JPMorgan Global Research raised its estimate for global AI-related capital expenditures to $5.5 trillion through 2030, up from $5.1 trillion, in its midyear outlook released June 24. The revision reflects accelerating data center buildouts, chip procurement, and supporting infrastructure spending concentrated overwhelmingly in US markets.
The Numbers
The bank now projects hyperscaler capital expenditures will reach $650 billion in 2026 and surpass $1.1 trillion in 2027, according to Fortune. Operating cash flow is expected to exceed $900 billion by 2027, which JPMorgan cites as evidence the cycle remains profitable.
On the debt side, the outlook is equally aggressive. JPMorgan raised its estimate for AI-related debt financing to $4.1 trillion, citing higher loan-to-cost ratios. The bank forecasts high-grade corporate debt will account for more than $2.1 trillion in data center financing over five years, with $150 billion in US hyperscaler debt issuance expected in 2026 alone and another $100 billion equivalent abroad, per Fortune’s detailed breakdown.
An additional $170 billion is expected from data center and chip issuers outside the core high-grade market.
Where the Money Goes
The US accounts for roughly 85% of AI and machine learning venture capital, with spillover benefits reaching China, South Korea, and Taiwan through their roles in semiconductor supply chains. Microsoft alone expects to invest roughly $190 billion in capital expenditures in calendar year 2026, a 61% increase from the previous year, according to Fortune.
The technology companies driving this cycle are also expanding beyond their traditional markets. Qualcomm unveiled a data center strategy at its June 24 Investor Day targeting more than $15 billion in annual data center revenue by fiscal 2029. Memory chipmaker Micron reported a 346% surge in quarterly revenue on June 24, with quarterly profit reaching $28.2 billion.
The Tension
JPMorgan’s bullish revision arrives alongside warnings from the other direction. The Bank for International Settlements flagged in its 2026 annual report that the AI investment boom could trigger credit market disruption comparable to the 2008 financial crisis. High-grade US tech bond credit spreads widened to 0.79% in June, signaling bond investor skepticism about whether enterprise AI adoption will keep pace with infrastructure spending.
Loan-to-cost ratios averaging above 85%, with some exceeding 90%, suggest lenders are pricing in continued appreciation rather than hedging for a correction. JPMorgan’s own framing acknowledges the concentration risk: the investment cycle remains dependent on a small group of technology companies, and any slowdown could have an outsized impact on industry growth.
The central question for builders and operators: does $5.5 trillion in capex produce proportional demand for AI services, or does it produce overcapacity that drives down the unit economics everyone’s business models depend on?