Google to Continue Offering Anthropic AI Models on Cloud Platforms, Excluding U.S. Defense Projects
Google announced that it will continue to make artificial intelligence technologies developed by Anthropic available to its cloud customers, excluding projects linked to the defense sector. The decision follows a move by the United States Department of Defense to classify the company as a potential supply chain risk in federal contracts.
Google’s stance comes after a similar step by Microsoft, which confirmed it will keep offering Anthropic’s products — including the AI models known as Claude — to its commercial clients while excluding U.S. defense entities. Meanwhile, Amazon said through its cloud division Amazon Web Services that it will continue providing Anthropic’s technologies to customers, provided they are not used in projects connected to the U.S. defense sector.
Google clarified that the U.S. government’s decision does not prevent collaboration with Anthropic on non-defense initiatives. The company emphasized that Anthropic’s products will remain available through Google Cloud.
Anthropic’s AI models, including Claude, are offered through Google’s Vertex AI platform, which enables developers and businesses to integrate advanced AI models into their applications and digital services.
Google also maintains a strong investment relationship with Anthropic, being one of the company’s key financial backers. In January 2025, Google announced an additional $1 billion investment, bringing its total funding in Anthropic to approximately $3 billion.
Anthropic also relies on Google’s AI infrastructure to train its models. The partnership expanded recently after the company gained access to up to one million Tensor Processing Units (TPUs) developed by Google to support large-scale AI training.
The government classification followed Anthropic’s refusal to accept certain usage conditions requested by the U.S. Department of Defense. As a result, the administration directed federal agencies to stop using the company’s technologies, while the Pentagon announced it would phase out its collaboration with Anthropic over a six-month transition period.
Following the decision, several U.S. defense technology firms reportedly asked employees to stop using Anthropic’s Claude models and shift to alternative AI solutions available in the market, including technologies provided by OpenAI.
Anthropic CEO Dario Amodei stated that the company plans to challenge the government’s designation in court, noting that legal action is the only option available to contest the decision.
These developments highlight the growing tensions between major artificial intelligence companies and U.S. government agencies as competition intensifies to develop and deploy advanced AI technologies — particularly as these tools become increasingly linked to military and national security sectors.