On February 20, Alibaba Group CEO Eddie Wu stated at the earnings conference that Alibaba will soon release a deep reasoning model based on Qwen 2.5-MAX.At the end of January, Alibaba launched the AI foundational large model Qwen 2.5-Max, which is at the industry-leading level in several authoritative benchmark evaluations. Currently, the global derivative models of Qwen have surpassed 90,000, ranking first in the world.Eddie Wu indicated that looking to the future, Alibaba will continue to focus on three major business types: domestic and international e-commerce, AI + cloud computing technology, and internet platform products. In the next three years, Alibaba will increase investment in three areas centered around AI as a strategic core: AI infrastructure, foundational model platforms, and AI-native applications, as well as the AI transformation of existing businesses.First, invest in the construction of AI and cloud computing infrastructure. The AI era has clear and significant demands for infrastructure, and Alibaba will actively invest in AI infrastructure construction, with expected investments in cloud and AI infrastructure in the next three years surpassing the total of the past decade.Second, invest in AI foundational model platforms and AI-native applications. The foundational large models of AI have significant implications for the transformation of industry productivity, and Alibaba will significantly increase R&D investment in foundational AI models to ensure technological advancement and industry leadership, while promoting the development of AI-native applications.Third, invest in the AI transformation and upgrading of existing businesses. For e-commerce and other internet platform businesses, the upgrade of AI technology will bring significant opportunities for enhancing user value, thus Alibaba will continue to increase R&D investment in AI applications and computing power, using AI technology to deeply transform and upgrade various businesses, seizing new development opportunities in the AI era.Regarding the reasons, Alibaba stated that the enhancement of intelligent capabilities is the core axis of this round of AI technology productivity transformation. The standard definition of AGI is the ability to accomplish over 80% of human capabilities. In fact, 50% of global GDP expenditure is the labor wage expenditure of workers. Alibaba predicts that if AGI is achieved, industries related to artificial intelligence are likely to become the largest industry globally, potentially affecting or replacing about 50% of the current GDP composition.Alibaba believes that through the deep integration of cloud and AI, it will become the most important AI infrastructure. Most intelligence is the output tokens behind the models. We judge that 90% of tokens will be generated on cloud computing networks in the future, and only through Alibaba Cloud's globally distributed data centers can we deliver them to global application developers more quickly.Alibaba will adopt a more open mindset and deeply apply AI across the group's applications. Alibaba believes that the application of AI technology in upgrading existing businesses can significantly enhance the value of each business.On the same day, Alibaba announced its quarterly financial report, showing that the company's revenue growth rate in the third quarter reached its fastest pace in over a year, with a significant increase in capital expenditure, accelerated growth in core businesses, faster growth in Taotian revenue, and Alibaba Cloud's revenue returning to double-digit growth, marking the largest increase in about two years. AI-related product revenue has maintained triple-digit growth for six consecutive quartersAlibaba's stock surged over 10% in pre-market trading after its earnings report.The following is the full Q&A:Q1: We have noticed that Alibaba is a leader in cloud infrastructure services and strong foundational AI models, and is well-positioned for the upcoming AI application transformation era, especially in China and possibly beyond. Could management share some insights on how this will translate into cloud revenue growth and trends in cloud gross margins? We understand that the company's capital expenditures this quarter were nearly double that of the previous quarter, reaching 31 billion yuan. At the same time, management mentioned in prepared remarks that capital expenditures over the next three years are expected to exceed the total of the past decade. Could management clarify this comment and help us understand the spending amounts for the next few quarters? Is there a budget for the next three years that can be shared? How will this increase in spending affect overall profitability trends?A1: Before answering the question, I would like to take this opportunity to share why we are making such a significant investment in AI. How do we recognize the opportunities in the AI industry? How do we gradually apply AI in the development process, or how do we monetize during the AI development process? First, Alibaba Group is an important player in the AI field in the Asian market, with the largest cloud service in Asia and the fourth largest globally, advanced self-developed AI models, a thriving open-source ecosystem, and a rich To B application ecosystem. How do we view this AI opportunity? We see AI as a once-in-a-generation industry transformation opportunity, and the primary goal of our AI strategy is to pursue the realization of AGI (Artificial General Intelligence) and continuously push the boundaries of model intelligence capabilities. Why is AGI the primary goal? Because all current AI application scenarios, such as productivity creation, AI search, Chatbot, etc., are merely interim results in the process of enhancing AI capabilities. The enhancement of intelligent capabilities is the