Anthropic Scales Claude Mythos to Critical Infrastructure in 15+ Countries

Why It Matters

By securing codebases that support essential services, Anthropic helps prevent attacks that could affect over 100 million people, reinforcing global and national security. The move also positions the firm as a leader in AI‑driven cybersecurity ahead of its IPO.

Key Takeaways

• Project Glasswing adds 150 organizations in 15+ countries.
• Claude Mythos scans code for thousands of zero‑day flaws.
• Partners span power, water, healthcare, communications, hardware sectors.
• Potential attacks could impact over 100 million people worldwide.
• Anthropic’s IPO follows $65 billion funding, near $1 trillion valuation.

Pulse Analysis

Anthropic’s Project Glasswing, an industry‑wide effort to harden software against cyber threats, has entered its second phase with the rollout of Claude Mythos. The model, billed as the company’s most powerful, leverages large‑scale language‑model reasoning to comb through massive codebases and surface zero‑day vulnerabilities that traditional tools miss. In early‑April, a preview of Mythos was offered to 50 initial partners, including U.S. government agencies, proving its ability to identify thousands of previously unknown flaws within weeks. This AI‑driven approach marks a shift from reactive patching to proactive vulnerability discovery.

The latest expansion brings roughly 150 additional organizations from more than 15 nations into the Glasswing ecosystem, covering sectors that were under‑represented in the pilot—power grids, water treatment, healthcare delivery, telecommunications and hardware manufacturing. Companies such as Okta, Samsung, SK Hynix, SK Telecom, as well as strategic entities like NATO and the EU’s ENISA, now have direct access to Mythos for scanning the code that underpins critical services. Anthropic estimates that a successful breach of any partner’s code could affect over 100 million people, underscoring the model’s relevance to both national security and global supply‑chain resilience.

Anthropic’s aggressive rollout coincides with its confidential filing for an initial public offering, a move enabled by a $65 billion funding round that pushed its valuation close to $1 trillion. The timing signals confidence that AI‑powered cybersecurity will be a marquee growth engine for the company, especially as rivals like OpenAI have introduced GPT‑5.5‑Cyber for similar testing. By establishing Project Glasswing’s safeguards now, Anthropic aims to set industry standards and lock in early adopters before competitors catch up, positioning itself as the go‑to provider for AI‑enhanced protection of the world’s most vital digital infrastructure.