Alibaba's Major AI Restructuring: Tongyi Lab Upgraded, AI Offensive Begins?

On April 8th, an all-hands email from Alibaba went viral in tech circles.

No lengthy preamble, no vague language—just several blockbuster adjustments: establishing a new Group Technology Committee, upgrading Tongyi Lab to business unit level, reshuffling three key technical executives, and appointing Fei-Fei Li as Alibaba Cloud CTO.

My first impression from this letter: Alibaba is dead serious this time.

Over the past few years, Alibaba’s AI layout always gave the impression of “having resources and technology, but missing something.” DAMO Academy conducted plenty of cutting-edge research, but productizing it was always half a beat slow. Tongyi Qianwen released several versions, but never quite matched the mindshare of Wenxin Yiyan or Doubao.

This organizational restructuring centers on one thing: elevating AI to the group’s highest priority.

Let’s start with Tongyi Lab’s upgrade to business unit. Previously, Tongyi Lab was a research institution under DAMO Academy. Now it’s directly promoted to independent business unit, led by Jingren Zhou, who also serves as Chief AI Architect of the Group Technology Committee. What does this mean? It means Tongyi Qianwen is no longer just a “research project”—it’s the core carrier of Alibaba’s AI strategy.

Resource allocation will definitely be different. Before it was “do what you can with the budget you have.” Now it’s “give whatever budget is needed to get it done.” This shift is crucial for AI products in catch-up mode.

Next, Fei-Fei Li’s appointment as Alibaba Cloud CTO. This appointment is quite interesting. Fei-Fei Li is a top-tier AI academic, founder of ImageNet, Stanford professor. Her becoming Alibaba Cloud CTO sends a clear signal: Alibaba Cloud’s technical roadmap needs to transform toward AI-native—not just adding AI as a feature module to cloud products, but rebuilding from the ground up.

I saw some interpretations saying “this is just Alibaba hyping concepts, hiring a celebrity scientist for publicity.” I disagree. Someone of Fei-Fei Li’s caliber wouldn’t join Alibaba for just a title. What she sees is Alibaba Cloud’s scenarios and data, and the opportunity to deploy academic research at massive scale. This kind of “academia + industry” combination, if well-coordinated, is enormously powerful.

Another detail easily overlooked: Taobao Flash Purchase changed leadership.

On the surface this looks like e-commerce personnel adjustment, but it actually reflects Alibaba’s emphasis on AI application scenarios. Flash Purchase is a high-frequency, real-time business scenario—perfect for testing AI’s decision-making and recommendation capabilities. Putting the strongest people in this position shows Alibaba wants to inject AI capabilities into core business, not just build external AI products.

From a competitive landscape perspective, this adjustment is also necessary.

ByteDance has Doubao, Baidu has Wenxin, Tencent has Hunyuan—everyone is All-in AI. What’s Alibaba’s advantage? Cloud computing infrastructure, rich e-commerce and payment scenarios, years of accumulated enterprise service capabilities. But these advantages need a good organizational structure to actualize, otherwise they’re dispersed and internally competitive.

This adjustment is about consolidating dispersed forces into cohesive power.

Of course, the challenges are obvious too.

Alibaba’s biggest problem is “big.” For large companies doing innovation, the biggest enemy isn’t competitors—it’s themselves: layers of approval processes, complex departmental interests, fixed mindsets. Upgrading Tongyi Lab to business unit is good, but whether it can really achieve “small steps, fast iterations” remains to be seen.

Another point: Alibaba’s AI products have consistently been weak on consumer experience. Tongyi Qianwen has strong capabilities, but interface design, interaction flow, user onboarding—always felt a bit lacking. This isn’t a technical problem, it’s a product thinking problem. Fei-Fei Li and Jingren Zhou both have technical backgrounds. How to compensate for product capability gaps is a challenge.

Overall, this organizational restructuring is an important turning point for Alibaba’s AI strategy. From “testing the waters” to “full offensive,” from “scattered exploration” to “focused breakthrough.”

Domestic large model competition is moving from “model capability comparison” into a new phase of “ecosystem and application deployment competition.” Is Alibaba’s move too late? Honestly, yes, a bit late. But a late start doesn’t mean no chance—what matters is execution going forward.

In 2026, the AI battlefield is getting increasingly intense.