DeepSeek Unveils V4 LLM on Huawei Amid AI Theft Claims

DeepSeek, a startup from Hangzhou, has released its latest large language model (LLM) V4, which boasts 1.6 trillion parameters and a 1 million token context window. This model represents a significant technical shift as it is optimized for Huawei’s Ascend AI processors, marking a departure from reliance on Nvidia hardware. Concurrently, U.S. State Department issued a diplomatic warning about alleged intellectual property theft involving DeepSeek and other Chinese firms, intensifying scrutiny on the company’s methods of developing AI technologies.

The implications of this launch are multifaceted. By utilizing Huawei's chips, DeepSeek not only circumvents existing export controls on Nvidia's products but may also contribute to China's AI sovereignty by strengthening domestic hardware capabilities. This scenario highlights a growing tension in U.S.-China relations concerning AI technology and intellectual property. The diplomatic cable reinforcing warnings about potential IP theft indicates a proactive U.S. stance on protecting its technological advancements while underscoring the competitive landscape among major AI players.