Anthropic Expands Claude Mythos to Secure Critical Infrastructure Across 15+ Countries

Anthropic is significantly scaling Project Glasswing, its joint industry initiative to identify and remediate critical software vulnerabilities using artificial intelligence, bringing the total number of participating organizations to approximately 150 across more than 15 countries. The expansion, announced Tuesday, extends access to the company’s most powerful AI model, Claude Mythos, to sectors that were previously underrepresented in the program’s initial pilot phase.

Expanding Beyond Government to Critical Sectors

In early April, Anthropic granted 50 initial partners, including the U.S. government, access to Claude Mythos Preview to scan their codebases for security flaws and zero-day vulnerabilities. The new cohort now includes organizations operating in power generation, water utilities, healthcare systems, telecommunications networks, and hardware manufacturing — industries that Anthropic described as not well-represented in the first group.

Many of the newly onboarded partners are companies or nonprofits that maintain codebases relied upon by other organizations and governments globally. According to Anthropic’s blog post, what unites these partners is the potential scale of a successful cyberattack on their systems. The company estimates that a major breach at most of these organizations could affect more than 100 million people, with significant ramifications for both global and national security.

Geographic and Organizational Scope

The expanded list of participating organizations spans countries considered friendly to the United States, including Australia, Canada, France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, India, Japan, New Zealand, and South Korea, according to The Financial Times, citing a person familiar with the matter. The FT also reported that notable organizations now granted access include U.S.-based identity and security management firm Okta; South Korean companies Samsung, SK Hynix, and SK Telecom; NATO, the U.S.-led military alliance headquartered in Brussels; and the European Union’s cybersecurity agency ENISA. Bitcoin World has reached out to Anthropic to confirm these specific partnerships.

Why This Matters for Global Security

The expansion reflects a growing recognition that critical infrastructure software supply chains represent a concentrated attack surface. By deploying Claude Mythos — which Anthropic has described as its most powerful model yet, capable of identifying thousands of zero-day vulnerabilities over several weeks — the initiative aims to proactively close security gaps before they can be exploited. The model’s ability to analyze codebases at scale offers a significant advantage over traditional manual auditing methods, which are often slow and resource-intensive.

Anthropic has acknowledged that rival AI companies are likely to develop similarly capable models in the near future. This expectation is already materializing: since the release of Mythos, OpenAI launched GPT-5.5-Cyber, a cybersecurity-focused model that has been rolled out to a large group of partners for testing. Anthropic’s strategy appears to be one of establishing protective frameworks and trust networks within Project Glasswing before the competitive landscape becomes more crowded.

Conclusion

Anthropic’s expansion of Project Glasswing represents a notable step in the application of advanced AI to national and global cybersecurity challenges. By extending Claude Mythos to a diverse set of critical infrastructure operators across more than 15 countries, the initiative aims to preemptively address vulnerabilities that could have catastrophic consequences. As both public and private sector entities increasingly rely on interconnected software systems, the success of such collaborative AI-driven security efforts may become a benchmark for the industry. The coming months will reveal whether this model of proactive vulnerability scanning can scale effectively and whether it will influence how other AI companies approach cybersecurity partnerships.

FAQs

Q1: What is Claude Mythos and how does it differ from other AI models?
Claude Mythos is Anthropic’s most powerful AI model, specifically designed for cybersecurity applications. It is capable of scanning large codebases to identify zero-day vulnerabilities — previously unknown security flaws — over a period of several weeks. This makes it distinct from general-purpose AI models by being optimized for deep, systematic code analysis.

Q2: Which types of organizations are included in the expanded Project Glasswing?
The expanded group includes organizations in critical infrastructure sectors such as power, water, healthcare, communications, and hardware manufacturing. Partners range from large corporations like Samsung and Okta to international bodies like NATO and ENISA, all of which maintain codebases that, if compromised, could affect over 100 million people.

Q3: Why is Anthropic expanding this initiative now?
Anthropic is racing to establish safeguards and trust networks within Project Glasswing before other AI companies develop similarly capable cybersecurity models. The recent release of OpenAI’s GPT-5.5-Cyber underscores the competitive urgency. By expanding now, Anthropic aims to set industry standards for responsible AI use in critical infrastructure protection.