AI Policy Wiki
Dashboard
⚲

Eddie Wu

medium confidence · updated 2026-09-23

Eddie Wu Yongming, co-founder and chief executive officer of Alibaba Group since September 2023 and head of its cloud business. Has set out Alibaba's stated goal of artificial superintelligence at successive Apsara Conferences, including a three-stage path to ASI (2025) and, in September 2026, a 20 GW data-centre target for 2032 and a Qwen roadmap toward 5–10 trillion parameters.

Eddie Wu Yongming is the chief executive officer of Alibaba Group Holding and one of the company's co-founders. He has served as CEO and a director since September 2023 and also leads the group's cloud business, the division that develops the Qwen models covered at Alibaba / Qwen Team (Source: home.alibabagroup.com; en.wikipedia.org).

Background

Wu graduated from the College of Information Engineering of Zhejiang University of Technology in June 1996 and was Alibaba's technology director at the company's inception in 1999; he is a member of the Alibaba Partnership (Source: home.alibabagroup.com). Business Insider, reporting his appointment in June 2023, described him as Alibaba's first programmer and said he was then chairman of Taobao and Tmall Group (Source: businessinsider.com). He succeeded Daniel Zhang as CEO on September 10, 2023, when it was also announced that he would lead the cloud business (Source: en.wikipedia.org). KrASIA identified him in September 2025 as CEO of Alibaba Group and chairman and CEO of Alibaba Cloud Intelligence Group (Source: kr-asia.com).

Role in Alibaba's AI strategy

Apsara Conference 2025

In his keynote at the Apsara Conference opening on September 24, 2025 in Hangzhou, Wu set out artificial superintelligence as Alibaba's goal and described the path to it in three stages: "emergent intelligence," in which AI develops reasoning by learning from humanity's collective knowledge; "autonomous action," in which AI acquires tool use and programming abilities to assist humans, which KrASIA reported as roughly where the industry stood; and "self-iteration," in which AI connects to raw data from the physical world, learns autonomously and "ultimately surpasses humans." He predicted that global AI investment would exceed $4 trillion over the following five years. Alibaba had announced a RMB 380 billion three-year investment plan in AI infrastructure in February 2025 (Source: kr-asia.com). See Superintelligence.

Model development task force

In March 2026 Alibaba formed an internal task force headed by Wu to direct more resources to foundation-model development; the South China Morning Post reported that in an accompanying letter Wu affirmed Alibaba would "uphold our open-source model strategy" and scale up AI investment (Source: scmp.com).

Apsara Conference 2026

At the Apsara Conference on September 22, 2026, Wu named "AI models, AI chips and the AI cloud" as the three cornerstones of what he called the machine intelligence era, and said that "machines are becoming the primary force behind thinking, turning intelligence into a commodity supplied at scale." Alibaba set a target of more than 20 GW of Alibaba Cloud data-centre capacity worldwide by 2032. Wu said Alibaba expected a "significant" increase in annual AI chip shipments without giving a figure, and that it would "commit fully to AI infrastructure development" with partners (Source: amp.scmp.com). Alibaba's own post on the keynote quoted him saying that machine thinking represents under 3% of human capacity and that scaling toward 1,000 times that level "opens an unprecedented runway for exponential growth" (Source: x.com).

The same conference introduced the T-Head Zhenwu V900 chip, which Alibaba said delivers three times the performance of the Zhenwu M890, carries 216 GB of high-bandwidth memory and 1.2 TB/s of inter-chip bandwidth, can be clustered up to 500,000 chips, and is scheduled for mass production in the first quarter of 2027. Qwen project lead Liu Dayiheng presented a roadmap for the Qwen 4, Qwen 4.5 and Qwen 5 series; the SCMP reported plans to train Qwen 5 at 5 to 10 trillion parameters, against 2.4 trillion for the then-current Qwen 3.8-Max, while Techwire Asia reported Wu as saying Qwen 4 was in training and that the Qwen 4.5 and Qwen 5 series were projected to reach that range (Source: amp.scmp.com; techwireasia.com). The SCMP attributed the reported progress on recursive self-improvement to Liu's presentation (Recursive Self-Improvement (RSI)). Alibaba owns the South China Morning Post.

Reuters attributes the recursive self-improvement remark to Wu himself, reporting that he said the Qwen team had made "meaningful progress" on it, and gives the planned model size as 5 trillion to 10 trillion parameters, two to four times Qwen 3.8 Max (Source: reuters.com). Techwire Asia adds that the V900 natively supports FP8 and FP4 and that the M890 it succeeds was released in May 2026 (Source: techwireasia.com).

Congressional letter on pacing

Wu and chairman Joe Tsai were the addressees of a letter dated September 16, 2026 from Rep. Ro Khanna, ranking member of the House Select Committee on the Chinese Communist Party, urging "China's leading artificial intelligence (AI) companies to enter into a binding, enforceable, and verifiable agreement alongside America's leading artificial intelligence (AI) labs to pace the frontier" and asking four questions with a September 22, 2026 response deadline (Khanna letter to Alibaba on pacing the frontier (September 16, 2026)). No response from Alibaba is recorded in the sources cited here.

Relationships