Anthropic expande o Projeto Glasswing, concedendo acesso ao Claude Mythos a 150 empresas de infraestrutura crítica em 15+ países

Key Points

• Anthropic adds roughly 150 new partners to Project Glasswing across more than 15 countries.
• Partners receive access to Claude Mythos, the firm’s most advanced AI model for finding zero‑day software flaws.
• Expanded sectors include power, water, health‑care, communications and hardware, previously under‑represented.
• Anthropic estimates a successful breach could impact over 100 million people worldwide.
• New participants span friendly nations such as Australia, Canada, France, Germany, India, Japan and South Korea.
• Key partners include Okta, Samsung, SK Hynix, SK Telecom, NATO and EU agency ENISA.
• The rollout follows Anthropic’s confidential IPO filing after a $65 billion funding round.
• OpenAI has introduced GPT‑5.5‑Cyber, signaling intense competition in AI‑powered cybersecurity.