Less than a month after the last intensive adjustment around Token Hub, Alibaba's organizational structure related to AI business has undergone another iteration.
On April 8th, Alibaba Group CEO Wu Yongming released an internal letter announcing organizational adjustments related to AI, including the establishment of a new group technical committee, upgrading the Tongyi Large Model Business Unit, and accelerating AI construction.
According to an internal letter, Alibaba has established a technical committee at the group level, led by Wu Yongming, with members including Zhou Jingren, Wu Zemin, and Li Feifei. Among them, Zhou Jingren serves as the Chief AI Architect of the Technical Committee, Li Feifei is responsible for Alibaba Cloud technology and AI cloud infrastructure construction, and Wu Zemin is responsible for the construction of the group's business technology platform and AI inference platform.
Three people, three lines pointing to the model, infrastructure, and inference platform respectively.
You should know that the traditional organizational structure of Alibaba emphasizes "specialization+BU system", but this time Alibaba has brought all the people who can "run the future" to a table and integrated the key links that were originally scattered in the cloud, business lines, and model teams.
A person close to Alibaba told Wall Street that the company used to be good at building people, resources, and business matrices, but this method is no longer feasible in the era of big models. Models need to be trained quickly, reasoning needs to be implemented quickly, business needs to be reused quickly, and any further dispersion of the organization will slow down the entire chain.
Therefore, in the industry's view, the newly established technical committee is a decision-making hub, where the model iterates in which direction, how computing resources are allocated, and how the inference platform is built are all decided at this level.
A noteworthy detail is that in this adjustment, Wu Zemin stepped down as CEO of Taobao Flash Shopping and was succeeded by Lei Yanqin. Wu Zemin is a veteran of Alibaba and also the CTO of the group. Allowing him to withdraw from frontline business management and focus on the construction of the group's technology platform and AI inference platform is a signal that the priority of AI infrastructure has been elevated to business operations within Alibaba.
A similar logic also appears in Li Feifei. He has been appointed as the CTO of Alibaba Cloud and is also responsible for the construction of AI cloud infrastructure.
Alibaba Cloud is the selling port of Alibaba's AI strategy - enterprises need computing power, inference services, and model calling platforms to use large models. What Li Feifei needs to do is to ensure that this pipeline is sufficiently unobstructed.
As the Chief AI Architect and also in charge of the upgraded Tongyi Large Model Business Unit, Zhou Jingren bears the core mission of keeping Alibaba's models at the forefront of the world. The explosive performance of Qwen 3.6 Plus proves the feasibility of this route, but the big model competition has no finale, OpenAI、Anthropic， And domestic companies like ByteDance and Tencent, no one will stop and wait.
Aggregating advantageous forces and resources, investing in the most critical battlefield, indicates that Alibaba has entered a comprehensive combat state of AI. In fact, this is Alibaba's second major organizational change around AI in less than a month.
On March 16th, Alibaba just announced the establishment of the ATH business group - officially known as Alibaba Token Hub - under the direct leadership of CEO Wu Yongming. It includes the Tongyi Laboratory, MaaS Business Line, Qianwen Business Unit, Wukong Business Unit, and AI Innovation Business Unit. A '; Create Token, Transport Token, Apply Token; The complete chain is tied together at the organizational level.
This is a judgment on future business models: the core of a big model is not capability, but consumption. Whoever can make the token flow faster, wider, and more stable will have control over the future of AI cloud.
At the recent Alibaba Group earnings conference call, Wu Yongming mentioned that since 2026, the company has seen some very obvious trends, and large models are beginning to have the ability to complete complex To B workflows. As more and more enterprises begin to use large model driven agents internally to complete end-to-end work tasks, the entire AI and cloud oriented IT budget market undergoes fundamental changes.
Wu Yongming stated that when companies consume tokens, they no longer consider them as IT budgets, but rather as production or research and development costs, as part of the means of production. This is the fundamental internal factor supporting the long-term growth of AI.
Faced with the huge and long-term growth momentum of the AI market, Wu Yongming announced the business goal of Alibaba Group's AI strategy, which is to achieve an annual revenue of over 100 billion US dollars for cloud and AI commercialization, including MaaS, in the next five years.
Regarding the goal of exceeding $100 billion in annual revenue for AI and cloud related businesses in the next five years, the path towards this goal is highly visible from the current market growth space, our existing business and product foundations. ”
Of course, Alibaba is not the only one who has adjusted its course in response to the opportunities of the times. At the same time as Alibaba was undergoing intensive adjustments, Tencent was also reshaping its AI organizational structure.
On March 20th, Tencent issued an internal notice to revoke the AI Lab and merge some personnel into the Large Language Modeling Department, reporting to Chief AI Scientist Yao Shunyu. AI Lab was established in 2016 and is one of Tencent's earliest enterprise level AI labs. Its revocation is aimed at concentrating scattered AI research and development efforts on the main line of hybrid big models.
Tencent President Liu Chiping revealed at a media briefing in mid March that Tencent has undergone intensive team adjustments and workflow restructuring around AI in the past few months.
He said that the next two to three quarters will present; Quantifiable progress;. The new version HY 3.0 of Tencent Hybrid is currently undergoing internal testing and reportedly has significantly improved inference and agent capabilities.
The actions of the two giants are almost synchronized, but the paths are different. Alibaba's approach is to; Institutionalization "; Build a complete business group centered around tokens from scratch, with the CEO personally leading and the five major business units advancing together; Tencent's approach is more like '; Intensive "; Consolidate the scattered AI research and development forces into a unified technological base, making hybrid elements the only entry point for basic models.
Both companies are doing the same thing: eliminating internal AI silos and directing resources in one direction.
This is not a coincidence. The AI competition in 2026 has entered a new stage - it is no longer just about; Should I do AI; The strategic choice question is; Can AI be pushed to the extreme; Comparison of execution power. The ceiling of model capability is still rapidly rising, agents are moving from concepts to products, and enterprise demand is shifting from; Try it out; Becoming '; Comprehensive deployment;.
During this window period, whoever has higher organizational efficiency and faster resource integration will be able to eat the biggest cake.
