Alibaba (Minutes): The Leader of China's AI Narrative
The following is $Alibaba(BABA.US) the minutes of the conference call for the third quarter of fiscal year 2025. For the interpretation of the financial report, please refer to《 AI to the Rescue, Is Alibaba Resurrected? 》
I. Core Financial Information Review
1. Overall Financial Performance: Q4 (excluding Alibaba's consolidated subsidiaries) grew by 11% year-on-year. Total revenue increased by 8% year-on-year, reaching RMB 280.2 billion; EBITDA grew by 4% year-on-year, reaching RMB 54.9 billion, mainly due to revenue growth and improved operational efficiency, although some growth was offset by increased investments in e-commerce; excluding the long-term cash incentive plan, adjusted EBITDA would have increased by 5% year-on-year; Non-GAAP net profit grew by 6% year-on-year, reaching RMB 51.1 billion; GAAP net profit increased by 333% year-on-year, reaching RMB 46.4 billion, primarily due to growth in operating income, gains from equity investments measured at fair value, and returns on equity investments; operating cash flow grew by 10% year-on-year, reaching RMB 70.9 billion. Free cash flow decreased by 31% year-on-year, amounting to RMB 39 billion, mainly due to cloud infrastructure capital expenditures. As of December 31, 2024, the net cash position was RMB 378.5 billion, or USD 51.9 billion.
II. Detailed Content of the Financial Report Conference Call
2.1 Key Information from Executive Statements
1. Strategic Focus and Business Momentum:
Adhering to a user-first AI-driven strategy, focusing on e-commerce and AI + cloud as the two core businesses.
1）Core Business: E-commerce and AI + cloud business are accelerating growth, with strong fundamentals in core business performance. The divestiture of offline assets is nearly complete. Q4 continued to advance the AI + cloud integration strategy in the cloud business, leveraging an industry-leading AI product portfolio.
2）AI + Cloud Strategy: Utilizing industry-leading AI products (such as the Q1 2.5 MAX foundational model), driving AI-related revenue to achieve triple-digit quarter-on-quarter growth for six consecutive quarters. Over 90,000 derivative models and 290,000 companies/developers are using the Q1 API, with deep reasoning models based on Q1 2.5 MAX set to be launched soon 3) E-commerce: Taobao/Tmall focuses on user growth and experience enhancement, with 88 VIP members reaching 49 million, and CMR increasing by 9% year-on-year.
- Outlook: AIDC is expected to achieve its first profit in the next fiscal year. Other internet platform businesses continue to improve operational efficiency, with Amap achieving profitability this quarter. The goal for most loss-making businesses is to achieve breakeven within 1-2 years.
2. Future Investment Focus (3-Year AI Strategy):
Alibaba is confident in focusing on e-commerce and AI + cloud strategy and will continue to concentrate on three major business categories: domestic and international e-commerce, AI plus cloud computing, and internet platform businesses, while strengthening investments in the following three areas:
AI and Cloud Computing Infrastructure: Actively investing to meet the demand for AI/cloud computing, with total investment in the next 3 years expected to exceed the total investment of the past decade.
AI Foundation Models and AI Prototype Applications: Increasing R&D efforts to maintain technological leadership and develop AI-native applications.
Business Transformation through AI: Integrating AI into e-commerce and platforms to enhance user value and seize growth opportunities.
3. Asset Disposal and Shareholder Returns:
This quarter, an agreement was reached to dispose of all shares held in Gaoxin, amounting to $1.6 billion, as well as to dispose of Intime shares for $1 billion. Approval for the above sales has been obtained from the State Administration for Market Regulation of China, with the main financial impact expected to be reflected in the March quarter; in December 2024, a total of $1.3 billion worth of shares were repurchased, representing a net reduction of 0.6% of total shares, combined with approximately $10 billion worth of shares purchased in the first half of this fiscal year, achieving a net reduction of 5% in total shares over the past 9 months; in November 2024, a local currency bond issuance was completed, raising approximately $5 billion through a combination of notes with a face value of $2.65 billion and RMB notes with a face value of 17 billion yuan, to reduce overall financing costs.
