Jialong Shi

Jialong Shi with Nomura asked about the impact of AI tools like Google Genie on the online gaming industry's competitive landscape, specifically whether it lowers entry barriers or raises success thresholds. He also inquired about NetEase's primary AI investment focus and its differentiated competitive advantages in this field.

Answer

CEO William Ding and VP of Corporate Development Bill Pang explained that while AI lowers entry barriers for game development, it significantly raises the success threshold for top-tier games by shifting the core barrier from production capacity to integration capabilities. They emphasized that NetEase focuses on building game-specific AI expertise rather than general large language models, leveraging proprietary vertical models trained on years of gaming data, focusing on user experience and commercialization, and utilizing versatile talents proficient in algorithms, graphic engines, and game design.

Jialong Shi (Nomura) inquired about the impact of AI tools like Google Genie on the competitive landscape of the online gaming industry, specifically regarding lowered entry barriers versus increased success thresholds for top-tier games. Shi also asked about NetEase's primary focus for AI investments, given other Chinese internet companies' ramp-up in foundation models and computing power.

Answer

CEO William Ding and VP of Corporate Development Bill Pang clarified that while AI lowers entry barriers for game development, it significantly raises the success threshold for top-tier games by shifting the core barrier from production to integration capabilities and valuing soft skills. They noted that world models are still far from practical application for traditional games. NetEase focuses its AI investments on building game-specific AI expertise and highly efficient applications through deep integration, prioritizing high-quality private data and vertical models over general large language models and raw computing power. They highlighted AI-native gameplay intelligence and UGC tools as key areas for commercial value.

Jialong Shi with Nomura observed a smaller number of new games in NetEase's pipeline compared to past years and asked about the current strategy for launching new games. He also inquired about the growth drivers for the online gaming business if fewer new games are launched annually.

Answer

William Ding, CEO of NetEase, stated that the company will focus on refining and concentrating on existing successful products, approaching new projects more prudently to ensure confidence in the content market. He emphasized that being focused is a core competence and that NetEase has a clear vision for future innovations and products, expecting grand breakthroughs in the coming years.

Jialong Shi observed a decrease in the number of new games launched by NetEase annually compared to previous years. He asked about the company's current strategy for new game launches and the overall strategy for driving gaming business growth in the coming years if fewer new games are introduced.

Answer

William Ding, CEO, NetEase, clarified that the company is focusing on refining existing successful products and will be prudent and focused on new projects, ensuring competitive content. He emphasized that focus is a core competence and that NetEase will concentrate resources on high-quality flagship products with strong conviction. Bill Pang, VP of Corporate Development, NetEase, added that NetEase has a clear vision for future innovations and expects significant breakthroughs.

Jialong Shi from Nomura asked for an update on the performance of 'Mystery and Mayhem' in overseas markets, the expected domestic launch timeline, and the company's forward-looking expectations or targets for the game.

Answer

Bill Pang, VP of Corporate Development, reported that 'Mystery and Mayhem' was well-received at its June 26 overseas launch due to its unique art and design. However, he acknowledged that the game faces long-term growth challenges, and the team is actively working on strategies to ensure its sustainable success.

Jialong Shi from Nomura asked about the latest trends in shopping frequency and ARPU for SVIP members, the trend for product return rates, and whether the company's capital return guidance remains intact despite industry challenges.

Answer

CEO Eric Shen (interpreted by Jessie Fan) noted that SVIP metrics are stable, with a slight ARPU decline due to new member mix, and that the product return rate has stabilized with only a low single-digit annual increase. CFO Mark Wang reaffirmed the commitment to return no less than 75% of 2024's non-GAAP net income to shareholders in 2025, highlighting over $400 million returned year-to-date.

Jialong Shi from Nomura asked about the application of generative AI within Vipshop's business to improve profitability or reduce costs. He also inquired about the recent trends in ARPU and shopping frequency for Super VIP (SVIP) members.

Answer

CEO Eric Shen detailed that Vipshop is actively deploying AI internally for personalization, recommendations, and Q&A to improve productivity and efficiency in areas like customer service and brand partner analytics. Regarding SVIPs, he noted that annual ARPU saw a slight decline due to the dilutive impact of a 16% increase in active SVIP members, as new members take time to ramp up spending. However, he emphasized that the ARPU for the core, long-term SVIP cohort remains very resilient.

Jialong Shi of Nomura sought more color on the reasons for the decline in ARPU for SVIP members and asked whether Vipshop's business benefited from government-funded trade-in subsidies.

Answer

CEO Eric Shen, via interpreter Jessie Fan, explained that the small decline in SVIP frequency is being addressed through enhanced membership privileges, expanded discounts, and events to encourage retention and cross-category shopping. Regarding subsidies, he stated the impact was not meaningful, estimating only RMB 400-500 million in additional sales, as the programs target electronics and home appliances, not Vipshop's core apparel category. The company is exploring opportunities should the programs expand.

Jialong Shi inquired about JD.com's strategy regarding artificial intelligence, specifically its deployment of large language models and the expected near-term and long-term impacts on the business.

Answer

CEO Sandy Xu detailed a broad, bottom-up adoption of AI across the company, leveraging its deep supply chain data. On the user front, JD launched an AI shopping assistant, 'Jingyan,' and restructured its search and recommendation systems. For merchants, it provides AI-powered tools for operations, marketing, and customer service. In logistics, AI is used for demand forecasting and robotics in fulfillment centers. Ms. Xu stated that AI is already enhancing efficiency, driving growth, and improving user experience, with its role expected to grow significantly.

Jialong Shi asked if competitors lowering fees poses a risk of merchant loss for Alibaba and also inquired about the GMV impact from government trade-in subsidy programs and the potential for future subsidies.

Answer

Executive Toby Xu addressed the first point, asserting that Alibaba is the most 'merchant-friendly' platform, balancing take rates with merchant health through rebates and supportive policies. Executive Yongming Wu then confirmed that trade-in subsidies since September have significantly accelerated growth in applicable categories like home appliances, with policies and impact varying by region.

Jialong Shi from Nomura asked about potential future divestments of assets like Freshippo, the monetization strategy for the Tongyi Qianwen model beyond compute, and the long-term competitive landscape for AI models in China.

Answer

CFO Toby Xu confirmed the strategy to exit non-core businesses but stated that for a successful asset like Freshippo, the company is not planning a sale but is open to introducing strategic investors. CEO Yongming Wu added that the Tongyi Qianwen model generates revenue via API access and creates significant cross-selling opportunities for other cloud services. He believes the future AI market will see high demand for post-training and customization of foundation models.