According to former OpenAI researcher Will DePue on November 13, six AI models from DeepSeek, Moonshot AI, MiniMax, and Ant Group contain hardware fingerprints matching Nvidia Blackwell chips in their public weights. The models are DeepSeek-V4-Flash, DeepSeek-V4-Pro, Kimi K3, Kimi K3.1-Pro, MiniMax-M3, and Ling-3.1-flash. DePue's research identified GPU architecture and random seed patterns that match Blackwell systems, validated through testing with physical B300 chips.

Estimates suggest Kimi K3 and DeepSeek-V4-Pro each used approximately 2,000 to 3,000 Blackwell GPUs, with total estimated cluster costs of $150 million to $200 million per model. However, the research cannot determine whether chips entered China or if companies rented computing power from overseas data centers.

View Source

Disclaimer: This content is for informational purposes only and does not constitute investment advice or encourage the buying or selling of crypto assets. It is intended to provide information about industry developments and is not directly related to the services offered by Gate TR. Although information obtained from third-party sources has been carefully reviewed, Gate TR assumes no responsibility for any errors or omissions. Crypto assets involve high risks and may experience significant price fluctuations. Before making any investment decisions, you should assess your financial situation and risk tolerance, conduct your own research, and make decisions independently.