Alibaba (09988) Q2 earnings call: Strong growth in GMV of Taobao and Tmall on Double 11 for 24th consecutive year, will continue to invest in e-commerce AI computing power.

EBITDAGMV GMVLong-term investment made by business services. Some measures have been taken to reduce the burden on businesses. Do these measures make it more attractive for competitors to attract merchants? If so, is it possible that it will lead to the loss of our 88VIP merchants? First of all, Taobao and Tmall have always placed a strong emphasis on the rights of merchants. From the perspective of platform development and operations, we believe that our commercialization rate and merchant friendliness are among the best on all platforms. This is also why we believe there is still room for improvement in commercialization rate, but we will balance the relationship between platform health, merchant operation health, and commercialization rate. Secondly, we have implemented many measures to alleviate the burden on merchants. For example, while charging technical service fees, we have also implemented exemption plans for small and medium-sized merchants and canceled the annual fee for Tmall merchants. In addition, we have introduced products such as "Return Guarantee", which greatly reduces the costs for merchants and protects their legitimate rights in refund optimization and other aspects. In fact, we have been continually investing in enhancing user experience, including enhanced member benefits, in the past and future stages. These measures will increase users' willingness to make purchases on our platform, helping merchants increase their transaction rates and make their businesses more prosperous on the platform. I believe that these investments will greatly benefit merchants. Q7. Regarding the subsidy policy for trading in old products for new ones that started in September, I would like to know the contribution or proportion of this policy to the entire e-commerce platform ecosystem or GMV. You mentioned earlier that the September subsidy for trading in old products for new ones was just the beginning of the government's stimulus consumption policy, and there may be more similar subsidies in the future. How do you view the intensity of future subsidies and the possible focus on subsidizing or supporting consumption in specific areas? The subsidy policy for trading in old products for new ones has significantly contributed to the growth of related industries on our platform in the past few months, with a noticeable increase in growth rates in these industries since September. As for subsidy policies, although the specific measures vary by region, we have observed a trend: in addition to major appliances, many regions have increased subsidies for categories such as small appliances, home decor, household goods, and even digital products. Q8. Regarding the business of Taobao and Tmall, we began charging technical service fees starting in September. The next quarter will be the first full quarter of us implementing this fee. If our take rate has already remained stable in the previous quarter, is it possible for our commission income growth to exceed GMV growth in the future? In other words, will the take rate not only remain stable but also increase? In terms of capital expenditures (CAPEX), I noticed that our CAPEX increased this quarter, mainly for investment in Alibaba Cloud and AI fields. I would like to know how the management views the return on these investments? Especially in the application end and the infrastructure construction in cloud computing, how do we evaluate the return on these investments? Our revenue conversion rate is influenced by several positive factors, mainly the introduction of software service fees and the penetration rate increase in full-site promotion. However, we are also investing in some new models that have rapid initial growth but relatively low commercialization rates. Therefore, these rapidly growing categories may dilute the overall revenue conversion rate, leading to a trade-off relationship. We will consider the overall revenue conversion rate of Taobao and Tmall and seek a balance on the merchant side to provide a more favorable operating environment and a more balanced business atmosphere. Therefore, we will focus on the level of the overall revenue conversion rate and consider the interaction of these factors. Looking ahead, we are optimistic about the development of the company. Currently, our positioning in the market is relatively low, so we have a lot of room for growth. Our significant capital expenditures are concentrated in the field of cloud computing, especially in AI infrastructure. These investments are based on our in-depth understanding and judgment of short-term and long-term market demands. Currently, there is a rapid growth in customer demand for AI computing power and AI model API services, and these demands are not yet fully met. Therefore, we are quite proactive in our investments in the AI field. We believe that the current development of the AI industry, especially the rise of generative AI, is a rare historic opportunity, possibly a once-in-a-generation technological innovation. Whether from the capabilities of existing models or their applications across various industries, the demand is steadily increasing and very clear. We have also observed that new models like OpenAI have a significant increase in the demand for reasoning computing power through cognitive chain (COT) technology. Based on these technological trends, we are investing early in AI computing power infrastructure. Therefore, we hold an optimistic view on the short-term and long-term demand for AI computing power, which is the fundamental reason for our active investment in AI infrastructure. Q9. Regarding shareholder returns, first of all, congratulations to Alibaba for successfully repurchasing $100 billion of stocks in the past 6 months. I would like to ask, when executing the repurchase, do you differentiate between ADR and Hong Kong-listed stocks (code 9988)? Secondly, the People's Bank of China recently launched a swap project that allows companies to borrow money to repurchase their shares. Since Alibaba has joined the southbound Hong Kong stock connect program, can this tool be used to increase share repurchases? Our stock repurchases are mainly conducted in the US market, as the liquidity is stronger there. Although we also have repurchases in the Hong Kong market, the focus is currently on the US market. We have been exploring different repurchase opportunities and actively raising funds to support repurchase activities. As for the swap project by the People's Bank of China, we have not yet utilized this channel as it mainly targets A-share listed companies. For example, in May, we issued $5 billion in convertible bonds and used the funds in June to conduct a $5.8 billion stock repurchase. This demonstrates that we will explore various avenues to increase repurchases. Additionally, our company holds ample Chinese yuan cash in mainland China, so there is less interest in repurchasing with Chinese yuan funds. However, for offshore yuan or US dollar funds, we may be more interested. Q10. Regarding the opportunities in the AI field, currently in cloud computing, is the main DRIVER of AI revenue model training or inference services? In the US, there is a lot of discussion about AI agents; will this drive growth in related workloads? In the AI field, the main driver of revenue for AI in cloud computing is currently a combination of model training and inference services. The trend in the US towards AI agents is expected to drive growth in related workloads, as these agents require both model training and inference services to function effectively. I hope this helps. Let me know if you need any more assistance.In the past, the demand for cloud services in the field of generative AI was mainly driven by model training in the early stages. However, we have observed a continuous growth in the demand for inference services and computational power. In the future, the demand for model training may be concentrated in a few large companies, especially in the area of foundational large models. At the same time, we also notice that various industries, such as autonomous driving and finance, have their own vertical domain model training needs.Overall, the demand for both model training and inference services is increasing. However, in the future, the AI inference demand may account for a larger proportion of computing power demand, leading to a greater growth. The application of AI in various industries, as you mentioned, is indeed widespread. Different companies and enterprises are developing new AI skills, utilizing AI to automate workflows, and even using AI to reconstruct models that previously relied on CPUs. These changes are actively happening around us, including within our company. In general, the technological development trends we observe are similar to those in the United States, with a significant shift from CPU-based computing to GPU restructuring for a large number of computing needs. This shift is driven by the widespread application of AI models.