IT之家 6 月 27 日消息，据外媒 The Information 报道，由于美国当局策略致英伟达 H20 芯片在中国内地供应短缺，继而导致 DeepSeek-R2 AI 模型开发工作延迟。
IT之家注意到，DeepSeek 使用了 5 万块 Hopper GPU 训练 R1 模型，其中包括 3 万块 H20、1 万块 H800 和 1 万块 H100，这些显卡由其投资方高毅资产（High-Flyer Capital Management）采购。
外媒援引两位知情人士透露，DeepSeek 团队近几个月一直在密集推进 R2 模型开发，但 CEO 梁文锋目前对模型的能力仍不满意，内部仍在继续提升性能，相应模型尚未准备好正式投用。
此外，相应报道同时指出，多数用户都是在英伟达的 H20 芯片上部署 / 运行现有 DeepSeek-R1 模型，随着 H20 芯片遭美国当局限制出口，除了令 DeepSeek R2 的研发更加困难外，也对现有 DeepSeek-R1 实际部署造成一定影响。
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.