A lengthy analysis by JPMorgan Chase & Co. said that more than $5 trillion is set to be spent on global data center and AI infrastructure and related power supplies over the next five years.

That investment "will likely require participation from every public capital market as well as private credit, alternative capital providers and even government involvement," JPMorgan said.

According to the company's base case analysis, 122GW of data center capacity is expected to be built from 2026-30, as "the scale of demand for compute remains astronomical." The number of data centers coming online would be even higher if it wasn't for power constraints, JPMorgan said.

"Current lead times for new natural gas turbines have ballooned to three/four years, and nuclear plants have historically taken 10+ years to build," the report said. "Adding 150GW of power in a timely manner is a remarkable challenge, particularly in light of grid upgrade requirements."

Funding this data center wave is expected to re-accelerate public bond and syndicated loan market growth after a post-covid dip.

"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," the report said. "However, 2030 funding needs are in excess of $1.4 trillion, which will likely require funding contributions from all capital providing markets."

The report added: "Hyperscaler cash flow will have to do an enormous amount of the heavy lifting. Hyperscalers are generating over $700bn of operating cash flow per annum, and reinvesting ~$500bn of that back into capex.

"Note that ~$500bn of reinvestment back into the collective businesses does not include ~$250bn of research and development spending per year. We are assuming approximately $300bn per year of hyperscaler capex reflects cash flow funded AI and data center capex (while AI is crowding out a lot of spending, it's not crowding out all spending)."

The next five years will see around $150bn come from leveraged finance, and up to $200bn in data center securitizations.

High-grade bond markets could bring in $300bn of AI/data center cash over the next year alone, and $1.5tn over five years.

Even with these different funding sources, JPMorgan believes that there's a $1.4tn funding gap. That's where private credit could pick up the slack, alongside "more aggressive financial support by governments... particularly if/when national defense concerns around AI grow."

However, JPMorgan hedged its analysis of the AI boom, noting the risk of this investment cycle being a "repeat of the telecom and fiber build-out experiences."

The report added: "Big picture, to drive a 10 percent return on our modeled AI investments through 2030 would require ~$650 billion of annual revenue into perpetuity, which is an astonishingly large number. But for context, that equates to 58bp of global GDP, or $34.72/month from every current iPhone user, or $180/month from every Netflix subscriber."

Even if the revenue comes in for some AI companies, there will likely be "spectacular losers as well, given the amount of capital involved and the winner-takes-all nature of portions of the AI ecosystem."