next three years will be more than the past 10 years.

So can management clarify the comment and help us crystallize the amount of the spend in the next few quarters and also do you have a budget that you can share for us for the next three years and how will the CapEx span actually impact overall profitability trend? Thank you.

[Foreign Language]

Thank you. And there were quite a few questions in there and perhaps before I get into each of those questions, I could begin by sharing with you our overall views on AI and why we're investing in AI so aggressively and the larger opportunities that we see in the sector, as well as the ways that we have been and will continue to monetize developments around AI progressively as they come out. First of all, we think that Alibaba is extremely well positioned in the AI space, in particular in the Asia market where we have several very important advantages. We are the number one cloud provider in the region and a number four cloud provider globally.

We have leading models, leading technology, we have proprietary AI models, as well as a thriving open ecosystem and we also have a multitude of application scenarios for AI across our B2C ecosystem. But to explain the Group's strategy around AI requires taking an even longer-term view because this is the kind of opportunity for industry transformation that really only comes about once every several decades. So when it comes to Alibaba's AI strategy, our first and foremost goal is to pursue AGI. The pursuit of AGI is our primary objective.

We aim to continue to develop models that extend the boundaries of intelligence. Why is that the primary aim? Well, it's because all of the visible AI application scenarios today that we see around content creation, search and so on and so forth have arisen precisely as a result of the ongoing extension of those boundaries and we want to keep pushing out those boundaries to create more and more opportunities.

[Foreign Language] So when it comes to the business value that AGI can create, why is the ability to use intelligence so important? The so-called standard definition of AGI is that it can achieve the ability of more than 80% of human beings. So from this perspective, 50% of the global GDP is actually the labor pay of employees, including brain and physical labor. From this perspective, we predict that if we continue to develop AI, it will probably be the largest industry in the world.

It is likely to affect or replace about 50% of the current GDP. Secondly, from our perspective, the deep integration of cloud and AI will become the most important AI infrastructure, and the integration of cloud and AI will become one of the largest cloud computing networks to output AI intelligence in the world. This is our goal. Why do we think that most of the so-called intelligence is the token output behind the current model? 90% of the tokens we predict will be generated and output on the cloud computing network in the future because such a huge model only has the highest efficiency on the cloud computing network. [Foreign Language]

So, we see the continued pushing out of the boundaries of intelligence and the pursuit of AGI as the key objective in our efforts. Secondly, the pursuit of AGI can contribute immense business value. There have been studies that indicate that when AGI is achieved, probably the standard for AGI is artificial intelligence that can replace or achieve 80% of human capabilities. Around 50% of global GDP is manpower salaries, including both intellectual or mental work and physical labor.

So, if AGI can be achieved, then that could have a tremendous impact in terms of the restructuring industry around the world. It could have a significant influence on or even replace 50% of global GDP. Second, we will continue to deepen the integration of cloud and AI. We see this as the most important kind of infrastructure.

In the future, we will continue to build AI on an integrated fashion across our cloud and in our own business applications. When we talk about intelligence, we're really talking about the tokens that are output by models. We think that in the future, 95% of those output tokens will be generated on the cloud and distributed by the cloud because only a cloud computing network can generate and distribute tokens with that highest level of efficiency. We can connect to developers on the most efficient and rapid basis worldwide through our globally deployed network of data centers.

Thirdly, we maintain an open attitude with respect to the deep integration of AI into our own scenarios to create value across all of our businesses. We expect that with the further integration of AI across our 2B and B2C offerings, we will achieve higher efficiency. We will increase user time spent and create more user value for our users. So that's precisely the thinking behind our determination, as you referenced to invest more in cloud and AI over the next three years than we did in aggregate over the last 10 years.

[Foreign Language]

The second part of your question had to do with CapEx on a quarter-by-quarter basis going forward. What I've laid out for you is our overall expectation for the coming three years. I would say that on a year-by-year basis, the annual level of CapEx will be more or less equal across these three years, but there could be a fluctuation within each year on a quarter-by-quarter basis, given the time that it takes for supply chains to provide what's needed as well as for the IDCs to get set up. But overall, we would expect it to be relatively even over the next three years.

[Foreign Language]

The other part of your question had to do with the potential impact of our CapEx plans on profitability. I would say that this next three-year period will likely be the single period in which we'll be making the most concentrated and highest level of investments in building out our cloud and AI-related infrastructure and, of course, the hardware infrastructure will have an impact in terms of depreciation, but behind that is our expectation of huge demand for take-up on the part of both internal and external customers. There's huge demand there, and we definitely see this being taken up very rapidly.

