Source: Reuters (via feed)
Chinese AI lab DeepSeek has withheld its latest AI model from U.S. chipmakers, including Nvidia, sources told Reuters. This decision breaks with the usual practice of sharing new models with hardware partners before a major performance update.
DeepSeek’s earlier low-cost AI model shook global markets last year. However, the company has chosen not to disclose its upcoming flagship AI model to U.S. chipmakers for optimization. The sources did not specify the reasons behind the move.
Typically, AI developers provide chipmakers with new models to help optimize processor performance. Without this collaboration, hardware manufacturers could face challenges in adapting their products for the AI workloads.
DeepSeek did not respond to requests for comment. Industry observers say the step is unusual, given the close link between AI developers and chip companies.
The move highlights ongoing tensions between U.S. and Chinese technology sectors, especially amid competition in artificial intelligence.
The withholding of AI model details limits U.S. chipmakers’ ability to optimize hardware, affecting AI market dynamics and technology development in both countries.
