“打破行业惯例，DeepSeek没让英伟达和AMD先测试优化”

Core Insights - DeepSeek, a Chinese AI startup, is making headlines by not providing early access to its next-generation model V4 to major US chip manufacturers like Nvidia and AMD, instead opting to collaborate with Chinese chipmaker Huawei for software optimization [1][3] - The move is speculated to be part of a broader strategy by the Chinese government to position US hardware at a disadvantage in the competitive landscape [1] - DeepSeek's model has gained significant traction, with over 75 million downloads on the Hugging Face platform, indicating a strong competitive presence against US AI labs [3] Group 1: DeepSeek's Strategy and Market Position - DeepSeek's decision to prioritize Chinese chip manufacturers over US counterparts marks a significant shift in industry norms, potentially impacting future collaborations [1] - The company plans to launch its flagship model around the Lunar New Year, with several other Chinese AI firms also set to release new models, maintaining a robust growth trajectory [3] - Analysts suggest that while DeepSeek's actions may not directly affect Nvidia and AMD's business, the rapid rise of Chinese open-source models is increasing concerns in Washington regarding the export of advanced AI chips to China [3] Group 2: US Government Response and Export Controls - The US Department of Commerce is actively monitoring potential violations related to DeepSeek's use of Nvidia's advanced AI chips, with officials stating that enforcement will be consistent regardless of the violator [5] - Despite a previous announcement allowing Nvidia to export its H200 chips to China, the actual approval process remains stalled due to ongoing national security reviews [6] - As of now, no H200 chips have been approved for sale to China, highlighting the complexities and uncertainties surrounding US-China tech relations [6][7]