A sharp rotation is unfolding in the US equity market. Since the start of June, the State Street Financial Select Sector SPDR ETF (XLF) has rallied more than 8%, while the S&P 500 has barely moved. Over the same period, the Roundhill Magnificent Seven ETF (MAGS) — a basket of the tech titans that once led the market — has dropped nearly 4%. Money is flowing out of artificial intelligence plays and into financials, particularly bank stocks.
The superficial drivers behind this shift are easy to identify. The interest rate backdrop is the dominant force. Although the federal funds rate has come down from its 2024 peak to just above 3.5%, it remains elevated by the standards of the past 17 years, and expectations are hardening that rates will stay higher for longer. That has translated into wider net interest margins for lenders. Bank of America, for instance, reported a 9% year-on-year jump in net interest income in the first quarter.
At the same time, AI stocks that were once must-own names have lost some of their shine. Commercial adoption has underwhelmed, and the enormous capital expenditure pouring into AI infrastructure still lacks a clearly defined path to return. Investors are trimming exposure and hunting for cheaper corners of the market. JPMorgan Chase trades at just 15 times forward earnings, while Bank of America sits at a little over 13 times — deep discounts that are luring capital. A revival in IPOs and dealmaking is adding to the appeal: global IPO proceeds surged more than 200% year on year in the first half of 2026, punctuated by SpaceX’s record-breaking public debut, fattening investment banking fee pools.
Yet treating this rotation as a straightforward swap — out of tech, into banks — misses a crucial truth: Wall Street banks are already deeply exposed to the very AI cycle investors think they are escaping.
Consider the latest numbers. Combined equity and debt trading revenue at JPMorgan Chase, Bank of America, Citigroup and Goldman Sachs hit $38 billion, up more than a third from a year ago and 60% higher than two years ago. Investment banking fees topped $10 billion in the quarter. That extraordinary performance was fuelled by the churning markets and the wave of capital-raising unleashed by the AI boom. The stock-price gains tell a similar story: Goldman Sachs, Morgan Stanley and Citigroup have all more than doubled over the past 24 months, while even JPMorgan and Bank of America — which house large commercial and consumer operations — have comfortably beaten the S&P 500. The financial sector’s rally is not a hedge against AI; it is a side effect of it.
That is why investors trying to de-risk by rotating within the sector — selling pure-play investment banks such as Goldman Sachs and Morgan Stanley and buying more diversified universal banks — are achieving far less than they hope. A sharp unwind of the AI trade would not only crush trading and underwriting revenues. It would erode credit quality in commercial lending, shrink client assets in wealth management units, and, if the Federal Reserve were to cut rates to stabilise the economy, squeeze net interest margins. Put simply, the big banks have become unadulterated AI-correlated assets.
To be sure, the strength in financials rests on solid economic footings. The New York Fed’s model puts the probability of a US recession over the next 12 months at just 16%, and the IMF expects global GDP growth to accelerate to 3.4% in 2027. A healthy economy will continue to support loan demand and capital markets activity. But investors should be clear-eyed about what this rotation actually represents. It is not a divorce from the AI theme; it is a repositioning within it. If the AI wave recedes, two sets of assets that look entirely different today could end up exposed to exactly the same risk.
