The headline says JPMorgan is poaching Nomura's International Head of AI Strategy. Whether or not that specific person has accepted an offer, JPMorgan posted a job last week for Head of AI Strategy and Commercialization in its APAC division. The kind of role a Nomura AI strategy exec would fit.
But here's the thing that the headline gets wrong. JPMorgan already employs about 2,000 AI and machine learning experts. Its technology budget for 2026 is $19.8 billion, up nearly $2 billion from last year. CEO Jamie Dimon recently said the bank has already saved about $2 billion from its AI investment - roughly $2 billion spent, $2 billion back in cost savings, with what he calls "the tip of the iceberg" still to come. Dimon also told Bloomberg last month that JPMorgan will hire more AI specialists and fewer traditional bankers going forward.
The more interesting question isn't who the next hire is. It's whether any single hire matters at all when you're throwing $20 billion a year at the problem.
Think about the scale. $19.8 billion is bigger than the entire operating budget of most Fortune 500 companies. That is not a department. That is the entire technology infrastructure of a $4.4 trillion bank. At that size, hiring an AI strategy chief from a competitor doesn't give you an edge - it gives you a meeting calendar. The actual work is done by 2,000 people who already show up every day.
What JPMorgan is really doing is something slower and less headline-friendly. The bank has already displaced workers with AI, Dimon has admitted. It has "huge redeployment plans" for the people whose jobs automation replaced. The AI is embedded in lending, trading, client services - not sitting in a strategy office waiting for direction.
Here's the thing about banking AI that gets lost when the news cycle moves on to the next headline. This is not the kind of competition where one hire or one product wins the category. Every major Wall Street bank is doing the same thing. Goldman Sachs put $6 billion behind its own technology spending. Citi and the others are increasing tech budgets by roughly 10 percent this year. Even Morgan Stanley published a report predicting AI could eliminate 200,000 European banking jobs by 2030.
When everyone is doing the same thing, you don't win by hiring the strategy person. You win by having the most to automate.
That's JPMorgan's actual advantage. It has 318,000 employees. It operates across lending, trading, asset management, consumer banking, commercial real estate - a vast set of processes where even a small percentage improvement compounds into billions. A 10 to 11 percent productivity gain across that scale, which is roughly what Dimon has cited, is where the $2 billion in savings comes from. On a smaller bank, the same percentage doesn't pay for the investment.
This is superlinear returns in practice. Double the number of processes you can run AI on, and you don't just double the savings - you hit categories you couldn't touch before. The first 10 percent of AI projects are the obvious ones. The next wave requires the kind of scale where you can afford to experiment across dozens of business lines at once.
There's a limitation worth noting. Goldman Sachs Research published a piece just last week questioning whether the economics of AI investment are actually working out right now. The firm's Jim Covello says the case is more questionable today than it was two years ago. Annual AI capital spending is tracking above $800 billion for the year. A lot of that is infrastructure buildout that hasn't produced clear payback yet.
JPMorgan is different from most of those spenders because it is spending on internal use, not on building chips or data centers. The savings are already visible in its books. But the question remains whether the $2 billion Dimon cited is genuine cost reduction or reclassification of work that would have been automated anyway. He calls it the "tip of the iceberg." Icebergs are famous for being hard to see until you hit them.
The way to evaluate this isn't to watch personnel moves. It's to watch JPMorgan's operating expenses as a percentage of revenue over the next two quarters. If AI is doing what Dimon says it's doing, that ratio should compress even if revenue stays flat. If the ratio stays steady despite $20 billion in tech spending and thousands of AI staff, then the savings are a story, not a mechanism.
Most people think the AI winner in banking is whoever hires the smartest people and spends the most. But look at the math. In an industry where every bank is hiring and spending, the winner is the one with the most levers to pull. Size is the moat. Everything else is table stakes.
The test is simple: over the next year, watch expense ratios, not headlines. If JPMorgan's margin keeps expanding while competitors spend the same percentage and don't see the same result, you'll know it wasn't about the hire. It was about the scale all along.
