NSA Deploys Anthropic's Mythos With 6 Engineers for Cyberattacks Despite $1tn IPO Push
Updated
Updated · Financial Times · Jun 4
NSA Deploys Anthropic's Mythos With 6 Engineers for Cyberattacks Despite $1tn IPO Push
3 articles · Updated · Financial Times · Jun 4
Summary
Half a dozen Anthropic engineers have been embedded inside the NSA to tailor Mythos for offensive cyber operations, according to people familiar with the arrangement.
Mythos is valued for finding and exploiting software vulnerabilities and could help penetrate networks in countries such as China or Iran, reflecting a push to match adversaries' AI-driven attack tools.
The deployment comes even as Anthropic is suing the Pentagon over a "supply-chain risk" label tied to disputes over mass surveillance and lethal autonomous weapons.
Earlier this week, Anthropic widened Mythos access to 150 organisations in 15 countries after an April launch, as the company pursues an IPO that could value it above $1tn.
The NSA rollout also fits a broader US effort: Trump this week ordered voluntary AI security reviews and directed agencies to build methods for testing models' cyber capabilities.
Why is Anthropic suing the Pentagon on ethics while building cyberweapons for the NSA?
Is a private AI's hunt for cyber flaws creating an unpatchable global security crisis?
Can voluntary government oversight truly control AI firms building next-generation cyberweapons?
The NSA, Mythos AI, and the Patch Gap: How Advanced AI Is Reshaping Cybersecurity and Policy Worldwide
Overview
In June 2026, the NSA confirmed it is using Anthropic's Mythos AI, despite the Department of Defense's ongoing legal actions and blacklist against the company. This move shows that urgent national security needs are overriding official bans, creating a widening gap between policy and real-world operations. Mythos AI’s powerful offensive cyber capabilities have led Anthropic to restrict access to only 40 organizations, and the model is central to the exclusive Project Glasswing initiative. The NSA’s quiet adoption of Mythos raises serious concerns about governance, oversight, and accountability in the use of advanced AI for national security.