Home / Transcripts / Alibaba Group Holding Limited (BABA) · May 13, 2026

Alibaba Group Holding Limited (BABA) Earnings Call Transcript

May 13, 2026

New York Stock Exchange US Consumer Discretionary Broadline Retail earnings 77 min

Earnings Call Speaker Segments

Lydia Lu executive
#1

Hello, and welcome, everybody. I appreciate you being here for the Alibaba Group call covering the March Quarter and the full Fiscal Year 2026 results. Joining me today are Chairman Joe Tsai; Eddie Wu, our Chief Executive Officer; Toby Xu, our Chief Financial Officer; and Jiang Fan, who leads the Alibaba E-commerce Business Group as its Chief Executive Officer. Just so you know, there's a live webcast of this session. Later on today, a replay will be posted on our website. During this call, we might deliver forward-looking statements and go over some non-GAAP financial measures. Those forward-looking statements represent what management currently anticipates, and they carry risks along with uncertainties. You can find our GAAP results, plus the reconciliations between GAAP and non-GAAP measures, in the earnings press release and investor presentation issued today. Unless we say otherwise, whatever we discuss will be framed as year-over-year comparisons. With all of that said, I'll hand things over to Eddie.

Yongming Wu executive
#2

< > Greetings, and thank you for joining the earnings call for Alibaba Group covering the fourth quarter of fiscal year 2026. During the quarter just ended, we poured substantial resources into two strategic priorities — AI plus Cloud, and consumption — and those efforts are quickly showing up as concrete business outcomes. Group revenue rose 11% compared with a year earlier. At Cloud Intelligence Group, external revenue growth sped up to 40% in this quarter, while revenue from AI-related products delivered triple-digit expansion for a 11th quarter in a row. On a like-for-like basis, China e-commerce CMR increased 8% versus the prior-year period, and in quick commerce we drove meaningful gains in unit economics while holding onto market share. At this moment, we find ourselves at a pivotal juncture. The sector is moving beyond chat-style bots and toward self-directed AI agents, and that transition is driving tremendous growth in three primary workload areas: training, inference, and agent orchestration. Against that background, Alibaba's AI has moved past its initial investment phase and is now expanding its commercial rollout. I'd like to walk you through four topics in depth: AI commercialization, cloud infrastructure, the AI application ecosystem, and our consumption business. First off, the turning point for AI and cloud commercialization has arrived. For this quarter, annualized revenue from AI-related offerings at Cloud Intelligence Group surpassed RMB 35.8 billion, and it continues to expand at triple-digit rates. That figure is now equal to 30% of Cloud Intelligence Group's external revenue. About a year from now, we expect AI-related product revenue to cross the 50% threshold and turn into the primary driver of revenue growth for the Cloud business. As a result, in the quarters to come, external revenue growth at Cloud Intelligence Group ought to continue speeding up beyond the 40% level we're seeing at present. Since long-term AI demand is a sure thing and our full stack technology advantages are obvious, we're confident this path will produce robust growth over the medium to long term. What this demonstrates is that AI is driving a sweeping upgrade of the whole Alibaba Cloud business, as its growth engine shifts entirely away from traditional compute and storage toward models, AI compute and agent services. We're also seeing AI model and application services revenue grow exponentially — that's a fresh revenue engine powered by both foundation model services and AI-native software working together. Over the last 3 months, token consumption volumes on our model services platform climbed