The Bank of England’s July 2026 Financial Stability Report raises the prominence of frontier AI cyber risk, noting it could change the speed, scale and economics of cyber risk. The Financial Policy Committee finds that advances in frontier AI since December 2025 have increased financial stability risks from cyber and operational vulnerabilities. AI Security Institute testing shows the latest frontier models can now execute multi-stage attacks with little human input, passing “cyber range” tests for the first time. The FPC sets out three indicative scenarios: sustained operational strain and new concentration risk around frontier AI vendors; correlated disruption through shared providers; and a rising backlog of unfixed vulnerabilities triggering a system-wide event. On the defensive side, Anthropic identified 23,019 candidate vulnerabilities in open-source software, disclosing 1,596 across 281 projects. Palo Alto Networks’ May 2026 update highlighted 26 publicly recorded vulnerabilities and 75 security issues, far above its usual monthly total of fewer than five. The report confirms an upcoming Bank/Prudential Regulation Authority consultation on cyber and ICT risk management, and the Bank is considering AI scenarios in cyber and operational testing, including CBEST and STAR-FS, with emphasis on simultaneous multi-firm disruption.