JPMorgan treats AI as core spend

- JPMorgan has started treating AI spending as basic bank infrastructure, not a side experiment — folding it into the same budget logic as core systems. - The signal came from CFO Jeremy Barnum in 2025 and 2026 investor materials, alongside a tech budget running near $19.8 billion. - That matters because once AI becomes “must-run” plumbing, the debate shifts from proving ROI to keeping the platform fast and reliable.

A big bank budget change can sound boring. But this one is not. JPMorgan is basically telling investors that AI is no longer a pilot project with a maybe-payoff someday. It now sits with the stuff a bank simply has to fund — compute, infrastructure, resilience, security, and the rest of the machinery that keeps a giant financial system running. ### What actually changed? The important shift is not that JPMorgan “likes AI.” Every large bank says that now. The shift is that AI spending is being talked about as core infrastructure spend rather than discretionary innovation spend. That changes who has to justify the money, how fast projects get approved, and what counts as necessary maintenance versus optional experimentation. You can see the setup in the bank’s 2026 company update, where Barnum grouped “strong infrastructure and AI-related spending” together instead of treating AI as a separate moonshot bucket. (jpmorganchase.com) ### Why does the accounting frame matter? Because budgets are strategy in disguise. If something lives in the “innovation” bucket, it competes with dozens of other experiments and gets judged on short-term proof. If it lives in the “core platform” bucket, the question becomes whether the bank can afford *not* to build it. That is a very different conversation inside a company that processes about $12 trillion of payments a day, serves more than 86 million U.S. customers, and operates in over 100 markets. (jpmorganchase.com) ### Why would JPMorgan do this now? Scale. JPMorgan is already one of the heaviest technology spenders in banking. Its 2026 technology budget has been described around $19.8 billion in commentary tied to the firm’s investor materials, and its 2025 investor day framed AI as a firmwide operating priority rather than a niche lab effort. Once AI tools are being used across coding, service workflows, fraud, research, and employee productivity, the infrastructure underneath stops looking optional. (jpmorganchase.com) ### Is this about chatbots? Not mainly. The flashy part of AI is the assistant sitting on top. The expensive part is the plumbing underneath — GPUs, data pipelines, networking, storage, model hosting, governance, and latency control. In a bank, that plumbing matters more than in a lot of other industries because slow or unreliable systems are not just annoying — they can hit trading, payments, fraud checks, and client workflows. JPMorgan’s own technology materials have spent years emphasizing infrastructure modernization and software delivery efficiency, so AI is landing on top of that existing push. (jpmorganchase.com) ### Where does Jeremy Barnum fit in? Barnum matters because he is the CFO. When a head of AI says the firm needs more compute, that is interesting. When the CFO starts talking as if AI belongs with unavoidable platform costs, that is capital-allocation policy. Barnum’s job covers the firm’s financial management, treasury, investor relations, and business management, so his framing is the one that tells the organization how to treat the spend. (jpmorganchase.com) ### What does this unlock inside the bank? It makes less glamorous work easier to fund. Not just new models, but host tuning, network upgrades, specialized accelerators, platform renewals, and the boring reliability work that turns demos into production systems. Think of it like moving from buying fancy kitchen gadgets to rewiring the restaurant. Once the wiring is the priority, the gadgets are not the main story anymore. ### Why should other banks care? (jpmorganchase.com) Because JPMorgan is usually early on operating model signals. If the biggest U.S. bank starts treating AI as table stakes, peers get pressured to explain why they are still booking it like an experiment. That does not mean everyone copies the exact budget line tomorrow. But it does mean the burden of proof flips. The question is no longer “why spend?” It is “how are you going to compete if you don’t?” ### Bottom line? JPMorgan’s message is simple — AI is becoming part of the bank’s plumbing. And once something becomes plumbing, it stops being optional. (jpmorganchase.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.