BreakthroughsTuesday, June 2, 2026· 2 min read

Anthropic Deploys Claude Mythos to Safeguard Critical Infrastructure in 15 Countries

TL;DR

Anthropic is scaling access to Claude Mythos and expanding Project Glasswing to 150 organizations across 15 countries, focusing on power, water, healthcare and communications. The move aims to bolster defenses for services that reach roughly 100 million people, improving detection, response, and resilience against cyberattacks.

Key Takeaways

  • 1Anthropic expands Project Glasswing and Claude Mythos access to 150 organizations in 15 countries.
  • 2Target sectors include power, water, healthcare and communications — services affecting ~100 million people.
  • 3Project Glasswing strengthens vulnerability discovery and remediation in critical infrastructure.
  • 4Deployment emphasizes practical resilience: faster detection, coordinated response, and continuity of essential services.
  • 5This real-world scaling could set a positive example for defensive AI use in high-stakes public systems.

Anthropic brings Claude Mythos and Project Glasswing to frontline infrastructure

Anthropic has announced a major expansion of its security program Project Glasswing alongside wider access to its Claude Mythos model, enabling 150 organizations across 15 countries to use the tools to protect critical infrastructure. The initiative focuses on sectors where cyberattacks can have severe societal consequences — power, water, healthcare and communications — covering services that reach an estimated 100 million people.

By combining Mythos’ capabilities with the structured vulnerability testing and remediation pipelines of Project Glasswing, Anthropic aims to help operators find, prioritize and fix security gaps more quickly. That means faster detection and response to threats, reduced downtime for essential services, and stronger overall resilience for systems communities depend on daily.

Key practical benefits include:

  • Improved vulnerability discovery through automated and human-in-the-loop analysis.
  • Faster, coordinated incident response that reduces service disruption.
  • Stronger continuity planning for utilities and healthcare providers serving large populations.
  • A collaborative model for deploying defensive AI safely in regulated, high-stakes environments.

This scaling represents a concrete, real-world deployment of AI focused on public safety and reliability rather than only consumer features. By prioritizing security and partnership with infrastructure operators, Anthropic’s move could meaningfully lower the risk of disruptive cyber incidents and encourage other AI developers to follow suit in building protective tools for essential services.

Get AI Wins in Your Inbox

The best positive AI stories delivered to your inbox. No spam, unsubscribe anytime.