JPMorgan Holds AI Crown as Bank Deployment Speeds Up
A fifth straight win in the Evident AI Index lands as banks keep hiring and tech budgets climb
Photo by Etienne Martin on Unsplash
JPMorgan took the top spot in the Evident AI Index for a fifth year as banks speed up deployment. The real test is whether it shows up in margins.
Five years at the top, and the gap is widening
JPMorgan Chase finished first in the 2026 Evident AI Index, published October 6, for the fifth year running. The index scores 50 of the world's largest banks on AI maturity. Capital One held second, Royal Bank of Canada held third, and Commonwealth Bank of Australia stayed fourth. UBS, in sixth, is the only European lender in the upper tier.
JPMorgan ranked first or second on all four pillars: talent, innovation, leadership and transparency. It was first on leadership and transparency, and second on talent and innovation. Capital One leads on talent and innovation, so the top spot is a close contest on the inputs and a clearer win on disclosure and executive commitment.
The pace matters more than the podium. Evident found AI capabilities across the group advanced nearly three times faster this year than the average of the prior three. Every bank in the top 10 has placed there before. Leadership is stable while the field behind it compresses.
Why early movers keep pulling ahead
Evident CEO and co-founder Alexandra Mousavizadeh argues that early movers are compounding their lead. That fits what the numbers show. JPMorgan has earmarked roughly $19.8 billion for technology in 2026, and it now puts annual benefits from AI at around $2 billion, up from an earlier range of $1 billion to $1.5 billion. By that estimate, about a tenth of the tech budget is already being returned each year in measurable value.
That is a scale advantage rivals cannot copy quickly. Proprietary models, internal data pipelines and a deep bench of specialist talent take years to build. A bank starting now buys the same vendor tools, but it does not have the same history of deployment to learn from.
The disclosure trend supports the point. Twelve banks in the index now report actual or targeted AI value, up from eight a year ago. Bank of America said this month that its generative AI investment has turned positive. When the second tier starts publishing return targets, the conversation moves from experimentation to accountability.
The headcount puzzle
The most useful detail from Mousavizadeh's comments is that the banks investing most heavily in AI are still growing headcount. Jobs are changing, but the large spenders are not shrinking payrolls yet.
There are two historical analogs worth holding side by side. After ATMs spread through the 1980s and 1990s, branch economics improved, branches multiplied, and teller employment kept rising for years before it eventually fell. In the late 1980s, the Solow productivity paradox described computers showing up everywhere except in the statistics, until the productivity pickup of the second half of the 1990s. Technology adoption in banking has usually shown up first as role redesign and only later as lower headcount.
It would be premature to read today's hiring as proof that AI leaves labor untouched. The more defensible reading is that expense savings are lagging deployment, which is what both analogs would predict.
What it means for bank margins and sector positioning
For a macro desk, the question is whether AI can carry part of the earnings load as other drivers fade. Net interest income depends on the rate path and the shape of the curve, and neither is within a bank's control. The expense line is. If AI lowers the cost to serve, process and underwrite, the efficiency ratio improves even in a flat revenue environment. That is the case for the largest banks keeping a premium within financials.
The counterargument deserves weight. A $19.8 billion budget is also a $19.8 billion cost, and near term monetization remains uncertain. The index measures maturity, talent and transparency, not returns on invested capital. Transparency in particular favors banks that disclose more, and JPMorgan discloses a lot. Credit quality is the larger swing factor for bank earnings in any slowdown, and no model improves loan performance if the consumer turns.
For positioning, the [Sector Rotation dashboard](/sector) is the place to see whether money is treating large banks as a quality compounder or as a rate trade. Those are different holders with different exit triggers.
What to watch over the next two to four weeks
The third quarter earnings season for the big banks is the next real test. Listen for expense guidance, any change in headcount commentary, and whether management puts a dollar figure on AI benefits that matches or exceeds the $2 billion estimate. A rise in that number would support the early mover thesis. A flat number alongside rising tech spend would weaken it.
Outside the banks, credit spreads and loan loss provisions will say more about the financials trade than any AI ranking. If spreads stay tight and provisions stay contained, efficiency gains can drive upside. If spreads widen, the market will price credit first and technology second.
The setup changes if the second tier starts closing the gap on measurable returns, or if the large spenders begin cutting headcount faster than they have so far. Either would tell us the lag between deployment and savings is shortening.