Jamie Dimon says AI spending will reach $1 trillion by 2027

JPMorgan Chase CEO Jamie Dimon expects global AI infrastructure spending to hit $1 trillion next year, a figure that would reshape capital allocation across the technology sector and test the limits of the current build-out cycle.

JPMorgan Chase reported its biggest-ever quarterly profit by a US bank on July 14, posting net income of $16.9 billion and adjusted earnings per share of $7.70 — 38.7% above the $5.55 consensus estimate. Revenue surged 27.7% year-over-year to $57.35 billion, driven by an 86% jump in equities trading revenue and a 30% increase in investment banking fees. The results underscored how the AI infrastructure build-out is filtering through the broader economy, with banks benefiting from heightened capital markets activity tied to technology sector financing.

"AI went from $400 billion last year to $700 billion this year," Dimon said on the earnings call. "People project, which so do our people, it will be like a little over a trillion next year." The CEO's forecast, which covers spending across hyperscaler data centers, GPU procurement, networking equipment, and power infrastructure, positions AI as the dominant driver of corporate capital expenditure for the foreseeable future. He noted that total US CapEx runs at roughly $4 trillion annually, meaning AI-related investment could account for a quarter of all corporate spending by 2027.

The scale of the forecast carries significant implications for the technology supply chain. Nvidia, whose H100 and Blackwell GPUs power the majority of AI training and inference workloads, has seen its data center revenue grow from $15 billion in fiscal 2023 to an expected $100 billion-plus in fiscal 2026. AMD, with its MI300X and upcoming MI400 accelerators, and Broadcom, which designs custom AI chips for hyperscalers, are competing for a share of the build-out. Microsoft, Amazon, Alphabet, and Meta collectively spent more than $200 billion on capital expenditures in the trailing twelve months, with the majority directed toward AI infrastructure.

The $1 trillion question

Whether the spending trajectory is sustainable depends on whether enterprise AI adoption generates returns that justify the investment. Dimon acknowledged the uncertainty, telling analysts the environment is "getting close to as good as it gets" and that "we just don't know how long it's going to last." His caution echoes a broader debate on Wall Street: Goldman Sachs published research in June questioning whether the $1 trillion in AI spending would produce commensurate revenue, while Sequoia Capital has estimated that the AI industry needs to generate $600 billion in annual revenue to justify current infrastructure investment.

JPMorgan itself is investing heavily in the technology. The bank raised its full-year adjusted expense guidance to about $107.5 billion, with a portion directed toward AI tools, and Dimon said the company has "almost 1,000 use cases today" across risk management, fraud detection, marketing, hedging, and document processing. He cautioned, however, that the benefits of AI would ultimately accrue to customers rather than shareholders. "The ultimate beneficiary of AI will be our customers," Dimon said. "In a competitive capitalist world, we always use AI to do a better job for the customers, and we can't just say it's going to increase our margins."

Who wins, who loses

The $1 trillion forecast is most bullish for companies that supply the physical infrastructure of AI. Nvidia, trading at roughly 35x forward earnings, remains the primary beneficiary, though its dominance faces challenges from custom chips designed by hyperscalers and from AMD's growing product lineup. TSMC, which manufactures the advanced chips for Nvidia, AMD, and Broadcom, stands to gain from sustained high utilization of its 3nm and 2nm nodes. On the energy side, utilities and nuclear power providers are seeing unprecedented demand growth as data center power consumption accelerates.

The risk is that the spending cycle peaks before returns materialize. If enterprise AI adoption disappoints or if efficiency gains from model optimization reduce the need for compute, the $1 trillion figure could mark a top rather than a baseline. For now, Dimon's forecast — grounded in the capital allocation decisions of the world's largest companies — carries weight. JPMorgan's own results show a bank firing on all cylinders, and its CEO is betting that AI will keep the engine running.

This article is for informational purposes only and does not constitute investment advice.