Summary of Key Points
Alibaba has recently made another upgrade to its AI-related organizational structure, merging the Tongyi Large Model Business Unit with the Future Life Laboratory to establish the Token Foundry Business Unit (directly led by CEO Wu Yongming), as well as founding the Alibaba AI Future Research Institute (led by Zhou Jingren). These adjustments are crucial steps in consolidating AI resources and enhancing the coordination between models and applications, reflecting the company's strategic determination to strengthen its leading position in the large model industry and accelerate commercial returns. Alibaba's AI business has now entered a commercialization phase, with continuous breakthroughs in model capabilities. The company is advancing towards the goal of "rebuilding Alibaba with AI" through a comprehensive approach that combines organization, technology, commercialization, and talent development.
I. Organizational Structure Adjustments: Consolidating Core AI Forces
The core of these adjustments is to integrate AI business lines and reduce internal inefficiencies:
- Newly Established Token Foundry Business Unit: The former Tongyi Large Model Business Unit (responsible for large model research and development) and the Future Life Laboratory (possibly focused on exploring AI application scenarios) have been merged, under the direct management of CEO Wu Yongming. This moves the teams responsible for creating models and utilizing them closer together to avoid a disconnect between research and application.
- AI Future Research Institute: Led by Zhou Jingren (Alibaba's Chief Scientist), this institute focuses on cutting-edge AI technologies such as more advanced model architectures and AI security, serving as the company's "research brain" for overcoming long-term technical challenges.
- Collaboration with the Previous ATH Business Group: The Alibaba Token Hub (ATH), established in March, was responsible for "creating tokens (models), delivering token-based services (MaaS), and applying tokens to industry scenarios." With the addition of Token Foundry, a complete chain has been formed: "cutting-edge research (Future Research Institute) + core models (ATH) + scenario implementation (Token Foundry)," making the path from laboratory to commercialization smoother.
II. Strategic Intent: Creating an AI "Flywheel Effect" to Consolidate the Competitive Edge
The purpose of these adjustments is to make models, applications, and services mutually reinforcing, forming a growth cycle:
- Wu Yongming has previously stated that "the strongest model is the foundation of a competitive advantage." Therefore, Alibaba first focuses on developing high-quality models (such as the Qwen series) and then sells them through the MaaS platform (Bailian) to businesses. The more businesses use these models, the more data is collected for continuous improvement, which in turn attracts more users—this creates a positive feedback loop.
- By merging the Tongyi Large Model Business Unit with the Future Life Laboratory, model development teams can directly engage with real-world applications, quickly gathering user needs. For example, if businesses require AI to generate videos or code, the development teams can optimize the models accordingly, accelerating this growth cycle.
- The establishment of the Future Research Institute ensures that Alibaba remains at the forefront of AI technology, preparing for future challenges in areas like next-generation large models and multimodal AI (integrating text, images, video, and voice).
III. Commercialization Progress: Moving from "Expending Funds" to "Generating Revenue," Faster than Expected
Alibaba's AI business has moved beyond the stage of heavy investment with no returns and is now in a commercialization phase:
- Data Highlights: In the fourth quarter of the 2026 fiscal year (approximately January to March this year), Alibaba Cloud's external revenue grew by 40%, and AI-related product revenues have been growing at triple-digit rates for 11 consecutive quarters (for example, a 100% increase last year and a 150% increase this year).
- Stable Revenue Growth: The annual recurring revenue (ARR) from AI models and application services, including the Bailian MaaS platform, is expected to exceed 10 billion yuan by June and reach over 30 billion yuan by the end of the year. ARR represents stable annual income; for instance, if a business signs a one-year MaaS service agreement, the entire amount is included in this figure—indicating that Alibaba's AI business has established a steady cash flow.
- Driving Forces: Both underlying model services (such as businesses using Tongyi Qianwen APIs) and upper-layer AI-native software (such as AI-powered office tools and e-commerce solutions) are contributing to this revenue growth.
IV. Model Capability: Evolving from "Good Performance" to "Practical Results," Entering a New Stage of Competition
Alibaba's large models have advanced from being capable of simple tasks like chatting to performing more complex functions:
- Technological Leadership: The Qwen-3.7 model leads the world in coding capabilities, allowing developers to quickly write code; the HappyHorse-1.0 model achieved top rankings in global AI blind tests in text-to-video and image-to-video generation, demonstrating Alibaba's position in the AIGC (Artificial Intelligence Generated Content) field.
- Paradigm Shift: Zhou Jingren emphasizes that large models are shifting from aligning with human preferences to focusing on task objectives (e.g., booking flights, writing reports, analyzing data). Alibaba's models are evolving into intelligent agents that can autonomously plan tasks and collaborate across different tools, which is key for future AI competition.
V. Talent Strategy: Promoting Core AI Professionals to Key Positions
Alibaba's organizational changes also highlight its emphasis on AI talent:
- Promotion of Zhou Jingren: He has been promoted from the head of the Tongyi Large Model Business Unit to Alibaba's Chief Scientist, the highest academic title in the technology department, and is leading the Future Research Institute. His role reflects his critical contribution to the development of Tongyi Large Model and serves as a signal to the industry that Alibaba values AI professionals greatly.
- Internal Motivation: By aligning technical talents with business scenarios (e.g., the Happy Horse team at Token Foundry), innovation is encouraged, as they can more efficiently convert technology into market-ready products.
In summary, these actions aim to transform AI from a scattered set of business units into a central strategic focus for the entire company. Through organizational integration, technological breakthroughs, accelerated commercialization, and talent development, Alibaba is striving to reach the next growth milestone in the AI era. As Wu Yongming puts it, "rebuilding Alibaba with AI" is becoming a reality.