Skip to main content

The FPC's July warning: leverage, AI concentration, and a thinner margin for error

The Bank of England's Financial Policy Committee has flagged a sharper build-up of vulnerabilities across equity leverage, AI-driven market concentration, and cyber resilience, even as the UK system holds up. For senior leaders, the message is that the correlation of risks, not any single one, is what now demands board attention.

The Financial Policy Committee's July record does not announce a crisis. It does something more uncomfortable: it stacks the vulnerabilities it already named in December and tells the market they have become more entangled. Risky asset valuations, sovereign debt strains, private credit exposures, and a "substantial increase in the use of leverage in equity markets" now sit alongside a Middle East supply shock and rapid advances in frontier AI (Bank of England). The Committee's judgement is that the likelihood of these crystallising together has risen since December.

Leverage is quietly the headline

The FPC singles out a significant rise in hedge fund activity and increased equity leverage, at a point when global indices are being pulled upward by a narrow set of AI-related names (Bank of England). That combination, stretched valuations, concentrated exposures, and more borrowed money behind them, is the textbook setup for a disorderly unwind. Asset managers and prime brokers should read the record as a supervisory signal that margining practices, counterparty concentration, and non-bank leverage data will be probed more assertively in the second half of 2026. Boards that treated the December FSR as background reading no longer have that luxury.

AI has moved from opportunity to stability risk

The FPC explicitly links frontier AI advances to heightened cyber and operational resilience risk (Bank of England). That framing matters because it arrives days after the FCA's Mills Review set out how agentic AI will reshape retail financial services by 2030, with a fifth of UK adults, roughly 11 million people, already open to autonomous AI acting on their behalf (FCA). The prudential and conduct regulators are now converging on the same concern from opposite ends: the FCA on consumer trust and market power, the FPC on operational fragility and concentration in the infrastructure firms depend on. Executives running third-party risk, model governance, or cloud strategy should expect questions that assume both lenses apply simultaneously.

Conduct risk is not on holiday

While the FPC frames the macro picture, the FCA continues to press on the retail floor. Nine of the largest banks have been told to fix basic bank account provision after a mystery shop found a third of experiences rated poor or very poor, with firms pushing vulnerable customers toward unsuitable online journeys (FCA). Emad Aladhal, director of retail banking at the FCA, said the regulator will be "holding them to account to make sure change happens" (FCA). The Eldens Finance administration, announced 6 July, is a reminder that firm-level failure continues to land on customers holding pledged assets and on administrators managing orderly wind-down (FCA). Neither item is systemic. Both consume senior bandwidth precisely when macro-prudential attention is sharpening.

The implication

The operating assumption for the rest of 2026 should be that regulators expect firms to hold multiple risk narratives at once: leverage-driven market stress, AI-linked operational fragility, and conduct outcomes for the most exposed customers. Boards that force a single-issue agenda will be caught out. The firms that fare best will be those whose risk committees can articulate, in one sitting, how a disorderly equity unwind, a cloud or model outage, and a Consumer Duty failing would interact on their own balance sheet.

Polar Insight helps senior leaders in financial services understand what their key stakeholders actually think before significant decisions are made.

Book a conversation