The U.S. War Department has entered into an agreement with xAI to expand the capabilities of artificial intelligence on GenAI.mil, the Department’s custom-made AI platform.
According to a Dec. 22 press release from the War Department, the initiative will embed xAI’s frontier AI systems, based on the Grok family of models, directly into GenAI.mil, with an initial rollout planned for early 2026, allowing all military and civilian personnel to use xAI’s features at Impact Level 5, meaning they can safely manage Controlled Unclassified Information in their everyday tasks.
The War Department says it will continue to scale an AI ecosystem built around speed, security, and decision superiority, and it expects newly IL5-certified capabilities to empower every part of its workforce and turn AI into a daily operational asset.
GenAI.mil Puts Generative AI on Desktops Across the Force
GenAI.mil was launched on Dec. 9, 2025, as the War Department’s artificial intelligence platform, giving military personnel, civilians, and contractors access to generative AI tools across its networks.
The system is initially based on Google Cloud’s Gemini for Government, a specialized version of the Gemini model certified to handle Controlled Unclassified Information at Impact Level 5.
The launch forms part of a push to build an AI-first workforce and to deliver on the White House AI Action Plan announced earlier in the year, following a July mandate from President Donald Trump to secure a decisive lead in artificial intelligence.
The Department says AI services on GenAI.mil are now available on desktops throughout the Pentagon and at American bases worldwide. Secretary of War Pete Hegseth has urged all staff to adopt the platform and make AI part of their daily battle rhythm, while Under Secretary of War for Research and Engineering Emil Michael has framed the effort as a global race for AI dominance with no prize for second place.
How Deeply Has Washington Already Adopted Frontier AI?
Adoption of frontier AI inside the U.S. government is not new. In November 2024, Anthropic and Palantir Technologies said they had partnered with Amazon Web Services to give U.S. intelligence and defense agencies access to Anthropic’s Claude 3 and 3.5 models on AWS.
The models are available inside Palantir’s AI Platform, using Amazon SageMaker and an environment accredited at Impact Level 6, one of the strictest security categories approved by the Defense Information Systems Agency.
The companies say the setup is intended to help U.S. defense and intelligence personnel process large volumes of complex data, identify patterns, speed up document review, and support time-sensitive decisions while keeping final authority with human officials.
Palantir describes the deal as bringing the next generation of decision advantage to U.S. defense and intelligence customers and says it is the first industry partner to take Claude models into classified environments.
U.S. Approach Shows How Easily Frontier AI Can Be Adapted and Upgraded
The rollout of frontier AI across the War Department and U.S. security agencies shows how quickly officials are moving from small trials to making AI a normal part of defense and intelligence work.
This approach also reflects a decision to rely on powerful commercial models, which are then adapted and secured for government needs, rather than building everything in-house. This demonstrates how AI systems can be tailored to meet strict mission and security requirements as officials work to make AI a routine tool in national security rather than a side experiment.
Read More: Anthropic’s Opus 4.5 Sets New Benchmark in AI Engineering and Coding Performance