considerably quarter-over-quarter, since enterprise customers sped up their move from simple tasks toward production scale and complex workloads, which kept pushing up demand for model and application services on the Model Studio platform. We anticipate that model and application services annualized recurring revenue, or ARR, including the Model Studio platform, will exceed RMB 10 billion in the June quarter and reach RMB 30 billion by year-end. This revenue stream is showing a higher margin profile more and more clearly, and that makes it a source of healthy growth of high quality. Second, our AI infrastructure supports our entire technology stack and acts as a lasting moat. T-Head's proprietary GPU chips have reached scaled MaaS production, with more than 60% of compute capacity already serving external customers across the Internet, financial services and autonomous driving verticals. In China, we are the only AI cloud player able to provide homegrown AI chips at scale, and that gives us control over the compute portion of our supply chain while allowing us to deliver highly competitive AI inference and training services to customers. When compute is in short supply, this structural advantage benefits both revenue growth and gross margin improvement. At the same time, AI-driven changes are accelerating across our cloud services. Demand for standard cloud products built around CPU storage and containers has climbed steeply because agent workloads have surged, and we're turning those offerings into infrastructure tailored for the agent era. Third, when it comes to the application layer, we've built a complete closed loop spanning everything from AI-native software through a full agent ecosystem. New products continue to launch from Alibaba Token Hub ATH, connecting consumer and enterprise environments, and we've achieved breakthrough progress in AI-native software as well as coding agents. Reasoning, coding and agentic capabilities keep getting refined with each iteration of the Q1 model. On the enterprise side, we rolled out a suite of offerings that span smart workplace solutions, AI-assisted coding, and tools for managing business operations — all aimed at helping companies tap into significantly higher productivity. Turning to consumers: on May 7, the Qwen app fully unified the commerce service capabilities of Taobao and Tmall. As a result, Qwen now runs as a deeply integrated thread through an ecosystem that reaches across Taobao, Alipay, Amap, and Fliggy — making it the first all-in-one personal assistant in China to seamlessly connect daily life, productivity, and learning. Fourth, when it comes to our consumption business and the group overall, long-term value sits at the top of our priorities. Beyond AI, our consumption strategy continues to move forward at a steady pace, with CMR growth showing a notable rebound. For this quarter, on a like-for-like basis, CMR was up 8% compared with a year earlier, driven by our continued work to improve how users experience the platform and how efficiently merchants operate. In quick commerce, we achieved substantial gains in unit economics while keeping our market scale stable. Taking everything together, the returns on the money we've put into AI + Cloud and into consumption are becoming increasingly evident. Revenue from AI + Cloud is accelerating, and margins are improving as well; ARR from model and application services continues to grow at a consistent pace; and across our consumption business, operating efficiency keeps improving. Because AI creates an opportunity unlike anything we've seen before, Alibaba is now at a critical turning point, one where the money we've spent on technology is beginning to deliver commercial results. We will hold to our strategic results and use our full stack AI capabilities to underpin long-term growth; that wraps up my prepared remarks. Now I'll pass it to Toby, who will walk you through our financial results. Thank you.

