Table of Contents
JPMorgan said its forecast expects 122 GW of new data center capacity coming online between 2026 and 2030
In sum – what to know:
$250bn annual R&D commitment underscores AI’s scale – Hyperscalers collectively spend a quarter-trillion dollars per year on R&D, highlighting the structural shift toward long-term AI investment.
$5tn build-out reshapes capital markets – JPMorgan expects every major funding pool—public debt, private credit, securitization, government—to be needed to meet AI and data center demand through 2030.
Power constraints threaten AI expansion – Lengthy turbine and nuclear lead times make adding 150 GW of new power difficult, slowing deployment despite “astronomical” compute demand.
U.S. bank JPMorgan Chase has published an analysis projecting that more than $5 trillion will be required over the next five years to build global data center capacity, AI infrastructure, and the associated power systems needed to run them.
In its analysis, the bank warns that such a scale of spending will reshape capital markets and test the limits of today’s energy and grid planning.
The report says that meeting this level of investment “will likely require participation from every public capital market as well as private credit, alternative capital providers and even government involvement.”
JPMorgan’s forecast expects 122 GW of new data center capacity coming online between 2026 and 2030. “The scale of demand for compute remains astronomical,” the analysis said. But the financial institution notes that the build-out is already constrained by limited access to power.
“Current lead times for new natural gas turbines have ballooned to three/four years, and nuclear plants have historically taken more than 10 years to build,” the report said. “Adding 150 GW of power in a timely manner is a remarkable challenge, particularly in light of grid upgrade requirements.”
JPMorgan also expects the data center wave to reverse the slowdown seen in public bond and syndicated loan markets since the pandemic. According to the report, “annual data center funding needs in 2026 are on the order of $700 billion, which could be entirely financed by hyperscaler cash flow and high grade bond markets. However, 2030 funding needs are in excess of $1.4 trillion, which will likely require funding contributions from all capital providing markets.”
The U.S. bank also highlighted that hyperscalers are generating over $700 billion of operating cash flow per annum, and reinvesting nearly $500 billion of that back into capex.
The report also raises caution about parallels with previous tech build-outs. JPMorgan notes the possibility of a “repeat of the telecom and fiber build-out experiences,” with uneven returns and the potential for a sharp correction if AI revenue growth fails to match expectations.
The firm wrote: “Big picture, to drive a 10% return on our modeled AI investments through 2030 would require nerly $650 billion of annual revenue into perpetuity, which is an astonishingly large number.”
Even if revenue materializes for some players in the AI business, JPMorgan cautions that there will be “spectacular losers as well, given the amount of capital involved and the winner-takes-all nature of portions of the AI ecosystem.”
