Pentagon begins replacing Anthropic, OpenAI AI services ready

The US Department of Defense has initiated initial steps to replace AI services from Anthropic in its processes, such as current military operations in Iran, with alternatives. This is a consequence of the escalation in the dispute over the use of AI technology for the US military. According to high-ranking Pentagon officials, the replacement of Anthropic products like Claude will take more than a month, but work is currently underway to integrate other large language models. This likely includes AI services from OpenAI, which are now also being sold to US government agencies through Amazon's cloud offerings.

In an interview, however, Cameron Stanley, head of the Digital and AI department at the Pentagon, did not mention ChatGPT as an alternative to Claude. “The department is actively seeking to integrate multiple LLMs in the relevant government agencies,” he stated merely regarding large language models (LLMs) for the US Department of Defense. “Development work on these LLMs has begun, and we expect them to be ready for deployment very soon.”

Pentagon and Anthropic at odds

Although Anthropic was promised hundreds of millions of US dollars by the Pentagon last summer, like other AI companies, the use of AI technology became a point of contention early this year. Anthropic refused to allow the Department of Defense to use the AI models for mass surveillance of the US population. Furthermore, weapons that fire without human intervention were not to be developed with them. Following this, the Pentagon threatened Anthropic with far-reaching sanctions. Ultimately, the Pentagon blacklisted Anthropic and classified the AI company as a security risk. Anthropic is legally challenging this measure and suing the US government.

This appears to have severed ties between the Trump administration and Anthropic, and the multi-million dollar contract for the AI company is now severely jeopardized. According to Bloomberg, Defense Secretary Pete Hegseth and US President Donald Trump have already set a six-month timeframe for the US military and other federal agencies to switch to solutions from other AI companies. Until now, the Anthropic system was the only AI service permitted for use with confidential information and processes at the Pentagon, and Claude is very popular among Department of Defense employees due to its user-friendliness.

ChatGPT as a Claude replacement?

These employees will have to adapt and get used to ChatGPT, because after the break with Anthropic, the Pentagon has entered into an AI deal with OpenAI. This happened at the end of February, and now Reuters reports that OpenAI's AI services are already available to US government agencies via Amazon Cloud. This includes the US Department of Defense for both confidential and non-classified work. The ChatGPT developers expect contracts with larger enterprise customers from this, as such government contracts would demonstrate trust and reliability.

While it cannot be ruled out that Anthropic will be commissioned by the US government again, it is unlikely to be by the US Department of Defense in the near future. Shortly after Anthropic filed the lawsuit against the US government, Emil Michael, Under Secretary of Defense for Research and Engineering, stated that the US military is concluding this matter and he sees little chance for a resumption of talks. “I don't think there's a scenario where this gets resolved in this manner,” Michael said.

(fds)