Wu Yongming Appointed to Lead Wukong (悟空) as Alibaba (阿里巴巴) Stakes AI Future on B‑to‑B Breakout

A make‑or‑break move

Alibaba has appointed CEO Wu Yongming (吴泳铭) to personally oversee a new Wukong (悟空) business unit inside the freshly created Alibaba Token Hub (ATH). The move signals urgency. After months of heavy AI spending and few reliable revenue engines, Alibaba is shifting from “sell models” to “sell workflows” — and it has been reported that the company views Wukong as the linchpin for that pivot.

From model abundance to scenario scarcity

The architecture change bundles Tongyi Lab, MaaS, Qianwen (千问) and the new Wukong unit under ATH to force cross‑BU co‑ordination and speed commercialization. It has been reported that China’s daily token consumption surged dramatically over 2024–25 — from roughly 100 billion tokens to a marketwide daily peak in the tens of trillions — exposing a paradox: massive compute use but weak monetization. Alibaba’s dual pressure is acute: monetize cloud compute while using AI to revive e‑commerce efficiency, all amid rising capital expenditure on chips and infrastructure.

Wukong’s bet: enterprise agents inside DingTalk

Wukong is pitched as an “enterprise‑grade AI native work platform” focused on agent‑style autonomous execution and deep integration with DingTalk (钉钉). Unlike consumer‑facing chat demos, the product claims to prioritize enterprise needs — permission inheritance, sandboxed execution and measurable token costing — and to convert natural‑language commands into automated backend API calls. It has been reported that Alibaba expects Wukong to lift per‑enterprise token consumption several‑fold by tapping DingTalk’s 200M+ users and internal skills across Taobao, 1688 and Alibaba Cloud.

Real tests ahead

Putting Wu in direct charge and consolidating AI assets is a shock‑therapy response to organizational inertia. But consolidation carries risks: internal incentives, cost allocation and the hard work of convincing firms to pay for new workflows. Geopolitical headwinds — export controls and chip supply dynamics that raise compute costs — add another layer of pressure. Can Wukong turn token volume into sustainable revenue? Investors and competitors will be watching three hard metrics: DingTalk activation and token growth, B‑enterprise paid conversion, and the emergence of a clear single‑user “One Person Team” case that proves AI actually replaces weeks of work in a single afternoon.