HANGZHOU — In a major escalation of the global AI race, Chinese technology giant Alibaba Group has announced plans to develop and train a massive new artificial intelligence model featuring between 5 trillion and 10 trillion parameters.
The announcement was made by Alibaba Chief Executive Eddie Wu during his keynote address at Alibaba Cloud's annual Apsara Conference. The planned model will scale well beyond the company's previous flagship Qwen 3.8 Max, which relies on 2.4 trillion parameters, making the upcoming system roughly two to four times larger. According to leadership, the heavy-duty model is designed to tackle advanced, long-horizon tasks and pave the path toward artificial superintelligence (ASI).
Alongside the sweeping model roadmap, Alibaba’s T-Head semiconductor unit introduced the Zhenwu V900, billed as China's most powerful custom AI processor. The new chip delivers triple the performance of its predecessor, the M890, and is engineered to support massive clusters of up to 500,000 cards for frontier model training and inference.
The twin announcements underline how major Chinese tech firms are aggressively building out domestic, full-stack ecosystems spanning proprietary models, hardware accelerators, and cloud data centers to bypass ongoing Western restrictions on advanced semiconductor imports. Alibaba also set a long-term infrastructure goal to expand its global cloud data center capacity past 20 gigawatts by 2032.
.png)
The opinions posted here do not belong to 🔰www.indiansdaily.com. The author is solely responsible for the opinions.
As per the IT policy of the Central Government, insults against an individual, community, religion or country, defamatory and inflammatory remarks, obscene and vulgar language are punishable offenses. Legal action will be taken for such expressions of opinion.