Next question, please.

Thank you. Your next question comes from Alex Yao with JPMorgan. Please go ahead.

Thank you, management for taking my question, and congrats on a very strong quarter. My question is also related to AI in the cloud. So, based on my observation, the introduction of the latest DeepSeek large-language model families has brought high-quality models at affordable cost to the entire industry and because DeepSeek's model itself is a free-load charge, monetization of large-language model consumption has moved down one layer in the value chain to the compute power part. So, first of all, do you agree or disagree with this statement?

And secondly, to what extent is the compute power for large-language model or GenAI is not a commodity? And lastly, with the help from DeepSeek's high-quality and high-cost efficiency models, where do you see the most high-potential area in terms of AI-native application within Alibaba ecosystem and outside of Alibaba ecosystem? Thank you.

[Foreign Language] Thank you for some really questions. I think we're still in very early days when we're talking about the advancement of artificial intelligence. Technology, although it is developing rapidly. We are still in the initial stages of development.

So I think the future business models and the future ways in which these models will be monetized are not necessarily clear to anybody today. As the boundaries of intelligence get pushed out, as the models get smarter and smarter, there'll be more and more opportunities to monetize them, but in ways that may not be apparent to us today. I'm talking about a future where the models become sufficiently intelligent to be able to supplant engineers and scientists that kind of expertise, and that will really be a different stage of development. The second point is that if you look at current models today, the level of differentiation across models from different vendors is narrowing.

The differences are becoming less obvious. In fact, it's also becoming less apparent what the differentiation is between open source models and closed models, but the development of all of these different models, be they open source or closed, are beneficial for our cloud computing network offerings because all of these models, open or closed, will need to be hosted on a cloud computing network. So, if you ask me what the monetization pathway is today, it's definitely our cloud computing offerings that exist to host and to support the operation of these clouds. If I could offer an analogy, if AI, artificial intelligence is like electricity in this new era, then our cloud computing network is like the power grid that carries that electricity.

[Foreign Language] Your other question had to do with the AI applications that we think have the greatest potential within the Alibaba ecosystem. Models are becoming more and more powerful and are evolving faster and faster, which means that it is actually quite difficult for me to tell you definitively what are the areas with the highest potential. I think that all applications potentially could have huge potential but we are certainly internally looking at some very interesting opportunities. First, in terms of the Taobao app as a portal for lifestyle and for consumption.

We are doing a lot of internal development and many of these projects will be launched soon. You will see them when they are deployed. But these AI enhancements within Taobao will serve to increase consumer engagement and also drive higher transaction efficiency. Other kinds of AI-empowered applications will contribute to purchase-related decision-making and increase user time spent and create value for users.

So, beyond shopping, we can expect the further introduction of AI features on Taobao to create more kinds of value. We also have our B2C AI offerings. As you know, Quark and Tongyi Qianwen. Quark is an AI search product which has the largest number of users in China and AI is being deployed and will continue to be deployed there to improve search, productivity, creation and overall efficiency.

As you know, the most important asset we have on the 2B side is the DingTalk app as well. We are also deploying AI there to redefine the enterprise collaboration experience and the same is also true for AMAP where we are deploying AI to extend it. It is now currently mostly used as a navigation app. It has 170 million DAUs in China.

But as we more deeply integrate AI, it can become a portal into lifestyle and local services and that will increase user time spent as well.

Next question, please.

Thank you. Your next question comes from Kenneth Fong with UBS. Please go ahead.

Hi, good evening management. Thanks for taking my questions and congrats on the strong results. I have a question on the e-commerce side. So we see that revenue has been very strong and accelerating for both AIDC, as well as CML for TTG.

So can management share with us what have changed and the key initiative ahead for TTG as well as AIDC? And financially, we see the margin for TTG has also stabilized. So how should we think about the trend ahead? And for AIDC, we target for profitability next year.

So shorter term, what are the key drivers for driving profitability? And longer term, should we expect the international business being more profitable than domestic one, given it is less competitive? Thank you.

[Foreign Language] Thank you very much, Mr. Fong, and thank you, Mr. Fong, for your lecture.