2.2 Q&A
Q: How does Alibaba translate its cloud AI into financial revenue growth and profit margin trends? What is the Capex amount for the next few quarters, and is there a budget for the next three years that can be shared? How will Capex affect overall profitability trends?
A: Alibaba has a significant advantage in the AI field, especially in the Asian market, being the leading cloud service provider in the region, ranked fourth globally and first in Asia, with leading models, technologies, proprietary AI models, and a thriving open ecosystem, along with numerous AI application scenarios throughout the ecosystem.
We believe this AI opportunity represents an industry transformation that occurs once in decades. For Alibaba Group, pursuing AGI is the primary goal. The productivity of our existing scenarios, whether in creation, AI search, or Chatbot, are all opportunities arising from the enhancement process of AGI. Therefore, we believe that the enhancement of intelligent capabilities is the core storyline of this round of AI productivity transformation If AGI is realized, it could have a huge impact on the global industrial structure, potentially affecting or replacing 50% of the global GDP. It is estimated that over 90% of tokens will be generated on cloud computing networks, as the computational volume of such a large model can only achieve the highest efficiency on cloud computing networks. Through our globally distributed data centers, we can deliver to global applications in a faster manner. Furthermore, we will adopt an open attitude to deeply integrate AI with our business scenarios, improving efficiency, increasing user engagement time, and creating more user value. This is why the total investment in cloud and AI over the next three years will exceed the total of the past ten years. The annual Capex level over the next three years will be roughly equivalent, but there may be fluctuations between quarters each year due to supply chain delivery and data center construction time. The next three years will be the period of the most concentrated and highest level of investment in cloud and AI-related infrastructure, and while the depreciation of hardware infrastructure will have an impact, the demand from both internal and external customers is expected to be huge and grow rapidly.
Q: The recent launch of the DeepSeek large language model family has brought high-quality products to the entire industry at an affordable cost, and since the models are free, the monetization of large language models has shifted down the value chain to the computing power segment. Do you agree with this statement? To what extent is computing power a commodity or not? With the high quality and cost-effective models from DeepSeek, where do you see the most potential for AI-native applications both within and outside the Alibaba ecosystem?
A: Although artificial intelligence technology is developing rapidly, it is still in its early stages, and the future business models and monetization methods for models are not yet clear. As models become increasingly intelligent, there will be more monetization opportunities, but the specific methods are still unclear. From the current models, the differentiation between different suppliers' models is narrowing, and the differences between open-source and closed-source models are also not significant. However, all these models, whether open-source or closed-source, need to be hosted on cloud computing networks, so the most clear monetization path at present is the cloud computing services provided by Alibaba Cloud for hosting and supporting the operation of these models. If we compare artificial intelligence to electricity in the new era, then Alibaba Cloud is like the power grid that delivers electricity. Regarding the most promising areas for AI applications within the Alibaba ecosystem, it is difficult to pinpoint which areas have the highest potential due to the rapid development of models and their increasingly powerful functions; all applications may have huge potential.
Internally, we are focusing on several opportunities: 1) The entry point for consumer life, the transformation of AI technology. For example, in the Taobao application, a large amount of internal development is underway, with many projects about to be launched. The AI-enhanced features within Taobao will increase consumer engagement and transaction efficiency, and beyond shopping, AI functions will further create more value; 2) AI to C field, mainly Quark and Tongyi APP. Quark is the most widely used AI search product in China, and it is deploying AI to improve search, productivity, creativity, and overall efficiency; 3) AI to B assets, DingTalk is also deploying AI to redefine the enterprise collaboration experience; Amap is currently mainly used as a navigation application, with 170 million daily active users in China. As it integrates AI more deeply, it will become an entry point for lifestyle and local services, increasing user engagement time Q: What changes have occurred in domestic e-commerce (TTG) and international e-commerce (AIDC) businesses, and what are the key initiatives for the future? TTG's profit margin has stabilized; how should we view the subsequent trends? What are the short-term key drivers for AIDC to achieve profitability next year? In the long term, will international business be more profitable than domestic business?
