Anthropic Faces Controversy Over Cyber AI Model Access

Anthropic has been at the center of controversy, with the company denying EU access to its advanced AI model and raising national security stakes as a potential DOD vendor.

Anthropic's Mythos cybersecurity AI model — which can autonomously discover and exploit zero-day vulnerabilities — has become the center of a transatlantic access dispute. The company is withholding Mythos from EU markets while OpenAI has agreed to grant European regulators access to its own cybersecurity model, creating a visible divergence in how the two frontier labs are approaching government relations.

The DOD dimension adds complexity: Anthropic has not been selected as a Pentagon AI vendor. In May 2026, the Defense Department finalized a seven-vendor shortlist — OpenAI, Google, Microsoft, AWS, Nvidia, SpaceX, and Reflection AI — explicitly excluding Anthropic amid an ongoing dispute over military use of Claude models dating to a January 2026 designation of Anthropic as a "supply chain risk". The Trump administration has described a potential resolution as "possible," but no deal has been reached.

Separately, Anthropic warned investors in April 2026 against unauthorized trading of its equity through token markets, where the company's implied valuation has approached $1 trillion. The combination of regulatory friction in Europe, exclusion from U.S. defense contracts, and pre-IPO valuation pressures makes Anthropic's policy posture a material business risk — not just a headline controversy.