Musk's xAI and Pentagon reach deal to use Grok in classified systems
Add Axios as your preferred source to
see more of our stories on Google.
Elon Musk with President Trump at a SpaceX launch in 2024. Photo: Brandon Bell/Getty Images)
Elon Musk's artificial intelligence company xAI has signed an agreement to allow the military to use its model, Grok, in classified systems, a Defense official confirmed to Axios.
Why it matters: Up to now, Anthropic's Claude has been the only model available in the systems on which the military's most sensitive intelligence work, weapons development and battlefield operations take place. But the Pentagon is threatening Anthropic in a dispute over safeguards and may soon need a replacement.
- Anthropic has refused the Pentagon's demand that they make Claude available for "all lawful purposes," insisting in particular on blocking its use for the mass surveillance of Americans and the development of fully autonomous weapons.
- xAI agreed to that "all lawful use" standard, as Axios previously reported. The New York Times first reported that a deal had been signed. xAI did not respond to requests for comment.
- It's not clear whether xAI will be able to fully replace Anthropic, or how long that process would take. Claude was used in the Maduro raid, for example, through Anthropic's partnership with Palantir.
Driving the news: Defense Secretary Pete Hegseth will host Anthropic CEO Dario Amodei for what sources expect to be a tense meeting at the Pentagon on Tuesday.
- A Defense official said Hegseth would effectively be presenting Amodei with an ultimatum. The Pentagon is threatening to brand Anthropic a "supply chain risk," among other potential penalties, if it won't agree to lift all safeguards.
- Defense officials admit that offloading and replacing Claude would be a very difficult process.
State of play: Grok, Google's Gemini and OpenAI's ChatGPT are all available in the military's unclassified systems, and Google and OpenAI have also been in talks to move over into the classified space.
- The Pentagon has moved to speed up those negotiations as it prepares to potentially sever its relationship with Anthropic.
- One source said the Pentagon had "reached out to OpenAI to reignite talks with a new sense of urgency," though the parties were still "not close to getting a deal" due to the complex issues at play. A second source confirmed the Pentagon's outreach had been accelerating.
The intrigue: The NYT reported that Google was "close" to a deal to allow classified use of Gemini, while OpenAI was "not close."
- The Defense official disputed the characterization of Google as much closer than OpenAI to a deal, saying talks were ongoing with both and the department believes both will sign agreements.
- However, administration officials insist both Google and OpenAI will have to agree to the "all lawful purposes" criteria.
- One of the sources said it was not clear if OpenAI would agree to that standard, but "we're talking."
Go deeper: Pentagon-Anthropic battle pushes other AI labs into major dilemma
