In short
- The AI company Xai of Elon Musk has concluded a Pentagon contract of $ 200 million, only a few days after the chatbot groked anti -Semitic and homophobic reactions.
- The deal is intended to offer AI tools for defense and use of companies, despite concern about the reliability of grok in sensitive environments.
- Experts say that the contract reflects a strategic push to stay close to Top AI developers and to retain a lead over global rivals such as China.
Even a “mechhitler” melting was not enough to scare the Pentagon, apparently.
Just a few days after Xai’s Grok Chatbot generated a stream of anti -Semitic and homophobic reactions, the company of Elon Musk announced that it concluded a defense contract of $ 200 million to develop and deliver AI tools to federal employees.
The Chief Digital and Artificial Intelligence Office of the Ministry of Defense announced the contract on Monday and said it was designed to help the government use the latest AI tools for everything, from battlefield activities to business needs.
“The acceptance of AI is to transform the wealth’s ability to support our war fighters and to maintain a strategic benefit from our opponents,” said CDAO Chief Digital and AI officer Dr. Doug Matty in a statement. “The use of commercially available solutions in an integrated capacity approach will speed up the use of advanced AI as part of our joint mission -less tasks in our war fighting domain, as well as intelligence, companies and company information systems.”
According to Lucas Hansen, the co-founder of the AI Security and Education Non-Profit Civai, the deal with Xai is less about immediately use and more about retaining equal pace with worldwide rivals.
“The government wants to support the Frontier Labs and the development of AI, so that the US does not fall behind China,” Hansen said Decrypt. “Then there is the question of how those funds are distributed and how that support is offered. I think this is perhaps the specific way they have chosen to support the companies, instead of an expectation of internal implementation immediately.”
But Hansen warned that the underlying behavior of grock deeper worries expressed how it could work in government institutions with high commitment.
“Even if Grok is used by the government with the same version that previously showed the willingness to generate harmful reactions than – even if those use cases are not explicitly anti -Semitic or hateful – it is fundamentally more likely to behave unethically and autonomously,” he said.
Grok has been the center of various public meltdowns. In May, reports appeared that Grok has placed ‘white genocide’ claims in non -related instructions. Xai blamed the issue to a rogue employee. Last week, just before the launch of the Grok 4, the chatbot again criticized a wave of racist and homophobic output, which led to the resignation of X CEO Linda Yaccarino.
Although the worries get stuck about how Grok could behave in sensitive environments, Hansen said that the version that the Pentagon uses is probably not the same as those with which X users have had access since November 2023.
The price of $ 200 million is the newest deal between the military and top AI developers of the United States. Other AI developers who receive $ 200 million contracts are Google, Anthropic and OpenAi. In November Meta said it would make its open-source Llama AI model available for American defense agencies and contractors. In June, ChatGPT developer OpenAi received its own contract of $ 200 million with the Ministry of Defense, with the launch of its AI Suite, OpenAI for the government.
If the Pentagon grok would use a rival, Hansen emphasized that it would be because grok is “open weight”, which means that it can be performed on private hardware and not require API connection back to Xai.
Grok’s announcement for the government – a series of products that make our border models available to customers of the United States
We are especially enthusiastic about two new partnerships for our US government partners
1) A new contract from the US Department of Defense
2) Our …– Xai (@xai) July 14, 2025
Despite the recent controversies of Grok, in the heart of the deal, Hansen said, a simple strategic calculation can be: stay close to the most important players shaping the future of AI.
“I think most of these are the government that wants a relationship with all Frontier Labs, and they touch the four laboratories that really matter,” he said.
In a separate explanation in which the deal was announced with the Pentagon, Xai said that his mission is to create AI tools to “help humanity in our search for understanding and knowledge.” To do this, Xai announced the launch of Grok for the government, a series of AI tools, including the recently launched Grok 4 and Deep Search.
“America is the world leader in AI, and this is not least because of a tradition of innovation and strong investments in engineering and science,” Xai wrote. “We are delighted to return to the country that made Xai unique possible here.”
Xai did not respond to requests for comments from Decodeer.
Generally intelligent Newsletter
A weekly AI trip told by Gen, a generative AI model.