Anthropic Deploys Claude Mythos AI to Secure Critical Infrastructure Across 15 Nations

What happened

Anthropic has significantly expanded Project Glasswing, an industry initiative that utilizes artificial intelligence to identify and fix critical software vulnerabilities. The company announced Tuesday that it is granting access to approximately 150 new organizations across more than 15 countries. This move scales the deployment of Claude Mythos, Anthropic's most powerful model to date, to entities responsible for maintaining global infrastructure.

What's new in this update

The program is moving beyond its initial pilot phase with 50 partners to include sectors that were previously underrepresented, such as power, water, healthcare, and hardware. New partners include the U.S.-based security firm Okta, South Korean giants Samsung and SK Hynix, and international bodies such as NATO and the European Union’s cybersecurity agency, ENISA. The expansion covers a broad range of U.S. allies, including Canada, France, Germany, India, Japan, and South Korea.

Key details

Anthropic describes Claude Mythos as a defensive tool capable of scanning massive codebases for security flaws that human researchers might miss. For most of the partners included in this expansion, Anthropic estimates that a successful cyberattack on their infrastructure could affect more than 100 million people. The initiative is designed to protect 'load-bearing' code that both governments and private industries rely upon for national security and public safety.

Background and context

The rollout follows news that Anthropic has filed confidentially for an initial public offering. The company recently completed a $65 billion funding round, bringing its valuation to nearly $1 trillion. This scaling of Project Glasswing also places Anthropic in direct competition with OpenAI, which recently released GPT-5.5-Cyber for similar infrastructure-testing purposes among its own partner group.

What to watch next

Anthropic expects other AI companies to develop models with capabilities rivaling Claude Mythos in the near future. The firm is now racing to establish industry-standard safeguards within Project Glasswing before these advanced capabilities become more widely available to potential bad actors. Observers will be monitoring the effectiveness of these AI-driven scans as they are applied to legacy systems in the power and water sectors.