Japan’s three largest banks are set to gain access to Claude Mythos, the powerful artificial intelligence model developed by U.S. startup Anthropic, in a move that could significantly reshape the cybersecurity landscape in Asia. According to Nikkei Asia, the decision is expected to take effect as early as the end of May, marking the first time Japanese companies have been granted entry to the highly restricted AI system.
The banks expected to receive access are MUFG Bank, Sumitomo Mitsui Banking Corp. and Mizuho Bank. The development was reportedly communicated during meetings in Japan involving U.S. Treasury Secretary Scott Bessent on Tuesday.
Claude Mythos has emerged as one of the world’s most closely guarded AI models due to its advanced capability to identify and exploit software vulnerabilities at unprecedented speed. Until now, access had reportedly been limited to around 50 corporations and organizations globally, including select American firms, U.S. financial institutions and British government bodies.
According to Nikkei Asia, Tokyo had been actively lobbying Washington to secure access to the technology as concerns grow over cyber threats targeting critical infrastructure and financial systems. Japanese officials view the model as a strategic asset in strengthening national cybersecurity defenses amid an escalating global AI race.
The move also aligns with directives issued by Japanese Prime Minister Sanae Takaichi, who has instructed cabinet members to intensify efforts to identify vulnerabilities within Japan’s infrastructure and prepare for risks associated with next-generation AI systems. Access to Mythos was reportedly among the issues discussed during Takaichi’s meeting with Bessent earlier this week.
