Even the ugliest feuds with President Donald Trump seem to end with “the Art of the Deal.”
The U.S. government just inked a deal to put Elon Musk’s Grok AI inside federal agencies for 42 cents per agency — a bargain that the government called “unique” and could reset Musk’s rocky relationship with Trump and scramble the fight over which models dominate Washington.
It’s the latest in a string of deals that the U.S. General Services Administration (GSA), the agency responsible for technology procurement, has made with the top AI companies – Alphabet’s Google; the ChatGPT maker, OpenAI; and Anthropic — as part of its new initiative, the OneGov agreement. Each of these deals are short-term – to prevent one model dominating, the GSA said – but Grok’s is the longest, with an 18- month contract. On Sept. 22, the GSA announced that it would be working with Meta to get free access to its Llama models, while OpenAI and Anthropic agreed to provide their models for $1, and Google charged 47 cents.
Musk, according to the Wall Street Journal, picked 42 cents as a reference to the Hitchhiker’s Guide to the Galaxy, a science fiction novel.
It’s hard to calculate how much money Musk is saving the government by offering the model at only 42 cents a pop. xAI’s Grok 4 Fast is priced per output, and generally agencies might be on the hook for hefty API licencing fees.
“We really like the notion of having strong competition and market tension between these models and these companies,” Federal Acquisition Service Commissioner Josh Gruenbaum told the Wall Street Journal in an interview. “When someone goes and updates their model with a cool feature, that only encourages the others to go do the same thing.”
The deal could be a sign that the turbulent Musk-Trump relationship is in a period of thawing. After breaking with Trump in June over tariffs and spending — even calling for his impeachment — Musk has become one of the president’s most vocal critics. However, on Sunday, the two were spotted side by side at Charlie Kirk’s memorial in Arizona, shaking hands and chatting for the first time since their public split.
Now Musk is praising Trump’s leadership in official press releases, saying xAI looks forward to “rapidly deploying AI throughout the government.”
Whether this is a fragile truce or a genuine thaw, the timing is striking: Musk is still struggling to keep pace with rivals like OpenAI and Anthropic in the private market, but inside Washington, he just secured a coveted stamp of approval.
“MechaHitler” in the government?
Musk’s deal with Washington comes on the heels of embarrassing stumbles for Grok itself.
The chatbot has been caught spouting antisemitic comments, at one point dubbing itself “MechaHitler,” and even hurling slurs at Poland’s prime minister. xAI scrubbed the posts and promised tighter safeguards, framing the missteps as part of the messy process of training frontier AI.
“We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts,” the company said at the time, adding that its vast user base helps flag blind spots so the model can be retrained quickly.
More than 30 advocacy groups urged the Office of Management and Budget to keep the model out of federal systems, and several Democratic lawmakers pressed the GSA on its decision, according to news site Fedscoop. A GSA spokesperson has stressed the agency is weighing all vendors “equally” and that no single deal amounts to a final endorsement.
