source: techcrunch ai: anthropic scales claude mythos to critical infrastructure in 15+ countries
level: business
anthropic is scaling its project glasswing initiative, which uses the claude mythos model to scan code for security flaws. the company now grants access to about 150 new organizations across more than 15 countries. these include sectors like power, water, healthcare, communications, and hardware that were not well represented in the first group of 50 partners. many of the new participants maintain codebases that other organizations and governments depend on. anthropic says a successful attack on their systems could affect over 100 million people and threaten global security.
the expanded list covers countries friendly to the u.s., such as australia, canada, france, germany, india, japan, and south korea. some named recipients include okta, samsung, sk hynix, sk telecom, nato, and the eu cybersecurity agency enisa. anthropic first gave early access to claude mythos preview in april, allowing partners to find thousands of zero-day vulnerabilities over several weeks. the model is described as the company's most powerful yet for this task.
anthropic is moving quickly because it expects other ai firms to soon release similarly capable models. rival openai has already launched its own cybersecurity-focused model, gpt-5.5-cyber, for partner testing. the push comes as anthropic has confidentially filed for an ipo following a $65 billion funding round at a nearly $1 trillion valuation. the company aims to set safeguards and establish its role in protecting critical software infrastructure before the field becomes crowded.
why it matters: ai-driven vulnerability detection can help secure essential services that billions rely on, reducing the risk of catastrophic cyberattacks.
source: techcrunch ai: anthropic scales claude mythos to critical infrastructure in 15+ countries