← Back to stories Smartphone displaying AI app with book on AI technology in background.
Photo by Sanket Mishra on Pexels
凤凰科技 2026-04-11

Alibaba (阿里巴巴) bets its future on “Token” — ATH, a new tech committee and leadership reshuffle

Tightening control around Token as strategic center

Alibaba (阿里巴巴) has moved aggressively to centralize its AI efforts under a Token-driven strategy, it has been reported that. In the space of two months the group created the Alibaba Token Hub (ATH), upgraded Tongyi Lab (通义实验室) into a full business unit, announced a five‑year, $100 billion revenue ambition, and established a Group Technology Committee led by CEO Wu Yongming (吴泳铭). The changes follow the high‑profile exit of Qwen team lead Lin Junyang (林俊旸) and, reportedly, internal alarm about fractured technical coordination.

The signal is clear: Token — the usage unit that drives calls to large models and agent frameworks like OpenClaw — is now, in Wu’s words, “the electricity of the AI era.” It has been reported that daily domestic Token calls have surged to about 140 trillion, underlining why Alibaba is treating Token production (models), delivery (infrastructure) and application (B2B and B2C services) as a single supply chain.

New roles, clearer responsibilities

Alibaba has reallocated senior roles to map onto that supply chain. Zhou Jingren (周靖人) becomes chief AI architect and will lead the upgraded Tongyi large‑model business; Fei‑Fei Li (李飞飞) is reported to be Alibaba Cloud (阿里云) CTO, focusing on compute, inference services and the commercial pipeline for Token delivery; and Wu Zeming (吴泽明) has shifted from consumer operations to concentrate on group‑level AI infrastructure and inference platforms. ATH is now a first‑tier business group alongside cloud and commerce, folding in Tongyi, MaaS, and multiple application teams into a single AI “big tent.”

This is organizational warfare — Alibaba’s classic approach to group‑level strategy — intended to remove silos and speed model→product deployment. But structure alone won’t win. Engineering execution, platform cost-efficiency and business model fit remain the real tests.

Strategic context and the hard part ahead

Why does this matter beyond China’s tech press? The moves come amid an intensifying global technology rivalry. Export controls on advanced chips and mounting Western scrutiny of Chinese AI players mean local cloud and model stacks are strategic assets as much as commercial ones. Alibaba’s reorg is as much about internal alignment as it is about positioning in a geopolitically constrained supply chain: who controls compute, who owns the models, who monetizes Token calls.

Can Alibaba convert an e‑commerce giant into an AI platform company that sells Token at scale? That is the wager. ATH and the Group Technology Committee may provide faster decision‑making and clearer accountability. But in a market racing toward agent‑driven interfaces and hungry for cheap inference, execution speed, infrastructure economics and cross‑group coordination will decide whether Alibaba’s Token gamble pays off.

AIE-Commerce
View original source →