Nvidia Denies Plans for China-Specific AI Chip amid Export Restrictions
Crhk · 3 SOURCESabout 1 hour ago2 MIN
Summary
Nvidia has officially denied plans to supply a special AI chip to Chinese customers, contradicting a report by The Information that claimed the company was preparing to launch a China-specific product. The US chipmaker confirmed on June 20 that it has no LPU product currently being sold in China and no such chip in its development pipeline for the Chinese market.
Key Points
- The Information cited two Nvidia employees claiming the company planned limited shipments of a China-designed AI chip by year-end
- The chip, called a Language Processing Unit (LPU), would work alongside Graphics Processing Units (GPUs) to accelerate AI chatbot response times
- Specifications were reportedly modified to comply with US export controls, though it remains unclear whether Beijing has approved the sales
- Reuters reported in March that Nvidia was preparing compliant AI chips to remain competitive in China's AI market
- Nvidia CEO Jensen Huang stated in May that the company had "largely conceded" the China AI chip market to Huawei
Why It Matters
The denial highlights the ongoing tensions between US technology restrictions and China's ambitions in artificial intelligence development. For Hong Kong investors and tech professionals, Nvidia's retreat from the Chinese AI chip market signals a significant shift in the global semiconductor landscape, potentially opening opportunities for domestic champions like Huawei while affecting supply chains for AI-related businesses in the region.
The denial highlights the ongoing tensions between US technology restrictions and China's ambitions in artificial intelligence development. For Hong Kong investors and tech professionals, Nvidia's retreat from the Chinese AI chip market signals a significant shift in the global semiconductor landscape, potentially opening opportunities for domestic champions like Huawei while affecting supply chains for AI-related businesses in the region.