Nvidia denied a report claiming it was developing a China-specific AI inference chip, stating it has no LPU sales in China and no such product on its roadmap. The reported chip would have used technology licensed from Groq to accelerate AI inference alongside GPUs. Due to US export controls, Nvidia has lost significant ground in China’s AI market, which once represented at least one-fifth of its data-center revenue. The company has maintained limited access through H200 sales, with approximately 10,000 chips approved for ByteDance and Tencent each, but this remains modest and uncertain amid rising competition from Huawei and other domestic suppliers.
Source: Read the original article
