As Wall Street banks finance and adopt AI, risks mount on both fronts
- Wall Street titans are financing an AI data-center feeding frenzy while admitting they have no clue how the loan repayments will pan out.
- Goldman's Solomon, in peak CYA mode, says nobody knows how the capital and compute demands will evolve as the ecosystem matures—classic 'trust me, bro' from the guys who sold us mortgage-backed securities.
- JPMorgan's Dimon is busy building 'tools to navigate agentic AI risk,' which is fancy talk for installing a fire extinguisher while the house is already ablaze.
- Citi's Fraser claims AI is already paying for itself, promising 40% faster product development and 30% cheaper vendors—conveniently leaving out the part where the robots may replace her actual employees.
- Investors are crossing fingers that productivity gains magically appear by year-end 2025; otherwise, the AI hype bubble is going to pop louder than a Wharton MBA's ego.
- Morgan Stanley is 'very discerning' about risk—until the next guy loosens standards and they all race to the bottom, again.
Brief Summary
In the grand tradition of chasing the next shiny thing right off a cliff, Wall Street's biggest banks are now both financing and adopting artificial intelligence with the enthusiasm of a frat boy at a kegger—and about as much foresight. Goldman CEO David Solomon admitted the entire endeavor is based on guesswork, uncertain how the demand for capital and computing power will shake out, while Citi's Jane Fraser insists there's 'simply not enough compute,' which is apparently the 2024 version of 'there's no stopping the housing market.' The banks are tripping over themselves to fund data-center mega-projects, using enormous leverage, all while pretending they've learned lessons from 2008. Meanwhile, they're deploying AI internally to cut costs, claiming productivity gains that are, as JPMorgan's CFO delicately put it, 'not yet observable'—translation: completely invisible. And lest we forget, these autonomous AI agents are creating fresh security nightmares that require 'ringfencing,' because what could possibly go wrong with giving software the power to act on customers' behalf? It's a beautiful mess: unprecedented investment, wild-eyed optimism, and the same old song where banks can't resist a good speculative binge with other people's money.
Why This Matters
For the average reader, this is the financial equivalent of watching a slow-motion car crash with a potentially cataclysmic finale. These AI-infrastructure bets are not just about Silicon Valley's toys; they're absorbing a massive chunk of the economy's investment capital, which could crowd out lending to small businesses and everyday consumers or, worse, trigger another systemic banking crisis when these speculative data-center loans turn sour. The banks' cheerleading is a direct signal that they are betting the house on AI, and if that bet fails—whether through overcapacity, energy constraints, or just a fad fading—the collateral damage will hit pensions, 401(k)s, and job markets, not just executive bonuses. The 'existential' importance Morgan Stanley sees in AI financing means every taxpayer has a vested interest in these gambling addicts not blowing the whole casino again.