- Alibaba unveiled new AI infrastructure spanning chips, cloud computing, enterprise software and scientific research at WAIC
- The launches reflect the company’s push to build a full-stack AI ecosystem beyond foundation models
Alibaba used the opening day of the 2026 World Artificial Intelligence Conference (WAIC) to unveil a series of AI products spanning chips, cloud infrastructure, enterprise software, scientific research and AI hardware, highlighting its ambition to build a full-stack AI ecosystem.
The announcements began at the infrastructure layer. Alibaba chip unit T-Head open-sourced its self-developed T-Head SAIL AI software stack, providing the underlying software framework for the company’s XuanTie (Zhenwu) AI chips.
Supporting the hardware rollout, Alibaba Cloud launched Lingjun Zhenwu (灵骏真武) M890 supernode instances, offering public-cloud AI computing in a 64-GPU configuration capable of serving inference workloads for foundation models with tens of trillions of parameters.
The service has entered invitation-only testing.
Hardware, software stack
Alibaba said shipments of its Zhenwu AI chips had reached 560,000 units by April, serving more than 400 customers across over 20 industries.
Building on its computing infrastructure, Alibaba Cloud also introduced Meoo Team (秒悟团队版), an enterprise AI collaboration platform designed for businesses.
The platform provides centralized identity management, procurement controls and shared team assets to support AI-assisted workflows in e-commerce, content creation, marketing, finance and education.
Alibaba also expanded its AI hardware portfolio during the conference. Qwen unveiled its first AI agent earbuds, featuring offline voice interaction, real-time translation and integration with Alibaba’s broader ecosystem of AI services.
In scientific research, Alibaba’s DAMO Academy launched DAMO Lingshu (达摩灵枢), an AI platform designed to help researchers complete the full scientific workflow from hypothesis generation to experimental validation.
Advanced materials research
The platform has already been applied to superconducting materials research.
Assisted by DAMO Lingshu and the Elements foundation model for materials science, researchers screened 68,000 candidate superconductors, four of which have since been synthesized and experimentally validated.
In life sciences, DAMO Academy also introduced Lingshu-Cell, a cell world model aimed at accelerating drug discovery and virtual cell research.
Taken together, the announcements illustrate Alibaba’s evolving AI strategy. Rather than focusing solely on selling cloud computing resources, the company is assembling an integrated AI stack spanning chips, cloud infrastructure, enterprise software, scientific research platforms and intelligent agents.
Why it matters for global readers
For international observers, the product launches offer a snapshot of how Chinese technology companies are expanding beyond frontier models to build AI infrastructure across hardware, platforms and applications.
As China simultaneously advances national AI initiatives and international cooperation efforts, Alibaba’s strategy reflects a broader industry shift toward delivering complete AI systems rather than standalone models.
