The Department of Defense has awarded $200 million in contracts to Google, Anthropic, xAI, and OpenAI to ramp up its use of AI in national security.
According to its Chief Digital and Artificial Intelligence Office (CDAO), the deals bring top-tier tech muscle into defense operations to support US warfighters and accelerate deployment across mission-critical systems.
CDAO leans on top AI firms to move fast
CDAO structured the agreements with a ceiling of $200 million each, with a goal of building task-specific systems and agentic workflows for defense use at speed and scale.
The initiative targets command systems, data infrastructure, and core operations to embed AI tools into daily defense workflows. “Leveraging commercially available solutions into an integrated capabilities approach will enable the use of advanced AI,” said Chief Digital and AI Officer Dr. Doug Matty.
Anthropic, Google, OpenAI, and xAI will work with defense units to prototype frontier models aligned with mission requirements. AI systems will operate on platforms such as Advana, Maven Smart System, and the Edge Data Mesh, enabling in-place deployment at the infrastructure level.
To expand reach, CDAO is teaming up with the General Services Administration (GSA) to open procurement channels within the federal government. The commercial-first strategy is designed to streamline access to cutting-edge models.
What’s in the Pentagon AI stack
Each of the four companies awarded contracts by the Department of Defense has outlined what it plans to contribute. From secure cloud platforms to mission-specific models, here’s what Anthropic, Google, OpenAI, and xAI say they’re bringing to the table.
OpenAI: Custom models and operational tools
OpenAI was the first to announce its contract with the Department of Defense, as part of its OpenAI for Government program launched in June. The company is providing its newest models, including ChatGPT Gov and custom national security variants within secure, compliant environments. Use cases span healthcare access for service members and proactive cyber defense.
Anthropic: Responsible AI and Claude Gov models
Anthropic will prototype frontier AI systems tailored to national security needs, focusing on reliability, interpretability, and adversarial risk mitigation. Its Claude Gov and Claude for Enterprise models will support classified deployments and integrate into existing defense workflows. Anthropic also emphasized direct collaboration with DoD teams to fine-tune outputs on department data.
Google: Infrastructure, TPUs, and cloud AI
Google Public Sector will give the Pentagon access to its secure AI infrastructure, such as Cloud TPUs, Agentspace, and air-gapped systems authorized at IL6 and Top Secret levels. The contract builds on existing work with the Navy, Air Force, and Defense Innovation Unit. Google’s distributed cloud systems will help scale AI within the DoD’s enterprise architecture.
xAI: Grok 4, mission-specific tools, and forward-deployed support
xAI is delivering a suite of tools under its new Grok for Government platform, including Grok 4, Deep Search, and custom-built AI for classified and national security environments. The company is offering forward-deployed engineering, cleared personnel, and tailored deployments for use cases in science, healthcare, and defense. Products will be made available across federal agencies via the GSA schedule.
Uncommon ground but a common goal
With the contracts now in place, the Defense Department has pulled four rival firms into a shared operational plan.
In a field defined by one-upmanship, these four AI leaders, with their own models and ambitions, are now working in sync. If the coordination holds, the Pentagon may get something rare: The sharpest minds and the most advanced systems working in service of national defense.
Collaboration isn’t stopping at defense. Read about how Microsoft, OpenAI, and Anthropic are joining forces to launch a $23 million training hub for educators.