core of this round of AI technology productivity transformation.The realization of AGI will create tremendous commercial value. Why is intelligent capability so important? The standard definition of AGI is the ability to perform over 80% of human capabilities. From this perspective, 50% of global GDP spending is actually the labor wages of workers, including both mental and physical labor. Therefore, we anticipate that if AGI is achieved, the AI-related industry will likely become the largest industry globally, potentially impacting or replacing about 50% of the current GDP composition.Second, through the deep integration of cloud and AI, Alibaba Cloud will become the most important AI infrastructure and one of the largest cloud computing networks for outputting AI intelligence, which is Alibaba Cloud's goal. The so-called intelligence refers to the tokens output by modern AI models; in the future, 90% of tokens will be generated and output on cloud computing networks. Such a vast amount of computation can only achieve the highest efficiency on cloud computing networksThrough our globally distributed data centers, we can deliver to global application developers in a faster way.Third, the group will deeply apply AI in various businesses with an open mindset, as we believe that the application of AI technology can enhance the tremendous value of our existing business upgrades. AI technology will bring efficiency improvements, significant growth in user duration and user value to the group's ToB and ToC businesses. This is why we are making such a huge investment. This also reflects the group's strategic determination to invest in AI over the next three years, exceeding the total of the past decade.Regarding capital expenditures, overall, annual capital expenditures will be relatively average, but there will be fluctuations each quarter, mainly influenced by supply chain cycles and IDC (Internet Data Center) construction cycles, which will be flexibly arranged according to actual conditions. How will these capital expenditures affect overall profitability? In general, the next three years will be the largest and most concentrated cloud construction cycle in Alibaba Group's history, therefore, the capital expenditures during this concentrated construction cycle will have a certain impact on hardware amortization and other aspects. However, given the current strong user demand and predictions for the future of the industry, the group believes that the infrastructure investments will soon be absorbed by the demands of internal and external customers.Q2: The recent launch of DeepSeek has brought high-quality and cost-effective models to the entire industry. Since large language models themselves are difficult to monetize through charging, the consumption of large language models has shifted down the value chain to the computing power segment. First, do you agree or disagree with this view? Secondly, to what extent is the computing power of large language models or generative AI not a commodity? Finally, with the help of deep learning models providing high quality and cost-effectiveness, where do you think the most potential areas for AI-native applications lie within the Alibaba ecosystem?A2: Currently, AI technology is still in its early development stage, and the business model of foundational large models is not yet clear. However, as intelligent capabilities improve, there will be more opportunities in the future, and it may even replace some jobs of engineers and scientists, with business models differing from the current ones. Secondly, the differentiation among large models is narrowing, and the gap between open-source and closed-source models is also decreasing. This situation is actually favorable for cloud computing companies, as most models, whether open-source or closed-source, will be deployed on cloud computing networks. For Alibaba Group, whether it is the enhancement of future model capabilities or the current lack of differentiation among models, the cloud computing network is a clear business model. So, whether it is the so-called commoditization of intelligence, I think the most favorable aspect may be the cloud computing network. Imagine if artificial intelligence is the largest commodity in the future, analogous to electricity today, then the cloud computing network is analogous to the current power grid.Within the Alibaba ecosystem, the capabilities of AI models are becoming stronger and evolving faster, making it difficult to determine a single most promising area; every application has very high potential. Internally, we see several obvious opportunities in various aspects. The first is the entry point for consumer life, Taobao can significantly enhance interaction and transaction efficiency with consumers through AI technology transformation, while there is also huge space in many consumption areas such as shopping decisionsRegarding the user duration and value enhancement of Taobao, it is expected to bring greater user value beyond shopping. In the AI ToC field, Quark, as the AI search application with the highest user volume in China, along with Tongyi and other apps, can significantly enhance users' search experience, productivity creation, and work efficiency. In the ToB field, DingTalk, as the group's most important enterprise-level AI application aimed at ToB, will reshape internal collaboration and coordination scenarios in enterprises with the help of AI in the future. Additionally, Gaode has over 170 million independent users logging in or using the service daily. Gaode will also have the opportunity to transform from a navigation tool to a life service entry point through AI technology, further enhancing user duration and value.Q3: We have seen rapid growth in e-commerce revenue this quarter, particularly in AIDC (international