A: In terms of domestic e-commerce, the company has been pursuing improvements in user experience, innovating and optimizing around user experience to enhance user stickiness. In the coming year, further user growth is expected through investments. Recently, measures such as charging payment processing fees have been implemented to improve monetization levels, but the company will continue to invest in enhancing user experience and optimizing the merchant operating environment. The overall goal is to stabilize market share while improving user experience and optimizing merchant operational efficiency.
In the international e-commerce business, various business models have been integrated, including B2B cross-border and local platform operations. It is expected that the international business will maintain a stable trend in the coming years, moving towards achieving significant profitability at scale. After optimizing the business model, the unit economic benefits of B2C business have significantly improved, and profitability is expected to increase significantly in the coming quarters. In some countries, the company is collaborating with local platforms to enhance profitability. Currently, it is impossible to determine whether international e-commerce will be more profitable than domestic e-commerce in the long term, but there is confidence in seeing a noticeable improvement in the profit and loss of cross-border business in the coming quarters.
Regarding TTG's profit margin, the company's investment focus in recent quarters has been on acquiring a healthy and stable market share. This quarter, it continues to invest in enhancing user experience, acquiring new users, and the core user group of 88VIP, while also integrating new payment methods to attract new users. While investing, the company is also actively exploring ways to increase revenue, such as charging software service fees and launching smarter marketing products. The company is still in the investment phase and will continue to invest in enhancing user experience and acquiring new users; profit margin is always a balance between revenue and expenditure.
Q: Alibaba's AI revenue has seen triple-digit growth for six consecutive quarters; how should its scale be quantified? Does this revenue segment achieve double-digit profit margins? As the cloud business transitions from public cloud to training and now increasingly to inference, how should we view its growth and cloud business profit margin prospects compared to global peers in the long term?
A: Alibaba's AI-related revenue has achieved triple-digit growth for six consecutive quarters, with customer demand for AI and related products continuing to grow, and the growth rate exceeding expectations. Since the Spring Festival, demand for inference has surged, with approximately 60% - 70% of new demand being for inference. With the rapid expansion of demand, there is potential to broaden the customer base and industry coverage, which will enhance the profit margins of AI services. However, in the next three years, the company will make the highest Capex investment in its history, and amortizing these investments over the coming years will impact profit margins. Due to intense competition in the Chinese market and differing market dynamics, profit margins in the Chinese market will differ from those in the international market. Cloud services have strong economies of scale and network effects, with economies of scale being particularly important at this stage. As the scale expands and the number of customers increases, it can better optimize construction costs.
Q: Alibaba still has $20.7 billion available for stock repurchases. Considering the progress and investment plans in AI, how will these funds be deployed?
A: In the December quarter, the company conducted $1.3 billion in stock repurchases, resulting in a net reduction of 0.6% in the number of shares; over the past nine months, the number of shares has been reduced by 5%, achieving significant progress in stock repurchase. As a common market practice for companies conducting stock repurchase programs, current stock prices are considered when executing repurchase plans. Therefore, the pace of repurchases was accelerated in the first six months. For example, in the June quarter, $5 billion in debt financing was raised and fully used for stock repurchases, as the stock price was only $80 at that time, which the company believed was severely undervalued. The company has established a Capital Management Committee at the board level, aimed at optimizing capital allocation to enhance shareholder returns. This committee has consistently provided strong management for shareholder return initiatives and will continue to do so. In the future, the company will continue to use cash effectively and optimally to enhance shareholder returns, executing stock repurchases based on the board's allocation and guidance, while also considering stock price conditions. Additionally, shareholder returns will be enhanced through dividends, stock repurchases, and investments in high-growth, high-potential areas.
