The Financial Times report that landed on Monday is, on its face, about a trading firm's bad month. The Federal Reserve and the Bank of England have stepped up scrutiny of global banks' exposures to large trading firms including Jane Street and Citadel Securities, asking for details on the firms' risk appetite, on how each bank's exposure to them evolved during the trading day, and on how the banks' own risk controls operated. The proximate cause is a familiar one: heavy losses at Jane Street, reported at roughly $15 billion in July.
The last question in that list is the one that matters, and it is worth reading twice because it is unusual. Regulators normally ask where a bank's exposure stands. These regulators are asking how fast it moved, hour by hour, in the middle of a trading day. That is not a question about whether a trading firm is solvent. It is a question about whether a bank could see the risk it was carrying, in real time, while it was carrying it. The distinction is the whole story, and it is a story the rulebook was not built to tell.
The loss the probe is orbiting
The July event is the anchor. Situational Awareness, the AI-focused hedge fund run by former OpenAI researcher Leopold Aschenbrenner, was forced to sell most of its public equities portfolio to Citadel Securities after a sharp sell-off in AI and semiconductor stocks, and Jane Street took a reported loss of around $15 billion, partly through its investment in the fund and partly through other technology holdings. The scale is worth a benchmark: Jane Street reportedly generated about $40 billion in net trading revenue by early August, which means one of the world's most consistently profitable trading firms had a month that consumed a large share of its annual earnings in one event.
Every figure in that paragraph comes from unnamed sources in the Financial Times report, and the discipline of attribution matters here. Reuters said it could not immediately verify the FT's reporting. Jane Street and the Fed did not respond to requests for comment, and the Bank of England declined. What is established is the regulatory activity itself and its direction: two central banks asking banks, in detail, about their exposure to a class of counterparty that has grown faster than the rules written to contain it.
End-of-day capital is a lagging indicator
Bank capital rules are built on snapshots. A bank reports its positions, computes its risk-weighted assets, and demonstrates capital adequacy as of a point in time, generally the end of the day. The system assumes that a day is short enough that nothing material changes between snapshots. For most of banking history that assumption held: deposits, loans, and securities books did not typically rearrange themselves between sunrise and sunset in ways that a close-of-day number could not capture.
Electronic market making breaks the assumption. A trading firm's financing needs, margin requirements, and collateral calls move continuously, and a bank that finances, clears, or lends to such a firm can see its exposure swing by billions inside a single session, often in the direction it least expects. The Bank of England's Prudential Regulation Authority has already warned publicly about rising intraday exposures to electronic market makers. The FT report suggests the Fed has joined the conversation, and the specific question about how exposure evolved during the day is the tell: regulators have concluded that the close-of-day number is not merely imprecise but potentially misleading, because the risk they care about peaks at 3 p.m. and vanishes by the snapshot.
This is not an unknown gap, and that is part of why the scrutiny is notable. Global standard-setters put intraday liquidity on the supervisory map more than a decade ago, when the Basel Committee issued monitoring tools asking banks to track how payment and settlement flows move within a day. Those tools are supervisory diagnostics, not binding capital, and they were designed around payments, not around financing relationships with market makers whose positions can double against a bank in a single session. What the Fed and the Bank of England appear to be doing is extending the same lens to a counterparty class that has outgrown the assumptions the tools were built for, and the questions they are asking read like an inventory of what binding rules would have to capture: appetite, intraday evolution, and control failures in real time.
The transmission question is about banks, not funds
The deeper regulatory project has a name: non-bank financial intermediation, the broad category covering hedge funds, private credit, and specialist trading firms that now sit between savers and risk. The worry that animates it is transmission. If a leveraged fund fails, its lenders and prime brokers are on the other side of the failure, and losses travel through financing lines, margin calls, and clearing obligations into banks that never took the underlying bet. Jane Street's loss did not threaten Jane Street's survival, by all indications. The question regulators are asking is different: which banks were carrying that risk on their books during the worst days, and did they know it?
The Bank of England has separately examined the rapid growth in London prime brokers' financing of Asian equities, which shows the same concern through a different lens. Prime brokerage is where a bank lends to a fund against its portfolio, and the lending adjusts daily, sometimes hourly, as positions move. A fund in distress draws more credit exactly when its collateral is worth less, which is the mechanism by which a market sell-off becomes a bank problem. The regulators' interest in intraday exposure is an attempt to see that mechanism while it operates, rather than after it has closed.
That is also why the inquiry names Citadel Securities alongside Jane Street. A forced sale of a fund's portfolio, like the one that followed Situational Awareness's July troubles, concentrates risk inside the buyer rather than dispersing it, and the buyer in that transaction was a trading firm with its own web of bank financing. Regulators asking about both firms are following the same logic the market followed: when one large position moves, every balance sheet connected to it moves a little, and the connections are only visible if someone is looking at the right moment of the day.
The SEC is running the parallel track
The central banks are not the only regulators in the room. The SEC has subpoenaed Wall Street banks including Goldman Sachs, JPMorgan, Citigroup, and Bank of America as it examines Situational Awareness's trading activity and use of leverage, including the trades that triggered margin calls and the fund's communications with its lenders. JPMorgan has ended its lending relationship with the fund. The two inquiries are asking different questions: the SEC is reconstructing what happened inside the fund and its credit relationships, while the Fed and the Bank of England are asking what the banks knew about their own exposure while it was happening.
The parallel structure is itself diagnostic. When market risk materializes, the enforcement agency reconstructs the event and the prudential regulators interrogate the infrastructure. The fact that both are active, on the same event, through different legal channels, is the strongest available evidence that the July losses are being treated as a market-structure problem rather than a one-off trade gone wrong.
What a real answer would change
If the regulators conclude that intraday exposure is a genuine blind spot, the consequences would land on the banks rather than the trading firms. The menu is familiar from previous post-crisis cycles: tighter limits on intraday extensions of credit, more demanding monitoring requirements for financing relationships with market makers, and capital charges calibrated to peak exposure rather than end-of-day exposure. Each option costs the banks money, and each would be resisted on the grounds that prime brokerage and market financing are core businesses, not frills.
The counterargument deserves its place. Banks are not blind to intraday risk; they manage it with limits and margining systems that operate continuously, and no regulator has alleged that any bank failed to meet its obligations or suffered losses from the July event. Scrutiny is not a finding, and the central banks' questions could end in no action at all. What the questions reveal is where the supervisors' attention is pointing. In a market where a firm can lose $15 billion in a month and the regulator's first instinct is to ask banks what their exposure looked like at 3 p.m., the gap between how risk moves and how it is measured has become the subject of official curiosity. That is usually the stage before a rule.
The vehicle matters almost as much as the rule itself. The Federal Reserve just rewrote its stress tests, changing how the annual scenarios are built and what they demand of bank capital, and a supervisory interest in intraday counterparty risk would most naturally surface there, in scenario design that asks what a large trading firm's failure does to its financing banks over the course of a session. The stress-test cycle is where supervisory hypotheses become industry practice. If the intraday question makes it into a scenario, banks will answer it with systems long before any rule is finalized, which may be the point. Central banks ask questions like these in two registers at once: as investigators, and as instructors.
Primary sources
- Financial Times, the original outlet for the report on the Fed and Bank of England scrutiny; its reporting was used via syndicated copies because the original was not directly reachable.
- Reuters, whose syndicated copies supplied the details of the July losses, the SEC subpoenas, and the attribution caveats.
- Hedgeweek, for the analysis of the broader non-bank financial intermediation context and the PRA's intraday exposure warnings.