Menu Close

Alibaba plans 5–10T-parameter Qwen models, unveils Zhenwu V900 chip

Alibaba Qwen T-Head roadmap beside a Zhenwu V900 chip with chips for Qwen 4 in training, 5–10T params, Q1 2027 mass production, and over 20 GW by 2032.

Alibaba CEO Eddie Wu said the company plans to train a 5–10 trillion-parameter AI model as Qwen 4 trains and 4.5/5 roadmaps scale up, unveiled Zhenwu V900 as China’s most powerful AI chip (3× prior, mass production Q1 2027), and targeted more than 20 GW of global data-center capacity by 2032; shares jumped about 5%.

Alibaba Group plans to train a new AI model with 5 trillion to 10 trillion parameters, Chief Executive Eddie Wu said on Tuesday, as the company laid out a push across models, chips, and data centers at Alibaba Cloud’s Apsara Conference in Hangzhou, Reuters reported from Beijing.

In a company statement carried with the wire, Alibaba said next-generation Qwen 4 is already in training, with upcoming Qwen 4.5 and Qwen 5 series projected to scale to 5–10 trillion parameters. Current flagship Qwen 3.8 Max has 2.4 trillion parameters, so the planned models would be roughly two to four times larger. Wu said Alibaba’s proprietary M890 AI supernode already handles inference for models above two trillion parameters, a capability he said only “a handful of companies globally” possess. The Qwen team has made “meaningful progress” on recursive self-improvement, he said.

Alibaba also unveiled Zhenwu V900, a next-generation AI chip from its T-Head unit that Wu called the most powerful AI chip in China. It delivers three times the performance of its predecessor, the M890 (released in May), and a single cluster can support up to 500,000 cards for frontier training and inference, he said. Mass production and commercial release are scheduled for the first quarter of 2027. Existing Zhenwu chips already serve more than 650 customers across industries, company materials said. The launch lands as Chinese firms race to build domestic alternatives to Nvidia processors amid U.S. export restrictions.

Wu set a target for Alibaba Cloud’s global data-center capacity to surpass 20 gigawatts by 2032. He said AI demand remains “exceptionally robust” but that supply-chain shortages are limiting how fast the company can expand. Hong Kong-listed shares jumped about 5% on the announcements, per Reuters market coverage of the same day.

This brief covers Alibaba’s Tuesday model, chip, and capacity roadmap as reported. It does not treat 5–10T parameter targets or the 20 GW goal as completed milestones.

Sources

0 0 votes
Article Rating
Subscribe
Notify of
0 Comments
Inline Feedbacks
View all comments
0
Would love your thoughts, please comment.x
()
x