Toby Xu executive
#3

Eddie, thanks. Our strategic focus continues to be squarely on AI plus Cloud along with our consumption operations. A number of growth drivers — technological progress and business innovation among them — are coming together and producing powerful tailwinds. Within AI + Cloud, our full-stack strengths reach across models, cloud infrastructure, and applications; having established leadership at each layer, combined with robust expansion in our AI + Cloud operations and a well-defined route toward monetizing our MaaS platform, we feel confident putting substantial capital to work to broaden our lead. Turning to consumption, we posted solid CMR expansion on a like-for-like basis in the quarter, while our quick commerce unit continued to lift UE and AOV from one quarter to the next. Let's now review the financials for the quarter. Consolidated total revenue came to RMB 243.4 billion. If you strip out revenue from Sun Art and Intime, like-for-like revenue would have risen 11%. Total adjusted EBITA fell 84%, largely reflecting our strategic bets on technology businesses, quick commerce, and user experience, partially offset by better operating results driven by ongoing growth in consumer management service within the cloud business plus greater operating efficiencies across multiple businesses. GAAP net income reached RMB 23.5 billion, up 96%, mostly because of the year-over-year rise in net gain from mark-to-market changes on our equity investments and the disposal losses of Sun Art and Intime booked in the same quarter a year earlier, partially offset by the drop in adjusted EBITA. Operating cash flow came in as an inflow of RMB 9.4 billion. Free cash flow was an outflow of RMB 17.3 billion; we're plowing operating cash flow back into strengthening our competitive edge in AI. As of March 31, 2026, we held roughly USD 38 billion in net cash. If you exclude debt maturing beyond 5 years, our net cash position comes to about USD 59 billion. That balance sheet strength gives us the confidence to invest for growth; let's now turn to our consumption businesses. China E-commerce Group revenue was RMB 122 billion, up 6%. Customer management revenue rose 1%. To help merchants expand their operations and boost their willingness to spend on our platform, we upgraded our business development program for select merchants in the quarter, under which platform subsidies for those merchants are tied directly to their marketing spend on our platform. For accounting purposes, those subsidies, formerly booked as sales and marketing expenses, are now booked as a contra revenue item to CMR. As a result, CMR was up 1% year-over-year in the quarter. If you exclude the contra revenue effect of that program, CMR on a like-for-like basis would have been up 8% year-over-year. Revenue from our quick commerce business climbed 57% to RMB 20 billion. The quick commerce business improved UE further and raised AOV from quarter to quarter, mainly thanks to order mix optimization. Adjusted EBITA for our Alibaba China E-commerce Group came in at RMB 24 billion, down 40%, largely reflecting spending on quick commerce, user experience, and technology, though customer management service contributed positively. If we set aside the loss from quick commerce, EBITA for Alibaba China E-commerce Group would have held steady compared with a year earlier, and it will move up and down between quarters because of sizable outlays on merchant retention and user experience. AIDC revenue rose 6% in the quarter. AIDC's adjusted EBITA loss shrank considerably versus a year ago, getting close to breakeven, helped by both logistics optimization and operating efficiency. Unit economics for AliExpress' Choice business keep getting much better on a sequential basis. Now let's turn to the business update and results for Cloud Intelligence Group. Our cloud business posted yet another quarter of accelerating growth. Revenue from external customers sped up to 40% growth. AI-related products kept driving this momentum. We recorded our 11th quarter in a row of triple-digit growth in AI revenue. Its portion of external cloud revenue keeps rising and now stands at 30%; AI revenue for the quarter is RMB 9 billion, with an annual revenue run rate of RMB 36 billion, or USD 5.3 billion. That clearly shows the scale and acceleration of our AI business. The adjusted EBITA margin stayed fairly steady at 9.1%. All others segment revenue fell 21% to RMB 65.5 billion, mostly because of the disposal of the Sun Art and Intime businesses and lower revenue from Cainiao, partly offset by higher revenue from Freshippo and Amap. All others adjusted EBITA showed a loss of RMB 21.2 billion, mainly from greater investment in technology businesses, including foundation models and the consumer-facing Qwen app. As we wrap up this fiscal year, we stay committed to consistent shareholder returns. Our Board of Directors has approved an annual dividend of USD 1.05 per ADS. We will keep investing decisively in AI and consumption businesses, where we see substantial long-term growth potential and where our competitive advantages are compounding; we believe these investments will produce growth and returns over time, ultimately creating greater value for our shareholders. Thank you. We will now open for Q&A.

Lydia Lu executive
#4

Hello, everybody. Feel free to pose your questions in either Chinese or English. Consecutive interpretation will be handled by a translator from a third party. Should any inconsistency arise, what our management said in the original language takes precedence. <<foreign_lang>> Operator, kindly open the Q&A portion now. Thanks.

Ronald Keung analyst
#5

Thanks for giving us that first look at the quite substantial ARR figures for AI MaaS and applications, along with the target. So my question is this: what portion of that ARR comes from models we built ourselves, such as Qwen, and what portion comes from models made by other companies? And with token prices having gone up lately, how would that affect MaaS, and what would it mean for our Cloud margins?

