The $5.5 Trillion Bet: Inside JPMorgan's Massive AI Spending Forecast and Debt-Fueled Future

Wall Street’s largest banks are rewriting the rules of AI infrastructure finance. JPMorgan has just lifted its total capital expenditure forecast for the sector, not because of new optimism about a killer app, but because the sheer physical cost—data centers, power supplies, chips—is accelerating past even the largest tech companies' ability to pay with cash. The result is a debt cycle unlike anything the bond market has seen, and Morgan Stanley’s latest numbers show the pace is quadrupling.

The $5.5 Trillion Buildout

JPMorgan now projects that total global AI and data center capital expenditure will reach $5.5 trillion through 2030, up from its previous estimate of $5.1 trillion . The bank’s analysts believe the final figure could climb as high as $7 trillion . For context, the bank also expects 122 gigawatts of new data center capacity to be deployed between 2026 and 2030 .

This is not merely an extension of existing trends. The hyperscalers—Alphabet, Amazon, Meta, Microsoft, and Oracle—have already committed roughly $969 billion to data centers and chips, and their combined capital spending is on track to reach $805 billion in 2026 before crossing $1.1 trillion in 2027 .

The $4.1 Trillion Debt Machine

JPMorgan’s most striking revision is on the financing side. The bank forecasts that AI-related debt financing will total $4.1 trillion through 2030 . The buildout now costs more than the hyperscalers generate in cash flow, forcing them into the bond market at an unprecedented scale .

The debt will be sourced from every corner of the capital markets :

• Investment-grade bonds: Approximately $1.5 trillion over the next five years .
• Leveraged finance and private credit: Roughly $150 billion .
• Securitization of data-center assets: Up to $40 billion annually .

Even with these sources, JPMorgan identifies a significant shortfall of approximately $1.4 trillion that will require private credit and potentially government funding . For 2026 alone, the bank projects a record $1.81 trillion in total US investment-grade bond issuance, eclipsing the previous record of $1.76 trillion set in 2020 . AI-related capital spending is the primary driver, alongside $1 trillion in maturing debt that needs refinancing and a revival in M&A activity .

Morgan Stanley Tracks a Fourfold Surge in 2026

The scale of this shift is already measurable. Morgan Stanley estimates that global AI-related debt issuance reached nearly $236 billion in just the first five months of 2026—a fourfold increase over the same period in 2025 . For the full year, the bank forecasts AI-linked debt issuance will reach approximately $570 billion, more than doubling the total raised in 2025 .

By October 2025, AI-linked debt had already surpassed US banks as the largest segment in the investment-grade market, representing 14% of the J.P. Morgan US Liquid index . For the ordinary bond investors holding index and target-date funds inside 401(k) accounts, the AI buildout is no longer just a technology story—it is becoming the single largest position in their fixed-income portfolios .

The $650 Billion Revenue Hurdle

The uncomfortable question hanging over this debt cycle is whether the revenue will ever arrive. JPMorgan analysts have modeled the required return on the projected investment and concluded that the AI industry must generate approximately $650 billion in annual revenue in perpetuity to clear a modest 10% internal rate of return .

The bank translates that figure into consumer terms: it is equivalent to 58 basis points of global GDP, roughly $34.72 per month from every active iPhone user, or $180 per month from every Netflix subscriber, every year, indefinitely .

The analysis does not predict failure, but it stakes out the threshold. JPMorgan’s own asset management team has argued that using bond markets to fund long-term AI capex is a rational decision rather than a signal of financial strain, noting that it allows companies to match long-duration assets with long-duration liabilities . Morgan Stanley’s credit analysts concur, describing the supply expansion as orderly and primarily driven by structural demand for compute .

The numbers now carry the debate. The capital is committed, the bonds are being sold, and the revenue requirement is no longer a theoretical exercise—it is the benchmark against which every AI subscription, enterprise license, and advertising dollar will be measured through the end of the decade.