Anthropic recruits D.C. power players for new national security council

For the longest time, AI safety was mostly about preventing models from hallucinating or teaching someone how to synthesize toxic chemicals. Now, frontier labs are redefining safety to mean something much bigger: national security.

Anthropic has officially formed its National Security and Public Sector Advisory Council. The new group brings together high-profile, bipartisan veterans from across the defense, intelligence, and energy spheres to help steer Claude into public sector and defense operations.

A bipartisan roster of national security heavyweights

The inaugural 11-member lineup reads like an all-star roster of Washington defense policy, and the political balance looks deliberate.

The council features figures from both sides of the aisle and across administrations. On the defense side, there is former Acting Secretary of Defense Patrick Shanahan, who served under the first Trump administration and previously spent over three decades at Boeing. Joining him is former Senator Jon Tester, who previously oversaw an $800 billion defense and intelligence spending portfolio as head of the defense appropriations subcommittee.

The intelligence community is equally represented. Former CIA Deputy Director David Cohen and former NSA Cybersecurity Director Dave Luber have joined the council. On the nuclear and energy front, the team includes former NNSA administrators from both the Trump and Biden administrations: Lisa Gordon-Hagerty and Jill Hruby.

Other notable members include former Missouri Senator Roy Blunt, along with former top national security advisors to House Speaker Mike Johnson and Senate Majority Leader Chuck Schumer.

Clearing the runway for public sector deals

Anthropic says the council will focus on high-impact applications in cybersecurity, intelligence analysis, and scientific research. But this is not an isolated PR push; Anthropic has been aggressively building out its public sector pipeline for months.

The company previously rolled out Claude Gov, a dedicated model line built specifically for U.S. national security users. It also inked a $200 million prototyping deal with the Department of Defense.

Beyond military contracts, Anthropic deployed Claude to 10,000 scientists at the Lawrence Livermore National Laboratory and partnered with the Department of Energy to audit its models for potential nuclear proliferation risks. To get its foot firmly in the door, the company even offered Claude to all three branches of the federal government for just $1.

Why it matters

We think this advisory council marks a pivotal shift in Anthropic's evolution from a safety-first research outfit into an institutional defense contractor.

Government and defense budgets are massive and notoriously sticky once you are inside the wire. Selling enterprise software is tough right now, but pairing deep technical safeguards with influential former officials who understand federal procurement is a masterclass in enterprise sales strategy. It gives Anthropic an immediate edge over rivals vying for lucrative public sector grants and agency contracts.

For the wider industry, this cements a new reality: frontier AI is officially national infrastructure. Winning the frontier race is no longer just about benchmarks or consumer hype; it is about geopolitical alignment and strategic competition with foreign adversaries.

Still, there are obvious tensions ahead. Anthropic made a name for itself by pledging responsible, cautious development. As it embeds itself deeper into defense prototypes and intelligence workflows, it will have to prove that its strict guardrails can survive contact with the realities of military procurement.