Alibaba Group Holding Ltd. introduced its Zhenwu V900 AI accelerator, which the company claims is China's most powerful AI chip, tripling the performance of its predecessor. This launch is part of a broader strategy to underpin a massive expansion of data center capacity, aiming for over 20 gigawatts globally by 2032. The move positions Alibaba to compete with leading AI chip makers like Nvidia Corp. and is driven by the company's commitment to substantial AI investments, with CEO Eddie Wu emphasizing the vast growth potential of machine thinking.
The announcement led to a roughly 3% jump in Alibaba's shares in Hong Kong. The Zhenwu V900, developed by Alibaba's T-Head division, can be clustered in up to 500,000 units for frontier-model training. Alibaba also revealed plans for its next-generation Qwen 4 model, currently in training, and future Qwen 4.5 and Qwen 5 model series. The Zhenwu V900 is scheduled for mass production and commercial release in the first quarter of 2027, with existing Zhenwu chips already serving over 650 customers across various industries.
Simultaneously, Meta's consumer AI agent, Muse, rapidly climbed to the top of mobile-app charts, contributing to a rally in global AI-related equities and reinforcing investor optimism about AI adoption and monetization. This broader tech rally indicates a positive market reaction to advancements in AI infrastructure and consumer-facing AI applications. Alibaba's push into AI infrastructure is seen as a strategic move to secure its cloud offering and potentially reduce inference costs, while Meta's AI traction is viewed positively if it enhances engagement and monetization without disproportionately increasing inference expenses.
Alibaba's new chip and data center expansion align with increasing AI infrastructure investments by other tech giants. Nvidia recently outlined plans for up to 2 gigawatts of AI infrastructure in Australia by 2027, and Meta announced a $9 billion, 1-gigawatt data center in Canada. Alibaba's domestic competitor, Huawei, also unveiled new AI infrastructure capable of scaling to one million processors. These developments highlight the intensifying global race for AI computing power and infrastructure.
