Treasury and Fed Convene Emergency Meeting with Bank CEOs Over Anthropic Mythos AI Cyber Risk to Financial Infrastructure

On April 8, 2026, Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened an emergency meeting at Treasury headquarters in Washington with the chief executives of the country's largest systemically important banks — including Bank of America's Brian Moynihan, Citigroup's Jane Fraser, Goldman Sachs CEO David Solomon, Morgan Stanley's Ted Pick, and Wells Fargo CEO Charlie Scharf — to warn them about cybersecurity risks posed by Anthropic's newly launched artificial intelligence model, Claude Mythos Preview. JPMorgan's Jamie Dimon was the only major banking CEO unable to attend. The meeting was first reported by Bloomberg and confirmed by CNBC.

The gathering was triggered by the release of Claude Mythos Preview under Anthropic's "Project Glasswing" initiative — a restricted access program involving 12 partner organizations in major technology and financial firms. The model represents what Anthropic describes as a step-change in AI capability, and its safety evaluation revealed an emergent and unintentionally trained ability to autonomously identify and exploit zero-day software vulnerabilities across major operating systems and web browsers. Anthropic researchers disclosed that Mythos identified thousands of previously unknown zero-day flaws — many of them one to two decades old — raising immediate concern that the same capabilities enabling defensive security scanning could be weaponized by adversaries.

Bessent and Powell assembled the bank executives to ensure the largest U.S. financial institutions are aware of risks posed by Mythos and similar future models, and are taking precautions to harden their systems, according to people familiar with the matter who asked not to be identified citing the private discussions. The surprise meeting was described by The Street as a signal that the Trump administration views AI-driven cyber risk as a financial stability concern, not just a technology problem.

Anthropic confirmed it has been in "ongoing discussions" with U.S. government entities including the Cybersecurity and Infrastructure Security Agency and the Center for AI Standards and Innovation about Mythos' offensive and defensive capabilities prior to its limited release. CEO Dario Amodei wrote: "The dangers of getting this wrong are obvious, but if we get it right, there is a real opportunity to create a fundamentally more secure internet and world than we had before the advent of AI-powered cyber capabilities."

The meeting adds a new dimension to an already contentious relationship between Anthropic and the Trump administration. The Pentagon has designated Anthropic a supply-chain risk to national security — a classification Anthropic is contesting in court. A federal appeals court declined earlier this week to temporarily block the Pentagon's designation. President Trump and Defense Secretary Hegseth have publicly criticized the company for imposing limits on the use of its AI in military applications. The intersection of Anthropic's advanced offensive cyber capabilities, its contested Pentagon designation, and its restricted model release to financial and critical infrastructure partners represents a convergence of AI safety, national security, and financial system stability concerns at the highest levels of the U.S. government.