At the Cloud Expo (Yunqi Conference), Alibaba Cloud unveiled its new-generation in-house AI chip, Zhenwu V900. CEO Wu Yongming said it is “China’s strongest AI chip,” with performance three times that of the previous-generation M890, and mass production planned for Q1 2027. The goal is even more ambitious: by 2032, Alibaba Cloud’s global data center capacity will exceed 20GW, while on the model side it plans to train new models with 50 trillion to 100 trillion parameters.
I’m more interested in Wu Yongming’s own remark—he said that in the long term, demand far exceeds supply capacity, and that shortages in the global AI data center supply chain are limiting the pace of expansion. Translate this: whether Zhenwu V900 can go head-to-head with NVIDIA will only be answered when it reaches mass production in 2027; but the 20GW target is constrained by power, racks, and advanced packaging—not by chip naming. Alibaba Cloud’s three-year AI investment commitment has already exceeded $53 billion, and in August it raised about $10.2 billion again through share placements.
$BABAB $NVDAB
Let me ask you this: is this a real substantive threat to NVIDIA, or an independent “domestic self-sufficiency” story?
I’m more interested in Wu Yongming’s own remark—he said that in the long term, demand far exceeds supply capacity, and that shortages in the global AI data center supply chain are limiting the pace of expansion. Translate this: whether Zhenwu V900 can go head-to-head with NVIDIA will only be answered when it reaches mass production in 2027; but the 20GW target is constrained by power, racks, and advanced packaging—not by chip naming. Alibaba Cloud’s three-year AI investment commitment has already exceeded $53 billion, and in August it raised about $10.2 billion again through share placements.
$BABAB $NVDAB
Let me ask you this: is this a real substantive threat to NVIDIA, or an independent “domestic self-sufficiency” story?