Wedoany.com Report on Mar 17th, On the eve of AGI (Artificial General Intelligence), Alibaba is responding to the call of the times with a profound organizational transformation. On March 16, Alibaba Group CEO Wu Yongming officially announced the establishment of the Alibaba Token Hub (abbreviated as ATH) business group through an internal company-wide letter, unveiling a brand-new AI strategic formation.

According to information disclosed in the internal letter, this new business group is directly overseen by Wu Yongming, with its core mission succinctly summarized as "Create Token, Transport Token, Apply Token." The establishment of ATH is seen as a key strategic gathering for Alibaba in the era of AI Agents, integrating the core forces across Alibaba's AI landscape into a unified framework. The five core components incorporated into ATH include: Tongyi Lab, MaaS business line (Bailian), Qianwen Business Department, AI Innovation Business Department, and the "Wukong Business Department," which makes its first public appearance and is positioned as a B-end AI-native work platform.

After years of divergent exploration in the field of large models, Alibaba's current adjustment of "clenching fingers into a fist" sends a clear signal to the outside world: through vertical integration, all core AI forces are being brought under a single strategic perspective to face the upcoming AI Agent era with more concentrated firepower.

Judging from the composition of the ATH business group, Alibaba has incorporated almost all its core assets related to large models and AI applications into this new organization. Most notably is its reporting structure—directly overseen by Group CEO Wu Yongming. In the context of major internet companies, a CEO personally leading a single business group typically signifies that the business has become a "top-priority project" crucial to the company's future survival. In the past, Alibaba's AI capabilities were scattered across Cloud, DAMO Academy, and various business lines. Although the "Tongyi" series of models have gained considerable traction in both open-source and closed-source domains, they still faced challenges common to large companies, such as "departmental silos," in terms of application deployment and internal coordination. The establishment of ATH is, in essence, a shortening of the management path. By forcibly binding underlying model R&D, infrastructure, and upper-layer applications within the same evaluation system, Wu Yongming is attempting to establish a flatter decision-making mechanism. This structure, with the group CEO directly overseeing operations, grants ATH the highest priority in compute resource allocation and talent mobilization. This is not only to improve response speed but also to establish a more deterministic execution capability during the window of opportunity in Agent competition.

The naming of the new organization as "Token Hub" is highly significant. In the past, major companies were accustomed to naming business groups with terms like "Cloud," "Intelligent," or "Data." This time, Alibaba directly taps into the most fundamental concept of large language models—Token. Wu Yongming previously articulated this logic clearly in internal activities: "A vast amount of digital work will be supported by tens of billions of AI Agents, and these AI Agents will operate on Tokens generated by models." If traffic was the lifeblood of commerce in the mobile internet era, then in the AGI era, Token is the new "water, electricity, and gas." Alibaba defines ATH's goal as "Create, Transport, Apply" Token, essentially reconstructing a complete AI supply chain internally: Tongyi Lab is responsible for R&D of foundational models, continuously raising the upper limit of model capabilities, akin to a "power plant"; the MaaS business line builds a model service platform, enabling compute power and model capabilities to be distributed efficiently and cost-effectively across all industries, akin to a "power grid"; while the Qianwen, Wukong, and Innovation Business Departments are responsible for transforming model capabilities into specific C-end and B-end products, akin to "household appliances and industrial equipment." With Token as the main thread, Alibaba no longer views AI as an isolated technological tool but rather as a brand-new factor of production that runs through the entire lifecycle from underlying compute power to upper-layer applications.

In this organizational chart, the debut of the "Wukong Business Department" has attracted significant industry attention. The internal letter clearly defines its positioning: "to build a B-end AI-native work platform, deeply integrating model capabilities into enterprise workflows." As large model technology evolves, the industry's competitive focus is shifting. Although C-end personal assistants hold great potential for traffic, it is often the transformation of B-end enterprise workflows that can achieve a commercial closed loop and generate sustained consumption. Traditional enterprise software is undergoing an AI-driven rewrite. Enterprises no longer need cumbersome menus and forms but intelligent agents that can understand intent, automatically invoke tools, and execute tasks across systems. By launching "Wukong," Alibaba clearly sees this incremental market. Leveraging Alibaba Cloud's base of government and enterprise customers and DingTalk's office scenarios, "Wukong" shoulders Alibaba's core task of converting AI capabilities into B-end commercial returns. This layout complements the Qianwen Business Department. Qianwen focuses on accumulating scale in C-end interaction, while Wukong attempts to establish a higher consumption barrier in enterprise production processes. Their commonality lies in the fact that they are both outlets for Tokens at the application layer, and their growth directly determines the operational efficiency of the entire supply chain.

"We are currently on the eve of the AGI explosion." Wu Yongming's statement in the internal letter conveys a sense of urgency. In 2026, competition in the AI industry has moved from the early parameter race into the deep waters of Agent ecosystem and workflow implementation. Major internet giants are intensively focusing efforts on the application layer, trying to secure their positions in the Agent era. At this critical juncture, Alibaba has completed its formation switch for the AI 2.0 era by establishing ATH. The depth of this vertical integration determines Alibaba's reaction speed when facing scaled demand. While most companies in the industry still adhere to traditional product management logic, Alibaba has already begun experimenting with redefining itself using Token logic. This strategic realignment is not only about Alibaba's market position in the AI era but will also provide the industry with a sample reference for AI commercialization paths through this attempt at a full-chain closed loop.