e-commerce), including significant improvements in CMR (customer management revenue) for TTG (Tmall, Taobao, Juhuasuan). What are the key strategic plans for TTG and AIDC businesses in the future? The profit margin of TTG business has stabilized; what will be the trend going forward? Regarding AIDC business, the company expects to achieve profitability next year; what are the core drivers for profitability in the short term? In the long term, considering that the competitive environment in the international market is relatively less intense, will the profit margin of international e-commerce business be higher than that of domestic e-commerce?A3: In the domestic e-commerce sector, Alibaba will continue to innovate and optimize user experience to enhance user stickiness. In the new year, we have connected more user payments, and we will continue to invest more on the user side. Over the past year, we have made many moves in commercialization, driving significant revenue growth through payment fee optimization and new advertising models. We will continue to improve user experience. Additionally, the company will continue to optimize merchant operational efficiency and the business environment. Our medium- to long-term goal in the domestic market is to stabilize market share while continuously enhancing user experience and business efficiency.In terms of international business, Alibaba's operations cover B2B, cross-border business, and local platforms. The B2B business is expected to remain stable and bring considerable profits in the coming years. The cross-border business has significantly improved after optimization, and it is expected that the profit and loss situation will improve significantly in the next few fiscal years. At the same time, the company is actively cooperating with local platforms in some countries to improve profitability in the long term. It is currently difficult to determine whether international business will be more profitable than domestic business in the future, but the profitability path for international business is relatively clear.Regarding TTG profit margins, Alibaba's medium- to long-term goal is to maintain a healthy and stable market share. In the past few quarters, including this quarter, we are still in the investment phase, focusing on enhancing user experience, developing new users (including ATA VIP members, etc.), and seizing the window period for integrating other payment methods to expand the user base, so we are still in our investment phase. At the same time, the company is also exploring open-source methods, such as charging software service fees and improving commercialization rates through intelligent product recommendations across the site. Additionally, further investments will be made to optimize the merchant business environment. Therefore, from a profit margin perspective, at this stage, the company is firmly committed to continuously investing in user experience, new users, and the merchant business environment to achieve a balance between income and expenditure, so we are still in an investment phaseQ4: Alibaba's AI-related revenue has achieved triple-digit growth for six consecutive quarters. I would like to know how its profit margin is doing and whether it has reached a considerable double-digit growth? The reason I ask this is that I want to understand the impact on cloud computing profit margins, as our cloud business has consistently had relatively lower profit margins compared to international cloud computing companies. So, as the focus of cloud computing gradually shifts from public cloud to training and then to inference, what impact will this have on our profit margins? What is the outlook for cloud computing profit margins among global peers in the future?A4: Alibaba's AI-related revenue has achieved triple-digit growth for six consecutive quarters, with customer demand consistently exceeding expectations, especially after the Spring Festival. Inference demand has surged, with over 60% to 70% of new customer demand concentrated in inference scenarios, rapidly expanding the customer base, application scenarios, and industry coverage, thereby enhancing the profit margin of AI infrastructure. However, in the next three years, the group will be in a peak construction period, and the depreciation of large-scale investments will have a certain impact on profit margins. At the same time, due to fierce competition in the Chinese market and differences from the global market, domestic cloud computing profit margins may differ from those of other global companies. Nevertheless, because cloud computing is a business model that possesses both scale effects and network effects, the scale effect is particularly critical at this stage. Alibaba believes that large-scale continuous investment will significantly improve hardware procurement and construction costs, and continuous scaling may attract more customers on one hand and improve overall investment costs on the other.Q5: Regarding shareholder returns, as of March 27, we still have $20.7 billion in repurchase capacity. How does management consider our future shareholder returns in the process of investing in AI?A5: In the December quarter, the company repurchased approximately $1.3 billion in stock, reducing the number of outstanding shares by 3.6%. In the previous two quarters, the company had invested $10 billion in repurchases, cumulatively reducing the number of outstanding shares by 4.4% to 4.5%. Therefore, over the past nine months, a total of more than $11 billion has been invested, and the overall number of outstanding shares has decreased by 5.1%. As is common practice in the market, for companies participating in stock repurchase programs, we certainly consider the current stock price when executing the repurchase plan. This is also why you