Anthropic Plans Fourfold Expansion of AI Security Partnership to 200 Organizations

Anthropic on June 2, 2026 said it will expand its Glasswing partnership to roughly 200 organizations, up from about 50, granting more entities access to its Claude Mythos Preview AI for detecting software vulnerabilities. Partners must meet security requirements before accessing the system. Current members have identified over 10,000 highly or critically severe flaws since April. The rollout includes government bodies and companies in more than 15 countries across healthcare, power, water, and communications sectors. The move follows Anthropic's confidential IPO filing on June 1, 2026, which could be one of the largest ever, reflecting heightened demand for its security-focused AI amid concerns that such powerful tools could be used for cyberattacks if widely released.