Executive executive
#6

<<foreign_lang>> < > Appreciate that question. This quarter is the first occasion on which we've disclosed the newest number for revenue from model and application services. That figure is made up mainly of 2 components. First, it covers money brought in through API calls on MaaS via our Bailian platform. Second, it covers money brought in through subscriptions to our AI software. Right now, the bulk of the money is generated by the first of those 2 components, though this platform is an open one. That means we offer access to models we own as well as models from other parties, whether open source or closed. That said, at this moment, the majority of that money is generated by models we own, which includes Qwen, Tmall, and our models for generating voice and video. <<foreign_lang>> Your second question matters a great deal too, since during the last quarter, and across the last few months, the market has shifted enormously, with AI moving away from working as a conversational chatbot and toward delivering agentic capabilities. These agents are getting better and better at tackling highly complex problems, which means they must perform far more inferencing than before. And exactly because such agents can assist in resolving highly complex tasks, customers' willingness to accept higher prices is solid, and we have raised per token prices, while demand stays strong and keeps expanding. In fact, our capacity to serve this demand cannot match all of the growth and demand; we in fact have many customers who are still waiting for access to the service. By its nature, MaaS will carry a higher gross margin than IaaS. That point is worth noting. And I can add several significant points beyond that. The first is that reasoning, or inferencing, technology keeps progressing. Thus, each quarter brings fresh results when it comes to optimization in reasoning and in inferencing, with ongoing incremental gains in the token capacity of one server and one card. Meanwhile, as model capabilities keep getting stronger and model prices keep rising over the next year or 2, we expect this to be a process of ongoing price improvement. So from that angle, I believe the fast expansion of this business across the next few quarters will produce a strongly favorable effect on our gross profit margin overall.

Kenneth Fong analyst
#7

Congratulations on the really impressive AI momentum. Let me ask about the return on invested capital for the AI spending. Our AI investments have fueled a remarkable 40% increase in cloud, but at the same time they've placed a sizable burden on group free cash flow and on our EBITA. So how should investors think about the return on this investment? And what framework does management use to weigh aggressive AI spending against earnings stability?

Toby Xu executive
#8

Thanks for that question, Kenneth. This is Toby, and I'll take that first question first, since I believe it matters and is of interest to everyone. What's being asked is why free cash flow has turned negative and how we're handling it. To begin there, the reason is mainly the very substantial AI investments we've made during the past year. And we've stayed extremely resolute in committing to those investments, exactly because we've identified the historic opportunity that AI represents. So over the past year we've been very resolute in making these investments; and as we look toward the next 2 years, we plan to remain just as resolute in carrying them forward, once again because we view this as a critical window of opportunity that stays open during that period, over the coming couple of years. Beyond that, there's really been no major shift in how we think about cash flow. To start with, Taobao and Tmall are the biggest source of operating cash flow for the group, and that cash flow is highly stable. And as we look forward across the next 2 years with respect to quick commerce, losses will shrink very substantially. Meanwhile, AIDC will move from posting a loss to becoming profitable. So we regard these trends over the next 2 years as strongly favorable for our net cash flow. One more key point to add is that our continued investment in cloud infrastructure will lift the revenues we're able to generate from our AI and cloud offerings. Simultaneously, we'll raise gross margins on those very same offerings. Through those means, we anticipate achieving greater net cash flow from our cloud and AI business, and that cash flow can then be directed toward supporting the buildout of the related infrastructure. A further point is that our balance sheet is very strong. As of March 31, 2026, we had roughly USD 38 billion in net cash. And when you set aside debt maturing beyond 5 years, our net cash position comes to roughly USD 58 billion. That balance sheet strength, then, also gives us the confidence to reinvest for growth. On top of that, I'd point out too that our ability to go after financing through the capital markets is quite strong, and when we need to fund our development, we have what it takes to bring in capital from those markets. That's what I wanted to say up front in reply to what you asked about cash flows, and I'll stop here to check whether Eddie wants to chime in with anything.

