The White House has announced plans to strengthen collaboration with U.S. artificial intelligence firms in response to a growing threat of mass AI technology theft by Chinese companies, as detailed in an internal memo by Michael Kratsios, Director of Science and Technology Policy. The memo indicates that foreign entities, primarily based in China, are utilizing “distillation” tactics to replicate AI advancements from American firms. Notably, this revelation aligns with earlier reports from Anthropic, which highlighted similar activities by Chinese labs, including DeepSeek, Moonshot, and MiniMax. The White House’s intent is to share threat intelligence, develop best practices, and explore accountability measures for these foreign actions, all amidst escalating tensions between the U.S. and China in the tech sector.

OpenAI: OpenAI is a leading US AI research organization developing frontier large language models and tools accessible via APIs. It has accused Chinese firm DeepSeek of using distillation tactics to copy its technology, supporting the White House’s recent concerns over systematic exploitation by foreign actors.
MiniMax: MiniMax is a Chinese AI developer advancing multimodal models for consumer applications and enterprise use. It has reported strong revenue growth this year and was accused by Anthropic of participating in coordinated distillation efforts against Claude models.
DeepSeek: DeepSeek is a Chinese AI startup specializing in efficient open-source large language models trained on domestic infrastructure. It faces US accusations from OpenAI and Anthropic of distilling their models and recently previewed its V4 series optimized for Huawei chips amid heightened scrutiny.
Moonshot: Moonshot AI is a prominent Chinese AI firm behind the Kimi chatbot, focusing on advanced language models and open-source releases. It released its flagship Kimi K2.6 model earlier this week and has been named by Anthropic in distillation attacks on US AI systems.
Anthropic: Anthropic is a US-based AI safety company known for its Claude models, emphasizing reliable and interpretable AI systems. It recently identified industrial-scale distillation campaigns by Chinese labs DeepSeek, Moonshot, and MiniMax targeting its models, informing the White House memo on foreign AI theft.
White House: The White House Office of Science and Technology Policy advises the US President on scientific and technological matters, coordinating federal efforts in innovation and policy. In a memo dated April 23, 2026, OSTP announced plans to share intelligence and develop best practices with US AI firms to combat industrial-scale distillation campaigns attributed to foreign entities primarily in China.
Michael Kratsios: Michael Kratsios is the Director of the White House Office of Science and Technology Policy and Assistant to the President for Science and Technology, focusing on advancing US leadership in emerging technologies. He authored the April 23, 2026 internal memo detailing new evidence of Chinese-led efforts to distill US AI models and outlined White House actions including enhanced coordination with industry to counter these threats.

Industry Alerts: Anthropic publicly detailed distillation campaigns by DeepSeek, Moonshot, and MiniMax, prompting broader US government warnings on AI model theft.
US Policy Response: The White House OSTP memo commits to sharing threat intelligence and best practices with American AI companies to identify and remediate distillation tactics.
Chinese AI Advances: DeepSeek previewed new V4 models running on Huawei chips just days after the White House memo, highlighting ongoing US-China tech tensions.