An Elon Musk company’s artificial intelligence chatbot that called itself “MechaHitler” last week will soon be helping the military as the Pentagon pens deals totaling $800 million with four U.S. tech firms to expand AI’s role in national defense.
Musk’s xAI — whose Grok chatbot came under fire for antisemitic comments — received contracts worth up to $200 million to help the Pentagon incorporate artificial intelligence into its operations. Three other AI companies also received such deals. They are Anthropic, Google and OpenAI,
The feds have listed the AI companies in the General Service’s Administration database to allow all federal agencies to invest in the technology.
The moves come as SpaceX, another Musk-led company, committed $2 billion to xAI as part of a $5 billion equity round.
READ MORE: Musk plans America Party to challenge Trump, two-party system
The deals speak to the government’s view of AI’s importance to national security.
“The adoption of AI is transforming the Department’s ability to support our warfighters and maintain strategic advantage over our adversaries,” Doug Matty, DOD’s chief digital and AI officer, said in a statement.
Last week Palo Alto, Calif.,-based xAI released Grok 4, the latest version of its AI chat interface. The rollout quickly drew criticism as the system posted the “MechaHitler” comment along with other antisemitic and hateful messages.
The company quickly apologized, blaming the behavior on a code update that made Grok overly reliant on X user posts, including those with extremist views.
They’re the latest problems for the chatbot Musk is constantly hyping. In May, the previous version of Grok came under fire for bringing up far-right conspiracy theories including white genocide in South Africa — Musk’s native country.
READ MORE: SpaceX is building a massive South Texas facility to help Musk colonize Mars
Despite the glitches, the DOD said the investment will “bring in the best U.S.-based frontier AI talent to help apply cutting-edge AI to solve DoD use cases.”
Matty’s office has provided access to the generative AI models to combatant commands, the secretary of defense and the military’s joint staff. It’s working to integrate AI into its workflows and data systems.
xAI announced its “Grok for Government” service Monday. The program will provide its AI tools, including Grok 4, Deep Search, Tool Use and other capabilities for “federal, local, state and national security customers,” the company said.
It has built what it says is the world’s biggest supercomputer, known as Colossus, to support its massive computing demands. That site in Memphis has 200,000 graphics processing units. It also has a presence in London.
The other firms nabbing contracts include San Francisco-based Anthropic, which operates a chatbot called Claude; Bay Area-based Google, which has an AI model called Gemini; and OpenAI, which has ChatGPT and also operates out of San Francisco.
