Anthropic, developer of the Claude AI model, will immediately stop selling artificial intelligence services to organizations majority-owned by Chinese entities, the Financial Times reported on Thursday, citing an unnamed Anthropic executive as its source.
The restriction applies not only to companies directly headquartered in China but also to those operating under foreign subsidiaries, including those in Singapore.
The policy, which takes effect immediately, also applies to countries that have been hostile to the US, including Russia, Iran and North Korea.
Anthropic’s decision represents the first instance of a major US-based AI company denying access to its technology over national security concerns.
The media outlet noted the move doesn’t just block direct customers but also applies to groups accessing Anthropic’s technology indirectly via third-party cloud providers like AWS or Google Cloud.
The unnamed executive said the impact on Anthropic’s global revenues would be in the “low hundreds of millions of dollars”.
Earlier this year, Chinese AI start-up DeepSeek sparked global concerns by releasing its open-source R1 model, reportedly trained using unauthorized access to OpenAI systems.
OpenAI has accused DeepSeek of illicitly using its models, though DeepSeek has not publicly responded.
Anthropic, which was founded in 2021 by former OpenAI employees, recently announced $13 billion in new funding, which would value the company at $170 billion.