[Foreign Language]

Thank you. Again, there are quite a few questions there. Why don't I start with domestic e-commerce? So in the domestic e-commerce business, the strategy that we've been pursuing has been to enhance the user experience to make innovation and optimization around the user experience to achieve higher levels of stickiness.

And looking ahead to the coming year, we see potential to achieve further user growth by investing in users. At the same time, as you've said, we've done a lot to drive increased monetization recently, including, among other things, charging payment processing fees. But we will continue to invest in enhancing the user experience, as well as enhancing the operating environment for merchants, and those things will continue to require investment. So overall, we're looking at stabilizing market share while enhancing the user experience and optimizing merchant operations efficiency. [Foreign Language] Secondly, in terms of our international e commerce business, this is really an amalgamation of lots of different business models, including B2B cross border, as well as local platforms that we operate.

But overall, as we said, we expect to see a stable trend in our international business in the next few years working towards achieving profitability, significant profitability at scale. In our B2C business, as you know, we've done a lot to optimize the business model and unit economics have increased very significantly. And as we said in over the next few quarters, you can expect to see significant increase in profitability. In some countries, we are working on collaborating with local platforms where it makes sense to do so and that is also beneficial to increasing our profitability.

You asked in the long term if the international e-commerce business could be more profitable than our domestic e-commerce business. I don't think that's something I can judge at this point. What I can say is that at this particular point in time, the pathway to profitability in the international business is clearer.

[Foreign Language] This is Toby. I'll take the question regarding margin on as I think Jiang Fan has stated very clearly the strategy for TTG has been to invest in achieving healthy, stable market share. That's been the priority of investment for the past several quarters and indeed in this quarter, investing in enhancing user experience in acquiring new users as well as in our 88 VIP core user group, and those investments will continue. Also, we've integrated new payment methods as well, which is an important means also of engaging with new users.

And of course, while making these investments, we're also actively exploring ways to increase revenues. We've done so with the software service fee, as you've seen, as well as with a more intelligent marketing product, John Two, our QT which is driving higher levels of monetization. And we'll continue to invest in those areas while also seeking to increase monetization in those same ways while paying a lot of attention to optimizing the business environment for merchants on the platform. So margin is always a balance between revenue and expenditure and we do remain in an investment stage as we have been, and we'll continue to invest in enhancing user experience and acquiring new users.

Next question, please.

Your next question comes from Ron Keung with Goldman Sachs.

Thank you, Eddie, Tobi and Jiang. So further on the AI questions, we've seen the AI revenue at triple-digit growth as you mentioned for 6 quarters now. So how should we quantify the size of that? Are we reaching kind of more substantial kind of double-digit mark here for AI?

The reason is I want to know the implications for cloud margins given that historical cloud business is at a very significant margin gap versus some of the global cloud players. So as the cloud mix shifts from public cloud to training and then now increasingly to influence, how should we think about the growth and then more importantly, the cloud margin outlook for us versus our global peers in the longer term?

[Foreign Language] Thanks for those questions. Indeed, you're absolutely correct. Our AI-related revenues achieved over 100% growth, three-digit growth for the sixth consecutive quarter. And customer demand for AI and related products continues to grow.

In fact, that growth is turning out to be much higher than our original expectation. And what we've seen in particular from the Spring Festival, Chinese New Year onward is an explosion in demand for inference. In fact, around 60% to 70% of the new demand that we're seeing now is for inference. So we expect that with this rapid expansion in demand, we will grow our customer base and expand industry coverage across a wider range of sectors and all of that will certainly contribute to higher levels of margin in our AI services.

However, as I said earlier, we're committed to making the highest ever historical investments in CapEx in the coming three-year period. And if you take those investments and amortize them over the coming years that will certainly have an impact on margin. And a final point I would expect that given the fierce competition in China and the different market dynamics here, the kind of margin that you'll see in China will be somewhat different from what you'll see internationally. [Foreign Language] The other thing I would say is that cloud is a business that's characterized by both strong scale effects and strong network effects and the scale effects are particularly important at this stage as we're engaging in large capital investments, building out our hardware with CapEx.

So the point is that as we achieve greater scale and acquire more customers, we'll be able to better optimize the cost of this build-out.

Next question, please.

Thank you. Your next question comes from Yuan Liao with Citic Securities. Please go ahead.

[Foreign Language] Thank you, management. Congratulations on the strong results and all the progress you're making around AI. My question has to do with shareholder returns. I think we see the still USD 2.07 billion.