Yongming Wu executive
#9

Thanks. Your question was about the money we're putting into AI and what kind of return we'll see from that down the road. So building on what Toby already covered, I want to share a few extra points about the direction we're taking. Manufacturing strikes me as the closest comparison. Basically, so that we can produce more and sell more later on and bring in more revenue, what we're doing right now is putting capital to work to construct 2 factories, in a manner of speaking. We could label the first one the AI training factory. The second one could be called the inferencing factory. And our AI data centers have to supply the power for both of those factories, which calls for investing cash flow in the present. Still, when we look ahead, the route to earning a strong return on investment in those factories, in those areas, is quite obvious. On the 2B side, that comes from monetizing our 2B offerings, which include our cloud-based IaaS and, naturally, MaaS, plus our AI-native apps. And I can share with you that right now, not one card across our service sits idle; so in the coming 3- to 5-year period, we view the ROI from this investment as being very clear.

Thomas Chong analyst
#10

I want to ask about quick commerce. In the prepared remarks, there was discussion of the UE improvement. I'd like a bit more detail on what's driving that, specifically around AOV, the subsidies ratio, the fulfillment ratio, and so on. Also, I recall that last time, we discussed what quick commerce might look like over the coming couple of years. Has anything been updated or changed regarding how we view the landscape or the UE over the next 3 years?

Executive executive
#11

<<foreign_lang>> < > Thanks. To start with, because we invested heavily in quick commerce, we saw extremely fast growth in that business across the past year, which represents a very fundamental shift in where we stand in the market. Against the same quarter a year earlier — which, naturally, came before all of this investment at large scale — our order volume and our market share both rose substantially. Total order volume reached 2.7x the level of the same quarter a year earlier, while non-food orders hit 3x. Starting in April, even as order volume held steady, we kept pushing significant UE gains via better fulfillment logistics efficiency and through optimization of order mix. As a result, we're confident UE will become positive by the close of fiscal year '27. <<foreign_lang>> < > As we optimize UE, we'll keep innovating to make things better for consumers and merchants alike, and in that way sustain our long-term competitiveness in quick commerce; we're confident that at a new scale and market share, our quick commerce business will reach overall profitability down the road. During this quarter, quick commerce kept producing synergies with our conventional e-commerce business, as shown by driving customer acquisition, enhancing user engagement, fulfilling diverse consumer demands, increasing transactions, improving monetization and supporting logistics infrastructure. When it comes to categories, quick commerce kept driving sales across a range of categories, particularly food and fresh produce and healthcare, and it contributed to accelerated growth at Freshippo and Tmall Supermarket. So within our conventional e-commerce business, GMV and CMR showed strong growth momentum in the March quarter, and quick commerce was instrumental in driving that performance.

Jialong Shi analyst
#12

Following up on what you said at the start about MaaS, I'd like to ask a couple of things. First, what stands out as Alibaba's biggest edges relative to China's other large AI platforms and its AI start-ups? Over in America, agents — particularly for coding — represent the quickest-growing area within AI. When do you expect that same sort of expansion to arrive in China for coding? Another thing we're aware of is that buyers in China show less willingness to spend on SaaS than buyers in America do. Does that imply the eventual commercial path for Chinese coding offerings could end up weaker than what we observe among American players?

