Anthropic Expands Project Glasswing Security Initiative to 150 Organizations Across 15+ Countries

Summary

Anthropic has announced a major expansion of Project Glasswing, extending access to Claude Mythos Preview to approximately 150 new organizations globally. Building on an initial cohort of roughly 50 partners who have already identified more than 10,000 high- or critical-severity security vulnerabilities, the new phase includes organizations from more than 15 countries spanning critical infrastructure sectors previously underrepresented in the program, including power, water, healthcare, communications, and hardware.

The expanded group consists primarily of vendors and infrastructure providers whose codebases are relied upon by millions of organizations and governments worldwide. Anthropic estimates that a successful cyberattack on most partners could affect over 100 million people and have significant implications for global and national security. The expansion comes after weeks of collaboration with existing Project Glasswing partners, the security industry, open-source maintainers, and the U.S. government.

Alongside the expansion, Anthropic introduced Claude Security, a new product using its frontier models like Claude Opus 4.8 to help organizations scan codebases and suggest patches. The company emphasized that Project Glasswing is designed to prepare the cybersecurity industry for a near-term future where multiple AI companies will have advanced cyber capabilities. Anthropic plans to shift its support model from vulnerability detection toward disclosure, fixing, and deploying patched software, while also sharing tools developed for the program with trusted security teams.

• Anthropic launched Claude Security, a new product for codebase scanning and patch suggestions using frontier models
• The initiative is designed to prepare cyberdefenders for a near-term future where many AI companies will have advanced cyber capabilities