Alibaba Group Holding announced a significant technological advancement from its cloud computing division on October 18, 2025. The company introduced a new computing pooling solution called Aegaeon, which it claims resulted in an 82% reduction in the number of Nvidia graphics processing units (GPUs) required to serve its artificial intelligence models. The system was beta-tested for over three months in Alibaba Cloud's model marketplace. During this period, it reduced the number of Nvidia H20 GPUs needed to serve dozens of models from 1,192 to just 213, according to a research paper presented at the 31st Symposium on Operating Systems Principles in Seoul. This development is poised to create significant cost efficiencies and improve resource allocation in the capital-intensive AI sector. On a day with otherwise light news for the company, Alibaba's stock saw a slight increase of about 1%, with investors continuing to assess the company's recent performance momentum, which has seen the stock rise nearly 39% over the last three months.
Alibaba Cloud Unveils Aegaeon System, Slashing AI GPU Needs by 82%
BABA
Related News
BABA
Alibaba trades slightly higher amid broader market rally and Supreme Court tariff ruling
BABA
Alibaba Shares Tumble on Pentagon List Confusion and AI Margin Concerns
BABA
Alibaba Faces Rare Downgrade Ahead of Q3 Results as AI Costs Pressure Margins
BABA
Alibaba Shares Steady Pre-Earnings as Qwen3.5 AI Launch Boosts Retail Optimism
BABA