Executive executive
#13

Thanks. Our definition of the MaaS platform known as Bailian Model Studio frames it as an open platform for AI inference. Right now, most of Bailian's top line certainly comes from the proprietary models we built ourselves. Yet when you set us beside China's AI start-ups, my view is that our investment happens at a far bigger scale and spans a far wider set of model categories; by comparison, such start-ups often concentrate on one quite narrow vertical slice, and that focus lets them advance quickly. Looking purely at our MaaS operation, I'd characterize those AI start-ups as collaborators, not rivals. At Alibaba, though, we put particular weight on model capabilities and on growing them across every space and every vertical so we can serve needs that are extremely broad and varied — that covers our coding model capabilities, image-based models such as Wanxiang and HappyHorse, plus these new world models and, naturally, voice models too; our goal is delivering every kind of model for every kind of need. That, I believe, sets us apart from those start-ups; and those start-ups are, at the same time, collaborators of ours. Alright. Your other point concerned the timing for comparable growth in China to what's being observed in America around coding; from what we can see, drawing on trends we observe ourselves on Bailian and on what some of these Chinese AI start-ups who collaborate closely with us have experienced, my answer is that China has already arrived there. The bulk of the increase in utilization we've observed from, for instance, November or December of last year up through May of this year has come from capability upgrades related to coding. Such models cannot substitute for software engineers. Essentially, they can handle a broad range of highly complex tasks that go well past coding itself, in any digitalized productivity setting; as a result, we've witnessed AI coding capabilities advance markedly in both America And within China, these capabilities can do far more than simply coding, if you will. They're able to handle an entire broad spectrum of highly complex workplace tasks, provided those tasks are capable of being digitalized. So as we look toward the coming 2 to 3 years, we regard this as a tremendously significant growth driver. <<foreign_lang>> < > Regarding your other point, we've likewise observed the reduced willingness within China to pay for SaaS. That said, my view is this is set to shift as the models grow ever more powerful and can genuinely tackle extremely complex tasks, extremely complex problems. Since they deliver genuinely valuable intelligence. I believe we should anticipate seeing demand for that type of service in China matching what we see in the U.S; in a sense, once the value generated by tokens surpasses the cost of those tokens, demand for tokens becomes infinite, in a sense. Thus we view growth in AI demand as a long-term certainty. And I can also give you some figures regarding the growth we're witnessing on our own Bailian platform from November, December last year through to May of this year; it exceeds a 10x growth. On our ARR, it's already above RMB 8 billion. And I believe this quarter, it's highly certain that we can reach ARR of more than RMB 10 billion.

Ellie Jiang analyst
#14

I'd like to remain on the subject of those global comparisons. When you examine things worldwide, it appears that peers overseas have achieved the quickest returns in enterprise agentic workflows, while consumer and monetization efforts have tended to trail somewhat. So looking ahead, given that Alibaba is putting money into several fronts at once — infrastructure models, cloud, and the Qwen app — how should we weigh strategic priority and resource allocation across 2B versus 2C initiatives? Should the enterprise segment keep building momentum, might we think about steadily moving more resources from the Qwen app toward Cloud and MaaS?

Executive executive
#15

<<foreign_lang>> < > Appreciate that question. It's a solid one. Still, I believe that at bottom, when you consider AI development, what's really happening is a paradigm shift in computing, and it's about using this new technology so that users — no matter who they are — can finish tasks and resolve problems; that holds just as much on the 2C side as it does on the 2B side; of course, right now we observe a greater readiness to pay on the 2B side, since demonstrating a business case with attractive ROI is simpler for a business. And right now, that means the bulk of our infrastructure resources get directed toward the 2B side. Yet in the final analysis, AI is ultimately an invention meant to serve as a helper and assistant for humans, aiding people across a broad spectrum of activities — their everyday lives, their studies, and their jobs. What AI actually accomplishes is identical across every one of those situations. It comes down to resolving problems. That holds for 2B and for 2C alike. So yes, willingness to pay for 2C is lower at this point, yet we already observe a business model in which consumers pay for individual usage in the U.S. And I'm confident the same will eventually happen in China, particularly as the technology gets better and can help them solve genuine problems in their everyday lives more effectively. Thus we believe that in the end, the business model will be identical internationally; and my guess is we'll reach that point in China within roughly the coming 1 to 2 years.

Joyce Ju analyst
#16

I've got another question, this one about how the cloud business will grow going forward. What I want to get a sense of is where you see EBITA margins for cloud landing over the coming quarters. As the business speeds up, should we look for margins comparable to those of your international competitors?

