Megabanks Just Made AI an Operating Cost

At the five largest US banks, AI has stopped being a discretionary innovation bet and is now a permanent line in the operating model. JPMorganChase, Bank of America, Citigroup, Goldman Sachs, and Wells Fargo each beat Q2 2026 expectations, and four of the five July 14, 2026 earnings calls — JPMorgan, Bank of America, Goldman Sachs, and Citigroup — framed the technology spend as baseline budget, not strategic theme. That is a different object than a transformation narrative. It is a cost line the banks expect to carry forward, every quarter.

The mechanism generalizes: when a sector's biggest players treat a new technology as operating cost in the same reporting week, the capital cycle has already priced it in. The reframing is the news: 'strong megabank earnings' in 2026 and 2027 means earnings that absorb an AI line item, not earnings that exclude one.

Jamie Dimon's caveat travels with the thesis: the economic surplus, he told investors, will not stay inside the banks. David Solomon of Goldman Sachs called the buildout the 'very early innings' of a 'very, very significant' cycle and tied it to data center and energy financing, so AI spend is also a financing story, not only a cost story. Citigroup's Jane Fraser signaled the bank will step up technology investment in the second half even as expense growth draws scrutiny. PYMNTS' coverage shows the simultaneity: five CEOs, one week, one frame.

Watch four more cycles. If expense growth stays inside the beat, the operating model has changed. If the line hides in goodwill or chip depreciation, the picture shifts.