Key Takeaways
- Alibaba shares increased approximately 3% in Hong Kong trading following the introduction of its Zhenwu V900 AI accelerator
- The V900 chip provides three times better performance compared to the previous Zhenwu M890 model
- The company aims to reach over 20 gigawatts of worldwide data center capacity by the end of 2032
- Alibaba has pledged over $53 billion in AI investments spanning three years
- The company’s upcoming Qwen 4 model is currently undergoing training, with Qwen 4.5 and Qwen 5 in the pipeline
Alibaba shares rose approximately 3% during Tuesday’s Hong Kong trading session following the company’s presentation of an advanced AI accelerator and substantial infrastructure expansion strategy at its annual Apsara Conference held in Hangzhou.
Alibaba Group Holding Limited, BABA
The newly introduced accelerator, designated the Zhenwu V900, originates from Alibaba’s T-Head semiconductor division. It delivers triple the computational power of the earlier Zhenwu M890, which launched in May.
The V900 chip supports clustering configurations reaching 500,000 units to facilitate frontier-model training operations. Full-scale production and commercial availability are projected for the first quarter of 2027.
Eddie Wu, serving as Alibaba’s CEO, revealed objectives to achieve 20 gigawatts of worldwide data center capacity for Alibaba Cloud by 2032. He pointed to “exponentially increasing AI demand” as the primary motivation for this aggressive expansion.
Wu characterized machine intelligence as possessing an “immense growth trajectory,” drawing parallels between current AI progress and the initial phases of electrical power adoption. He noted that “AI coding represents merely the light bulb moment of the machine intelligence epoch.”
The e-commerce giant has pledged investments exceeding $53 billion across a three-year period to strengthen its AI infrastructure. Last August, the company secured approximately $10.2 billion through a secondary equity offering in Hong Kong.
The current Zhenwu accelerator portfolio serves over 650 clients spanning automotive, financial services, energy, and manufacturing industries.
AI Model Advancement Takes Center Stage
Beyond the hardware announcement, Alibaba revealed that its upcoming Qwen 4 model remains in active training. The organization also disclosed roadmaps for subsequent Qwen 4.5 and Qwen 5 model generations.
Alibaba intends to create a model featuring between 5 and 10 trillion parameters, significantly exceeding current industry standards. This advanced model would target extended and intricate computational challenges.
The organization additionally intends to pursue a separate listing for its T-Head semiconductor design subsidiary to capitalize on growing investor enthusiasm for AI accelerator technologies.
Intensifying Industry Competition
Chinese competitor Huawei introduced new AI infrastructure capabilities last week, claiming scalability to one million processors.
Nvidia disclosed plans earlier this month to enable up to 2 gigawatts of AI infrastructure throughout Australia by 2027. Meta announced in July a 1-gigawatt facility in Alberta, Canada, with an estimated price tag near $9 billion.
Alibaba’s revelations coincided with a broader rally in global AI equities, driven by encouraging early results from Meta’s newly launched personal agent technology, which boosted overall market confidence.
The Apsara Conference occurred just before a scheduled summit between US President Donald Trump and Chinese President Xi Jinping. Senior technology leaders including Nvidia’s Jensen Huang and Microsoft’s Satya Nadella plan to participate in a White House state dinner connected to the diplomatic meeting.
Artificial intelligence themes are anticipated to play a significant role in those high-level conversations.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
