🤖 AI Summary
Nvidia CEO Jensen Huang expressed disappointment after the Financial Times reported China’s Cyberspace Administration urged domestic firms, including ByteDance and Alibaba, not to use Nvidia’s RTX Pro 6000D — a variant made for the Chinese market. The report follows a fraught period of U.S. export controls on Nvidia AI chips (including the H20 server accelerator), a negotiated August deal that allowed H20 exports in exchange for the U.S. government receiving 15% of China-bound H20 sales, and a recent Chinese antitrust probe into Nvidia’s Mellanox acquisition. Huang said Nvidia will “only be in service of a market if the country wants us to be” and has instructed analysts to exclude China from financial forecasts because access is subject to geopolitical negotiations.
Technically and commercially, this is significant: restricting access to chips like the RTX Pro 6000D and H20 constrains China’s ability to run large-scale training and inference on Nvidia’s accelerators, accelerating demand for domestic GPU/accelerator alternatives and cloud-native workarounds. For Nvidia it increases geopolitical revenue risk and regulatory complexity; for the AI community it further fragments the global hardware ecosystem, complicates cross-border model deployment and supply-chain planning, and may spur localized silicon and software stacks in China.
Loading comments...
login to comment
loading comments...
no comments yet