have seen us accelerate the pace of repurchases over the past six months. For example, in the June quarter, we raised $5 billion and used all the funds for stock repurchases. At that time, we believed the stock price was only about $580, far below its intrinsic value. Alibaba effectively manages capital allocation to enhance shareholder returns through the Capital Management Committee established by the Board of Directors. In the future, the company will continue to enhance shareholder returns through dividends, stock repurchases, and investments in high-growth businesses. In the future, Alibaba will continue to use cash to provide returns to shareholders while flexibly adjusting the repurchase plan based on stock price trends. The repurchase capacity approved by the Board of Directors provides the company with greater operational flexibilityQ6: I have two questions, both related to AI and cloud computing. AI has the potential to replace 50% of global GDP, and the cloud is an important underlying infrastructure. In terms of commercialization, aside from infrastructure, how can we capture monetization opportunities in applications like apps or software? Additionally, could you elaborate on the situation of enterprises adopting Alibaba's AI services? I just heard the management mention that the next three years will be the heaviest investment cycle in capital expenditure in the past decade. How will these investments be allocated? How much will be invested in chips? Specifically, how do you view the structure of expenditures, especially how much of the chips will be imported from the United States? How much will be domestically produced? And if the U.S. imposes further restrictions on chips produced in the U.S., what plans do we have to continue implementing our investment plans?A6: We have already discussed significant opportunities for upgrading AI in the ToC space. Alibaba's approach in the ToB sector is twofold. On one hand, in terms of Alibaba Cloud services, we have considerable opportunities on the enterprise side, where internal software systems will gradually be driven by AI. With the development of AI Agent technology, the future SaaS software and internal systems on the enterprise side will become fully intelligent. Traditional isolated enterprise software will gradually transform into a system driven by multiple AI agents that can connect and call each other, helping enterprises optimize daily operational efficiency and deeply participate in key decision-making processes. This trend will bring tremendous upgrade opportunities for various software.On the other hand, DingTalk will become an important platform for enterprise collaboration, with future operations being completed more through natural language interfaces, while traditional ERP/CRM software will serve more as database support. Internal decision-making within enterprises can be achieved through natural language interactions on DingTalk, and enterprises using DingTalk will need to deploy private models to enhance efficiency and security. This combination of AI technology and DingTalk will bring significant changes to the enterprise collaboration market. As for the group's investment in AI infrastructure, Alibaba's cloud computing infrastructure has adapted to various chips, allowing it to flexibly respond to policy changes and ensure the continuous advancement of AI strategies.Q7: Regarding Alibaba's asset disposal, Alibaba has recently divested some non-core retail assets, such as Intime. What is the positioning of assets like Hema, Youku, and Ele.me? Does Alibaba have plans to dispose of or make significant adjustments to these assets like it did with Intime? Secondly, Alibaba Cloud's revenue growth is attributed to strong demand for cloud infrastructure, but the Tongyi model is open-source and free. What is the potential and model for revenue monetization in the future? Additionally, how does management view the future market landscape for large AI models? Currently, the large AI models in the Chinese market are relatively homogeneous; will this decentralized structure continue, or will one or two companies occupy a larger market share?A7: Regarding the exit from non-core assets, we have always strictly implemented a strategy focused on core businesses, actively exiting non-core assets and investments. In the December quarter, we completed the overall exit from RT-Mart and Intime and have obtained antitrust approval, so we expect to complete the related sales in the March or June quarterIn addition, we are gradually exiting some non-strategic minority equity investments, and the market often sees news about our related investment exits. In the future, we will continue to focus on our main business while evaluating existing assets and investments, on one hand exiting non-core investments, and on the other hand focusing on how to better enhance business value, so that it is reflected in the group's valuation.Taking Hema as an example, I believe that Hema has shown steady growth, good expansion, and has achieved decent overall profitability, fully demonstrating the success of the digital retail strategy that integrates online and offline. Currently, we have no plans to sell Hema, but we will consider introducing strategic investors and other options with an open mind, exploring how to better reflect Hema's value in the overall valuation of the group.Regarding Alibaba Cloud, its revenue growth is mainly due to strong demand from customers for AI-related products. Although our Tongyi model is open source, it is not free. We provide paid APIs on the Bailian platform, and although the revenue from paid APIs is still relatively small at present, we believe that as the model's capabilities continue to improve and customer demand grows, its monetization potential will gradually become apparent