JPMorgan leans into fee growth and 'Agentic AI'

JPMorgan’s new balancing act is simple: keep net interest income near $104.5 billion in 2026 while building more of the business around fees that do not depend on interest rates staying high. The bank raised that 2026 net interest income target in its February 23, 2026 company update, but management also pointed investors back to long-term growth and expense discipline. (jpmorganchase.com) That shift makes sense because the easy money from higher rates was never guaranteed to last. Bloomberg reported that JPMorgan’s $104.5 billion target included a boost from trading, while the bank still expected about $95 billion of net interest income outside markets, which is a clue that the core lending engine is no longer doing all the work by itself. (bloomberg.com) JPMorgan already looks less like a plain lender than many people assume. In 2024, the bank showed $173 billion of revenue excluding the Visa gain, split across consumer banking, commercial and investment banking, and asset and wealth management, with revenue coming from both net interest income and noninterest revenue. (jpmorganchase.com) By 2025, the numbers leaned even harder in that direction. JPMorgan reported $182.4 billion of total net revenue, $95.6 billion of noninterest expense, and said noninterest revenue rose to $87.0 billion, helped by higher asset management fees, higher payments fees, and higher investment banking fees. (jpmorganchase.com) Payments is one of the clearest examples of why fees matter. JPMorgan told investors in February 2026 that it processes about $12 trillion of payments a day, and in fourth-quarter 2025 it said payments revenue hit a record $5.1 billion on deposit and fee growth. (jpmorganchase.com 1) (jpmorganchase.com 2) This is where the artificial intelligence story stops sounding like a science project and starts sounding like a bank budget. JPMorgan’s own payments team said artificial intelligence was already being used for payment validation screening, queue management, and client cash-flow insights, with account-validation rejection rates cut by 15% to 20%. (jpmorgan.com) The phrase “agentic artificial intelligence” basically means software that can carry out a chain of tasks instead of just answering one prompt. JPMorgan’s artificial intelligence research group says it is working on “AI agents” for multi-step tasks that involve computer actions, asking for help, and learning from human input. (jpmorganchase.com) Inside the bank, that push is already large enough to measure. JPMorgan said its internal large language model platform, called LLM Suite, was rolled out in summer 2024 and reached 200,000 onboarded users within eight months, giving employees a controlled way to draft content and build new workflow tools without exposing customer data. (jpmorganchase.com) Management has been explicit that this is not a blank check for more spending. At 2025 Investor Day, JPMorgan said its core philosophy was to keep investing through the cycle but “constantly” increase efficiency, and at the 2026 company update Jeremy Barnum said the bank’s ability to find “a broad range of efficiency opportunities” had been critical to delivering growth and profitability together. (jpmorganchase.com 1) (jpmorganchase.com 2) So the bet is not that artificial intelligence will replace banking. The bet is that a bank with 318,512 employees, $4.8 trillion in assets under management, and $12 trillion in daily payment flows can use software to shave costs in the back office and make fee businesses like payments, wealth, and investment banking more productive at the same time. (jpmorganchase.com 1) (jpmorganchase.com 2) That is why JPMorgan’s 2026 story is not really about whether $104.5 billion in net interest income holds for one more year. It is about building a bank where fees and automation do more of the heavy lifting before the interest-rate tailwind fades. (jpmorganchase.com) (jpmorganchase.com)