Alibaba Unifies AI Operations Under New Token Hub Division

Alibaba Group formed a new business division called Alibaba Token Hub (ATH) on March 16, consolidating five AI units under the direct leadership of CEO Eddie Wu Yongming. The move signals the Chinese tech giant’s most aggressive structural bet on artificial intelligence to date.

The new group merges Tongyi Laboratory, which develops Alibaba’s Qwen series of foundation models; the model-as-a-service (MaaS) business line; the Qwen consumer AI assistant unit; a newly created enterprise division called Wukong; and an AI Innovation unit focused on emerging applications. ATH will sit alongside Alibaba Cloud and e-commerce as a top-level business group in the company’s corporate hierarchy.

In an internal letter shared with employees, Wu framed the restructuring around a single concept: the token, the fundamental unit of computation that large language models produce when processing and generating text, images, or code. “ATH is built around a single organising mission: create tokens, deliver tokens and apply tokens,” Wu wrote. He added that the industry is approaching an “inflection point” in artificial general intelligence, with advances enabling billions of AI agents to handle increasing volumes of digital work.

Talent Shake-Up Preceded the Restructuring

The reorganization follows a turbulent period for Alibaba’s AI leadership. Lin Junyang, head of Alibaba’s Qwen model team, submitted his resignation on March 3 and publicly announced his departure the following day. Yu Bowen, head of post-training, departed the same day. Another senior contributor, Huibin, who led Qwen Code, had already left for Meta in January 2026.

The departures triggered an emergency meeting between Tongyi laboratory leadership and senior executives. On March 5, Wu issued a letter to the Tongyi AI research lab reaffirming Alibaba’s commitment to open-source AI models and pledging to scale up research investment and computational resources. The company also recruited Zhou Hao, a former senior staff research scientist at Google DeepMind, to lead post-training research.

Alibaba then formed a Foundation Model Task Force consisting of Wu, Alibaba Cloud CTO Zhou Jingren, and Group CTO Wu Zeming to coordinate group-wide resources for accelerating model development. The Token Hub announcement two weeks later effectively institutionalized that task force into a permanent structure.

What ATH Means for Alibaba’s AI Strategy

The consolidation addresses a fragmentation problem that had grown over multiple restructurings. Alibaba’s Qwen consumer business was first spun off from Alibaba Cloud in November 2024, then merged with another unit in December 2025 under Vice President Wu Jia. The Token Hub now brings all AI-related teams under one roof with direct CEO oversight.

The new Wukong unit, named after the mythological Monkey King, will build an AI-native enterprise workflow platform around Alibaba’s DingTalk workplace app, which competes with Slack and Microsoft Teams in China. The AI Innovation unit will focus on rapid experimentation with new AI applications and business models, including consumer products with agentic capabilities for tasks like food ordering and travel booking.

The structural shift comes as Alibaba ramps up spending. In February 2025, the company announced plans to invest at least RMB 380 billion (approximately $53 billion) in cloud and AI infrastructure over three years, a sum exceeding its total spending in the category over the prior decade. That investment is already showing returns: Alibaba’s Cloud Intelligence Group posted 34% year-over-year revenue growth in the September 2025 quarter, reaching RMB 39.8 billion ($5.6 billion), with AI-related product revenue achieving triple-digit growth for a ninth consecutive quarter.

On the consumer side, the Qwen app hit 100 million monthly active users within about two months of its November 2025 public beta launch, making it one of the fastest-growing AI apps globally.

The question now is whether centralizing under ATH can stabilize a division that has lost key talent while simultaneously scaling across research, enterprise, and consumer markets. Alibaba’s Hong Kong-listed shares have fallen roughly 20% from a recent high, reflecting investor unease about the pace and cost of the AI transition. Wu’s decision to personally lead ATH suggests the company views this as an existential priority, not a side bet. Whether that direct CEO attention can hold together an ambitious AI operation mid-restructuring is the central test ahead.