Alibaba AI restructures again

```

Less than a month since the last round of intensive adjustments around the Token Hub, Alibaba's organizational structure related to AI business is being iterated once again.

On April 8, Alibaba Group CEO Wu Yongming issued an internal letter announcing organizational adjustments related to AI, including the establishment of a Group Technical Committee, the upgrade of the Tongyi Large Model Division, and accelerating AI development.

According to the internal letter, Alibaba has set up a Technical Committee at the group level, with Wu Yongming as the head and members including Zhou Jingren, Wu Zeming, and Li Feifei. Zhou Jingren serves as the Chief AI Architect of the committee, Li Feifei is responsible for Alibaba Cloud technology and AI cloud infrastructure construction, and Wu Zeming is responsible for the group's business technology platform and AI inference platform construction.

Three people, three paths—pointing respectively to models, infrastructure, and inference platforms.

It's important to note that Alibaba's traditional organizational structure emphasizes "specialization + BU system," but this time Alibaba has invited all the "future runners" to sit at the same table, integrating the key links originally scattered across the cloud, business lines, and model teams.

People close to Alibaba told Wallstreetcn that the company used to be good at stacking people, resources, and business matrices, but this method doesn't work in the era of large models. "Models must be trained quickly, inference must be deployed rapidly, business must reuse swiftly—if the organization remains fragmented, it will slow down the whole chain."

Therefore, in the eyes of the industry, the newly established Technical Committee is a decision-making hub: which direction the models should iterate, how computational resources should be allocated, how inference platforms should be built—all are decided at this level.

One detail worth noting is that in this adjustment, Wu Zeming resigned as CEO of Taobao Flash Sale, replaced by Lei Yanqun. Wu Zeming is a veteran at Alibaba and also Group CTO. Having him withdraw from frontline business management to focus on the construction of the group's technology platform and AI inference platform sends a clear signal: the priority of AI infrastructure at Alibaba has now been elevated above business operations.

A similar logic applies to Li Feifei. He is now Alibaba Cloud CTO, and also responsible for AI cloud infrastructure.

Alibaba Cloud is the "pick-and-shovel" port for Alibaba’s AI strategy—enterprises wanting to use large models need computing power, inference services, and model invocation platforms. Li Feifei's task is to ensure this pipeline is sufficiently smooth.

Zhou Jingren, as Chief AI Architect and head of the upgraded Tongyi Large Model Division, carries the most central mission: to ensure Alibaba's models remain in the world's first tier. The explosive performance of Qwen 3.6 Plus has proved the viability of this approach, but the large model race has no end—OpenAI, Anthropic, and domestic giants like ByteDance and Tencent—no one will stop and wait.

Aggregating key strengths and resources, focusing on the most critical battlefield, shows that Alibaba has entered a full-fledged AI battle mode. In fact, this is the second major organizational change around AI that Alibaba has made within less than a month.

On March 16, Alibaba announced the establishment of the ATH Business Group—all spelled out as Alibaba Token Hub—with CEO Wu Yongming personally at the helm, overseeing the Tongyi Lab, MaaS business line, Qianwen Division, Wukong Division, and AI innovation division. A complete chain of "creating Token, delivering Token, applying Token" is now organizationally bound together.

This is a judgment about the future business model: the core of large models is not capability, but consumption. Whoever can make Tokens flow faster, wider, and more stably will control the future of the AI cloud.

In the recent Alibaba Group earnings call, Wu Yongming said that since 2026, the company has seen some very clear trends—large models are starting to possess the ability to complete complex To B workflows. As more enterprises begin to use large model-driven Agents internally to handle end-to-end tasks, the entire IT budget market for AI and cloud fundamentally changes.

Wu Yongming pointed out that companies no longer regard Token consumption as an IT budget, but as production or R&D costs—seeing it as part of their production materials—which is the most fundamental internal factor sustaining long-term AI growth.

Facing the vast, long-term AI market growth momentum, Wu Yongming announced Alibaba Group’s AI strategic business goal: In the next five years, including MaaS, the annual revenue of cloud and AI commercialization will surpass $100 billion.

"Looking at the future goal of achieving annual revenue of over $100 billion for AI and cloud-related businesses in the next five years, from the current market growth space and our existing business and product foundation, the path to this goal is highly visible."

Of course, Alibaba is not the only one adjusting course in response to opportunities of the era. During Alibaba's intensive reorganizations, Tencent is also reshaping its AI organizational structure.

On March 20, Tencent internally announced the dissolution of AI Lab, with some staff merged into the Large Language Model Division, reporting to Chief AI Scientist Yao Shunyu. Established in 2016, AI Lab was one of Tencent's earliest enterprise-level AI labs. Its dissolution aims to concentrate scattered AI R&D onto the foundational path of Hunyuan large models.

Tencent President Liu Chiping revealed in a mid-March media briefing that in the past several months, Tencent has undergone intensive team adjustments and workflow restructuring around AI.

He said, “In the next two to three quarters, ‘quantifiable progress’ will emerge.” The new version of Tencent Hunyuan, HY 3.0, is undergoing internal tests, reportedly with significant improvements in inference and agent capabilities.

The two giants' actions are nearly synchronized, though their paths differ. Alibaba's approach is "institutionalization," building from scratch a complete business group centered on Token, with the CEO directly in charge and five divisions advancing together. Tencent's approach, meanwhile, is more "intensive," collecting scattered AI R&D forces onto a unified technological foundation, making Hunyuan the solitary base model entry point.

The common denominator is that both companies are doing the same thing: eliminating internal AI silos and channeling resources toward a single direction.

This is no coincidence. The AI race for 2026 has entered a new phase—no longer a strategic question of "whether to do AI," but an execution-level contest of "whether you can maximize AI." The ceiling for model capabilities is rapidly rising, agents are moving from concept to product, and enterprise demand is shifting from "try it out" to "full deployment."

At this juncture, whoever's organizational efficiency is higher, whose resource integration is faster, will get the biggest slice of the cake.

Risk Warning and DisclaimerThe market involves risk, and investment must be cautious. This article does not constitute personal investment advice, nor does it take into account the special investment objectives, financial status, or needs of individual users. Users should consider whether any opinions, perspectives, or conclusions in this article suit their specific situations. Investments based on this are at your own risk. ```