As AI development enters the second half of the race, Alibaba is adjusting its organizational structure at a very tight pace. On June 8, Alibaba announced a major strategic upgrade: merging the Tongyi Large Model Business Unit with the Future Life Lab to establish a new Token Foundry Business Unit. This new department will be directly led by the group's CEO Wu Yongming, indicating that Alibaba's layout in the AI field has risen to the highest level of strategic decision-making.

This structural adjustment goes beyond simply "merging similar units," but represents a deep business transformation. The core strength of the former Tongyi Large Model Business Unit will be integrated with AI business lines such as Happy Horse and Happy Oyster into Token Foundry, achieving deep collaboration between model capabilities and application scenarios. At the same time, Zhang Jingren, Alibaba's Chief Scientist, will lead the establishment of the AI Future Research Institute, focusing on original innovation and breakthroughs in cutting-edge technologies.

Looking at Alibaba's recent actions, it is not difficult to see that its "AI flywheel" is accelerating. From the establishment of the Alibaba Token Hub business group, which includes the Tongyi Lab and MaaS business line, in March this year, to the establishment of the Group Technology Committee and the upgrading of the Tongyi Large Model Business Unit in April, and now the establishment of Token Foundry, Alibaba has been constantly aggregating strengths through a series of "small steps and fast runs" in organizational changes, trying to gain an advantage in the most critical battlefield of AI - the synergy between model capabilities and application implementation.

Wu Yongming has repeatedly emphasized in various occasions that Alibaba's core moat in the AI era is to build the "most intelligent" model and drive application scenarios across all industries. With the coordination between Token Foundry and Token Hub, Alibaba aims to bridge the bidirectional influence between models and applications, accelerating the self-evolution of model capabilities through the closed-loop feedback of MaaS (Model as a Service).

The commercial performance also verifies the effectiveness of this strategy. In their letter to shareholders, Alibaba Group Chairman蔡 Congxin and Wu Yongming clearly stated that Alibaba's AI business has left the initial phase of "only investment without output" and has officially entered the commercial return cycle. Data is the most direct proof: in the fourth quarter of the fiscal year 2026, Alibaba Cloud's external commercial revenue growth rate significantly increased, while revenue from AI-related products has achieved triple-digit growth for eleven consecutive quarters. Alibaba expects that the annual recurring revenue (ARR) of AI model and application services, including the Bailian MaaS platform alone, is expected to exceed 30 billion yuan by the end of this year.

In terms of technical achievements, Alibaba has also performed impressively recently. The Qwen-3.7 series of models have established a reputation in the industry with leading programming capabilities; while the HappyHorse-1.0 model has even outperformed others on the global authoritative AI blind testing platform Artificial Analysis, topping both the text-to-video and image-to-video categories.

With the further consolidation of its structure, Alibaba has formed a "full-stack technology system" from the underlying computing power, cloud infrastructure, large model base to the upper-level native AI software. With enhanced organizational cohesion and a clear commercialization path, these adjustments undoubtedly send a clear message to the outside world: Alibaba is fully committed to redefining business boundaries through AI technology, striving to occupy key positions in the global AI competition wave.