European banks face a growing cybersecurity gap as talks between the EU and Anthropic over access to the Claude Mythos Preview AI have stalled, according to Spanish officials in mid-May 2026.
The AI model, released on April 7-8, 2026, can identify and exploit zero-day software vulnerabilities at a speed and scale unmatched by human teams. Access, however, has been restricted to a handful of US partners, including Amazon and JPMorgan Chase.
EU Economy Commissioner Valdis Dombrovskis confirmed ongoing discussions on May 4, 2026, but minimal progress has been reported since. European financial institutions remain unable to use the tool to stress-test their defenses against next-generation cyber threats.
The exclusivity creates a competitive disadvantage for European banks in global markets, potentially affecting risk profiles, insurance costs, and investor confidence. In response, EU institutions are exploring alternative investments in homegrown AI cybersecurity tools and potential regulatory mandates for AI-powered penetration testing.
For fintech and cybersecurity investors, the stalled talks signal unmet demand in Europe for advanced AI-driven security solutions, which could accelerate a compliance-driven market for cybersecurity AI in the region.