Pardon me, USD 20.7 billion of dry powder for share buybacks. I'm wondering if you can tell us how that will be deployed given all of the progress that you intend to make around AI and the investment there.

[Foreign Language] Well, in the December quarter, as you know, we conducted USD 1.3 billion of buybacks, achieving a 0.6% net reduction in share count. And during the first half of the fiscal year, we repurchased approximately USD 10 billion achieving a 5% net reduction in share count over the last nine months. So as you can see, we've already made very considerable progress with these share repurchases. [Foreign Language] As is common market practice for companies engaging in share repurchase programs, of course, we will consider the current share price as we execute our share buyback program.

And that's why in the first six months, you saw an accelerated pace of share repurchase. As an example, in the June quarter, we raised USD 5 billion in debt financing and used the entirety of the proceeds for share buybacks, and that was because at the time our share price was only USD 80, which in our view, is extremely undervalued. [Foreign Language] As you know, we have set up at the Board level a capital management committee precisely for the purpose of optimizing capital allocation so as to enhance shareholder return. The committee has been providing strong management to shareholder return initiatives, and we'll continue to do so, and we will aim to continue to elevate shareholder returns through a combination of dividends of share buybacks as well, of course, as investment in high-growth, high-potential areas. [Foreign Language] Going forward, we will continue to deploy our cash effectively and optimally to enhance shareholder return and we'll continue to execute share repurchases in accordance with the allocation and guidance given to us by the Board of Directors and with an eye on the share price.

Next question, please.

Thank you. Your next question comes from Gary Yu with Morgan Stanley. Please go ahead.

Congratulations on a strong set of results. I have two questions, both related to AI and cloud. The first 1 is I appreciate management comment about AI potentially representing 50% of the global GDP and cloud is important infrastructure. But when we think about how we monetize this potential sizable opportunity, besides the infrastructure, which is represented by cloud, how should we think about Alibaba's strategy to tap into the application or software layer?

And could management comment more about the enterprise adoption side because we kind of understand the previous comment about consumer-facing part, including consumption and local services? And my second question is related to CapEx. So our management comment about the next 2 years being the half year investment cycle in the past decade. How should we think about how Alibaba plans to allocate investment?

How much will be spent on chips and specifically, how should we think about the mix between spending on import chips from U.S. as well as domestic chips. And in the event of further export restriction from the U.S., how should we think about any contingent plan in order to continue with the investment?

[Foreign Language] Well, thank you for those questions. As we said, we definitely see some very clear opportunities for the application of AI on the B2C side, which I've already mentioned. When it comes to the 2B side, there are a host of different opportunities for Alibaba Cloud Intelligence to capture. I think as a general remark, SaaS software going forward will become more and more AI agent driven.

In other words, a lot of the internal systems employed by enterprises in the future, will become more like a network of multiple AI agents to collaborate and call on 1 another to provide services, including to assist the enterprise even in important decision-making. Along with that, there will also be a lot of opportunities to upgrade not just the software, the SaaS layer, but also the underlying supporting pass layer. Another really good example, I think, is Ding Talk, which is our flagship enterprise collaboration product. I think going forward with AI, a lot of functionality in a will be achieved through natural language interaction and today's CRM and ERP-type systems will become more like databases that feed in.

So enterprise meetings and decision-making can all be handled through natural language on Ding Talk. And again, there are many opportunities for the integration of all kinds of agents into Ding Talk. Another important thing to note is that a lot of companies internally have huge amounts of proprietary data and their own proprietary in-house processes and they'll need to leverage AI agents in order to enhance the efficiency of those kinds of processes. So I think overall, AI will be able to empower and reshape all kinds of corporate -- software corporate applications, and this is a huge opportunity. [Foreign Language] On to the other part of your question regarding CapEx in order to give you a breakdown as to what percentage of CapEx will go into what area that would really require giving you a detailed breakdown of the cost of all the different equipment in our IDCs.

So I don't think I'm really at liberty to get into that kind of detail with you. What I can tell you is that the way we've designed our cloud very deliberately is to be compatible with a whole range of all kinds of different ships. So no matter what kind of policy changes may be forthcoming in the future, that will not affect us. I mean we'll certainly be able to implement our investment plan.

Due to time limit, we will now take the last question.

