AI Officially Enters the Heart of the Pentagon

Tehran - BORNA - In the final days of summer 2025, a major development in artificial intelligence and U.S. national security was announced. Anthropic, one of the leading AI companies, revealed that it has developed a special version of its language model called *Claude Gov* for use within the government and security agencies.

At the same time, the company announced the formation of a National Security and Public Sector Advisory Council, comprising prominent officials and security experts. This council is tasked with guiding the integration of AI into America’s most sensitive institutions.

This was not merely a technical unveiling it signaled the deeper entry of AI into the realms of national security and politics, where billions of dollars in investment, ethical debates, geopolitical rivalries, and public concerns converge.

A $200 Million Pentagon Contract

The highlight of this development is a \$200 million agreement between Anthropic and the U.S. Department of Defense. The deal, structured as a two-year prototype project with the Pentagon’s Chief Digital and AI Office (CDAO), will see Anthropic’s advanced models tested in key national security domains from analyzing classified documents to supporting cyber operations.

The contract reflects the broader race among tech giants. OpenAI, Google, and xAI have each secured similar deals, but Anthropic’s move is notable: despite being smaller than its rivals, it has now entered the highest levels of national security competition.

Claude Gov: An AI for Secure Rooms

As its name suggests, Claude Gov is tailored for “government” use. Compared with Anthropic’s public-facing models, it brings distinct features:

Greater flexibility in response: Unlike public models that maintain stricter ethical and safety limits, often refusing to answer, Claude Gov is tuned to “decline less often.” This is considered essential for classified environments.

Security and confidentiality focus: The model operates on hardened infrastructure capable of handling sensitive and classified data.

Research and analysis applications: From Lawrence Livermore National Laboratory to other federal institutions, more than 10,000 researchers now have access to Claude Gov.

Anthropic stresses that the system still operates under strict safety controls to prevent misuse, but it is more adaptable than public versions.

The National Security Advisory Council: Where Tech Meets Policy

To ensure the deployment of Claude Gov serves U.S. national interests, Anthropic has formed a National Security and Public Sector Advisory Council.

This 11-member body includes high-profile figures such as Jill Hruby, former director of the National Nuclear Security Administration; Dave Luber, former cybersecurity leader at the NSA; David Cohen, former CIA deputy director; and Richard Fontaine, CEO of the Center for a New American Security.

The council’s composition highlights Anthropic’s ambition not just to be a tech provider, but to position itself as a strategic player in national security policymaking.

Infrastructure and Government Access

Anthropic runs *Claude Gov* on powerful infrastructure, including Amazon’s Rainier cloud clusters powered by Trainium2 chips, while also leveraging Google’s TPUs for training.

More striking is Anthropic’s agreement with the U.S. General Services Administration (GSA). Under this deal, federal agencies including the executive, judicial, and legislative branches can access Claude for as little as \$1 per license. This move positions Claude to become the common AI platform across government in record time.

Strategic Implications

Anthropic’s entry into national security carries significant consequences:

Strengthening America’s position in the global AI

While China emphasizes civilian applications, the U.S. is reinforcing its lead in military and security AI.

Expanding federal access to cutting-edge tools:The \$1 GSA deal suggests Anthropic aims to embed itself deeply in government operations.

Providing security agencies with tailored AI: From the Pentagon to the CIA and NSA, Claude Gov can serve needs unmet by public models.

Anthropic now stands at a critical juncture. If it turns these contracts into tangible results, its stature in the AI industry will rise, making it a central pillar of U.S. national security. But failure to navigate political, technical, and ethical challenges could expose it to public backlash and competitive pressure.

A Turning Point for AI in National Security

With the launch of Claude Gov and the creation of its security advisory council, Anthropic has signaled one of the clearest shifts yet: artificial intelligence is no longer just a tool for generating text or images—it is becoming part of the machinery of national defense and global politics.

This step offers major opportunities to strengthen U.S. defense capabilities and governance but also raises profound concerns around ethics, transparency, and geopolitical competition.

The future will reveal whether Anthropic and Claude Gov can strike a balance between “technological power” and “social responsibility.”

End article