Executive executive
#17

Regarding how deeply AI technology will penetrate every different industry, we are honestly still at the very beginning of what will be a lengthy process. Still, the goal we've set is unambiguous. That goal is to secure growth, to propel growth, to propel growth in token consumption, and to capture a bigger share of the market. We intend to keep growing faster than the market average so that we can take a larger share and firmly lock in our position as the absolute market leader. Those, then, are the main goals, while margin remains a secondary consideration. Another point worth making is that across the coming 3 or even 5 years, there will be physical constraints affecting production, affecting production capacity for chips, for memory, for the physical items required to sustain all of this growth in demand. One edge Alibaba Cloud holds is the scale of our customer base, along with the scale effect from all the CapEx we've invested over these years; yet in this climate of market scarcity, we're already observing that deploying one new server this year costs us double what that same server would have cost a year ago. So cost inflation has exceeded 100%. Given that higher replacement cost effect, we do have a certain pricing power when it comes to new customers and also old customers; so I believe over the long term, the asset pricing effect will work in favor of our revenues going forward. Second, we see extremely rapid growth in MaaS, as we reported to you. And by its very nature, as we've said, MaaS carries a much higher level of gross margin than IaaS or traditional types of IT operations. So as demand for inference keeps growing exponentially, we expect this will be very positive for gross margin; and thanks to the optimization of our reasoning technology, the output capacity, the productivity of a single card will keep rising. One more factor is that as we keep scaling up the deployment of T-Head, the T-Head chips represent the highest value for money compute power on the cloud platform, and that too will contribute to a better gross margin; but for several objective reasons that I've outlined, I think, overall in the next 2 to 3 years, we can expect to see a significantly higher gross margin for Alibaba Cloud, and we can expect to start to see that in the next 1 to 2 quarters.

Gary Yu analyst
#18

I've got a question about CapEx. What level of CapEx spending is needed to meet demand from both MaaS and the long-term cloud revenue? Also, management brought up the T-Head opportunity. Right now, what share of T-Head is deployed on AliCloud? And as that share grows, what margin uplift should we expect from our in-house chip?

Executive executive
#19

<<foreign_lang>> < > Thank you. That first question matters quite a bit. In the prepared remarks we gave during last quarter's earnings call, we laid out a revenue forecast covering the next 5 years, and it was a very target. Basically, though, if you look at where we stood in 2022, before AI models grew explosively, and compare it with what we think we'll need in 2033, that's roughly a 10x rise; in other words, we'll require 10x the data center infrastructure we had back in 2022. That said, there are multiple routes to obtaining such compute capacity. A portion of it can come as CapEx. Another portion can come as OpEx; in fact, we're currently picking up a fair amount of computing capacity through OpEx. For reasons we've already talked about, the situation today is complicated. Still, my sense is that with investment on that scale, we'll probably exceed the original CapEx number we gave of RMB 380 billion; yet we can secure some compute via OpEx, and since we have our own proprietary T-Head chips, we're also able to sell AI servers, using those chips, to other computing centers, or we can co-build computing centers alongside others. So there are various routes to reach the point we need to reach. The key point, though, is that compute infrastructure demand will be 10x what it was in 2022. <<foreign_lang>> < > Right. As for our proprietary T-Head chips, they can be deployed across a very large portion of our AI infrastructure, and that includes not only compute chips but a full stack, memory included. Currently, though, the ratio remains fairly low, and the reason is constraints on production capacity within China, which has been limited. Naturally, it has been rising. But as we deploy ever more of our own proprietary T-Head chips, those new chips will definitely add very, very substantially to gross margin expansion. It's fair to say that semiconductors made domestically in China trail the leading overseas ones when it comes to energy efficiency and production efficiency. Even so, if you look at the globally leading AI chip vendors today, their gross margins reach as high as 60% or even 80%. So as we scale up domestic chip production and capabilities get better, I believe there's considerable room for our chips to deliver very strong value for money relative to that 60% to 80% gross margin that other vendors are taking.

Lydia Lu executive
#20

Thanks. That concludes our earnings call for today. We’re grateful for the time you’ve spent with us and for taking part, and we hope to talk with you again before long.

Operator operator
#21

Thank you. You now may disconnect your lines.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Alibaba Group Holding Limited transcript - plus 132,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.

Get an API key View API docs →

For developers and AI pipelines

Programmatic access to Alibaba Group Holding Limited earnings transcripts and 132,000+ others is available through the EarningsAPI REST API and the hosted MCP server. Quarterly plans from $162 - full transcripts, speaker segments, full-text search, and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.