Thank you. Your next question comes from Jialong Shi with Nomura. Please go ahead.

[Foreign Language] Thanks, management for taking my questions. I have two questions. My first is following your successful divestment of physical retail assets in summer time and Sun Art. I'm wondering if you have any plans for potential future divestment of other assets, including Freshippo (ph) and Illuma.

My second question, coming back to cloud, it seems that a lot of the revenue growth that's being driven in cloud is as a result of demand for compute associated with uptake in AI. But I'm wondering if there's -- in addition to that, any future monetization pathway monetization model for Tony, the use of the model itself, is there any way to generate revenue and profits from the model not just from the compute. And as Eddie said, the Chinese AI market seems to be rather homogeneous in the sense that there's a lack of differentiation across the different models. Do you think that will continue to be the case in the future, will there come a day when there might just be one or two big models in the market that take up a large market share.

[Foreign Language] Thank you. This is Toby. I'll take the first part of the question, and then I'll hand over to Eddie. So as you know, we have been strictly implementing our strategy of focusing on our core businesses and exiting our non-core businesses.

And as has been announced in the December quarter, we completed the sale in full of Sun Art and in time, and we have already passed monopoly reviews. So we expect to complete those transactions either in the March quarter or in the June quarter. We will continue to look for opportunities to exit our nonstrategic minority-owned assets, and you'll continue to see announcements coming forth as we go forward. And in the future, we will be focusing as we have been on our core businesses, we'll be looking at the assets in which we're invested and exiting, as I say, the noncore ones.

But in respect of other assets, we'll be looking for ways to tap into their business value and to see that better -- that value better reflected in the valuation of Alibaba Group. And I think Freshippo, which you mentioned is a really good example of that. Freshippo is a business that's achieved very good growth, very good expansion. It's achieved very good profitability, and it's an excellent example of an innovative business model that integrates online and offline retail.

So given its success, no, we don't have any plan to sell Freshippo, but we would certainly have and take an open attitude for example, to introducing a strategic investor or other solar approaches that could enhance the value of Freshippo and in particular, result in a value being better reflected in the valuation of the Alibaba Group. So this is something that needs further exploration going forward.

[Foreign Language] This is Eddie. Let me take the other part of that question. So, yes, the growth the robust growth that we're seeing in cloud revenue is largely being driven by a huge customer demand for AI products. Our (indiscernible) model is an open source product, but that does not necessarily mean that it is provided free of charge.

In fact, we do charge for access to the model via the API on by Lean platform. Of course, we're not -- the revenues that we create that way are relatively low. But as the models improve, become more powerful as their capabilities become more sophisticated, it's certainly possible that we could charge more for the use of the models. But I should also say that Q1 is an open source model, which means that a lot of developers are using it as a foundation model on us to develop their own vertical models and their own applications and it's, therefore, only natural that they will deploy those models, those applications on Alibaba Cloud because that's the most efficient way to do it.

Thirdly, I would also say that when customers are accessing our Q1 model through the API. There are a lot of opportunities for us to cross-sell to them other of our cloud offerings. It's really excellent example of a cross-selling opportunity. So, although, Q1 in and of itself is not driving and may not drive huge revenues.

It's a very, very important part of our overall cloud offering that will help to drive overall cloud revenue. [Foreign Language] Finally, on your last question, which had to do with the future market landscape for AI in China. This is a period of time right now where things are iterating and developing rapidly. And at the same time, it's very early days in the development of AI. So I just don't think it's possible to judge what the end game will look like.

What I can tell you is the trends that we're seeing today, as I said, the gap between different foundation models today is narrowing. I think that going forward, you can expect to see a lot of demand around post training. So thinking of models like OpenAI, DeepSeek, the reasoning models as well as the reasoning model of Qwen that we will be releasing shortly, a lot of the demand will be around post training to customize those models, adapt them to different sectors, different use cases to adapt them to a private data sets. And I think that there's huge market potential for that kind of post-training.

I think there's also lots of opportunity to create value for a whole range of different companies, large and small. And all of these models, the specialized models, the vertical models I've referenced will all need to be hosted on cloud. So that's why we are very excited and confident in the prosperous development of an open cloud ecosystem.

Thanks, Eddie. That wraps up the Q&A session of today's earnings call. Thank you very much for joining us today. We look forward to speaking with you soon. Thank you. .

That does conclude our conference for today. Thank you for participating. You may now disconnect.