Anthropic is broadening its Project Glasswing initiative, deploying its AI-powered vulnerability scanner Claude Mythos to 150+ organizations across 15+ countries. This expansion targets power, water, healthcare, and infrastructure sectors where breaches could impact 100 million+ people. The move follows a $65 billion funding round that placed the company at a $1 trillion valuation, signaling its ambition to dominate AI-driven security.

The initiative builds on early partnerships with the U.S. government and now includes entities like Okta, Samsung, and the EU’s ENISA cybersecurity agency. Many new partners maintain codebases critical to global systems, making them high-value targets for attacks. Anthropic emphasized that Mythos Preview can identify thousands of zero-day flaws in weeks—a capability rivaled by OpenAI’s recent GPT-5.5-Cyber model. The company is racing to implement safeguards as competitors accelerate similar tools.

This expansion reflects growing reliance on AI for critical infrastructure protection. With adversaries exploiting software flaws, Mythos offers a proactive defense layer. However, Anthropic’s success hinges on balancing speed with security—ensuring its AI doesn’t introduce new risks. The $1 trillion valuation underscores investor confidence, but the real test lies in whether Mythos can prevent catastrophic breaches at scale. For now, the tool’s deployment in 15+ countries marks a pivotal step in securing the digital backbone of modern societies.