Alibaba unveils Zhenwu V900 chip and plans 5–10 trillion-parameter AI model

# What Alibaba announced

Alibaba used its annual Apsara conference in Hangzhou to lay out three linked moves: larger AI models, a new in-house AI chip, and expanded cloud capacity. CEO Eddie Wu framed the company's work as an effort to build every layer needed to train and deploy more powerful AI systems.

# New model scale: 5 trillion to 10 trillion parameters

# New chip: Zhenwu V900

Alibaba's semiconductor unit T-Head introduced the Zhenwu V900 AI accelerator. Alibaba claims the V900 delivers three times the performance of its predecessor, the M890, and can be linked in clusters of up to 500,000 chips to train and serve very large models. The V900 is slated for mass production and commercial release in the first quarter of 2027. The company expects "significant growth" in annual AI chip shipments driven by the new processor.

# Cloud expansion and capacity target

Alibaba Cloud set a target of exceeding 20 gigawatts of global data centre capacity by 2032. CEO Eddie Wu said customer demand for AI is "exceptionally robust," accelerating cloud revenue growth but constrained by supply chain limits that slow capacity expansion. Alibaba also said it will begin bringing AI supernodes online at commercial scale this quarter.

# Market reaction and positioning

Shares of Alibaba's Hong Kong-listed stock rose about 5% to a one-month high on the announcements. Alibaba positioned the V900 and larger Qwen models as parts of a domestic stack to compete with foreign AI processor suppliers amid export controls and the industry push to develop domestic alternatives.

# Practical implications

• For enterprises using Alibaba Cloud: expect higher-capacity AI training and inference options over the next 12–18 months as the V900 enters production and larger Qwen models mature.
• For customers evaluating cloud suppliers: Alibaba's announcements emphasize integrated vertical control (models, chips, and data centres), which could influence cost and latency for large-scale AI workloads.
• For investors and supply-chain stakeholders: growth in chip shipments and increased capital spending on data centres could affect hardware suppliers and logistics partners.

# Bottom line

Alibaba announced concrete hardware and model-scaling plans: an in-house accelerator it says triples prior performance and a roadmap to models four times larger than its flagship. The company ties these moves to commercial cloud demand and a plan to expand data centre capacity through 2032.