Q: Besides infrastructure represented by cloud services, how does Alibaba plan to monetize the huge potential opportunities brought by artificial intelligence, and what is its strategy for developing applications or software? Furthermore, can you elaborate on the adoption of artificial intelligence on the enterprise side?
A: On the consumer side (2C), there are clear opportunities for AI applications; on the business side (2B), Alibaba Cloud Intelligence also has many opportunities. In the future, SaaS software will increasingly be driven by AI agents, and internal enterprise systems will resemble a network of multiple AI agents collaborating and calling upon each other to provide services, even assisting enterprises in making important decisions. Not only does the SaaS layer of software have upgrade opportunities, but the underlying PaaS layer also has many upgrade opportunities. Taking DingTalk as an example, many functions will be achieved through natural language interaction with the help of artificial intelligence in the future. Today's CRM and ERP-type systems will become more like data input sources, and enterprise meetings and decisions can be processed through natural language on DingTalk, with many opportunities for integrating various AI agents into DingTalk. Many companies have a large amount of proprietary data and their own processes that need to leverage AI agents to improve the efficiency of these processes. Overall, artificial intelligence will empower and reshape various enterprise software applications, presenting a huge opportunity.
Q: How do you view Alibaba's investment allocation plan for the next two years? How much will be invested in chips, and what is the ratio of imports of chips from the United States to domestic chips? If the U.S. further implements export restrictions, what contingency plans does Alibaba have to continue its investments?
A: I cannot provide specific allocation percentages for Capex across various fields, as this requires detailed breakdowns of the costs of different devices in data centers (IDC). However, Alibaba Cloud's design is compatible with various chips, so regardless of how future policies change, it will not affect Alibaba's implementation of its investment plans Q: After successfully divesting the physical retail assets of GaoXin Retail and Intime, are there plans to divest other assets in the future, including Hema, Ele.me, and Youku?
A: The company has been strictly implementing the strategy of focusing on core businesses and exiting non-core businesses. In the December quarter, it was announced that the complete sale of GaoXin Retail has been finalized, and Intime has passed antitrust review, with the transaction expected to be completed in the March or June quarter. The company will continue to seek opportunities to exit non-strategic equity assets. As for Hema, it has achieved good growth, expansion, and profitability, serving as a model for the innovative business model of integrating online and offline retail. Currently, there are no plans for sale, but the company remains open to ways to enhance its value, such as introducing strategic investors, which requires further exploration.
Q: The revenue growth of Alibaba Cloud is largely driven by AI-related computing demand. Besides computing, does the Tongyi Qianwen model itself have other monetization pathways and profit models in the future? Currently, the Chinese AI market lacks differentiation among different models; will there be 1-2 large models occupying a significant market share in the future?
A: The strong revenue growth of Alibaba Cloud is primarily driven by the huge demand from customers for AI products. Tongyi Qianwen is an open-source product, but using the model through API access is charged, and the revenue generated from this is currently relatively low. However, as the model's capabilities improve, it may increase the charges. Additionally, many developers use Tongyi Qianwen as a foundational model to develop their own vertical models and applications, which will naturally be deployed on Alibaba Cloud; when customers access the Tongyi Qianwen model via API, the company also has many opportunities to cross-sell other cloud products to them.
Currently, AI development is in a rapidly iterating early stage, making it difficult to determine the final market landscape. The gap between different foundational models is narrowing, and there will be significant demand for post-training around models in the future, that is, customizing models based on different industries, use cases, and private datasets, which has huge market potential. Various professional models and vertical models need to be hosted on the cloud, so the company is excited and confident about the prosperous development of the open cloud ecosystem.
Risk disclosure and statement of this article: Dolphin Research Disclaimer and General Disclosure
