AI Infrastructure Spending Set to Top $1.3T in 2027
The world’s largest technology companies are preparing to spend more than $1.3 trillion on artificial intelligence infrastructure in 2027, creating an unprecedented financing requirement that could test corporate balance sheets and the credit market’s ability to absorb new debt, according to S&P Global Ratings.
The forecast covers Alphabet, Amazon, Microsoft, Meta, Oracle and SpaceX, six hyperscalers investing heavily in data centers, computing equipment, networking and power infrastructure. Their projected 2027 spending represents an increase of approximately 49% from the $870 billion S&P expects in 2026.
Strong demand for AI computing capacity is supported by the companies’ internal workloads and multiyear customer contracts, S&P said. The hyperscalers are pursuing several revenue models, including embedding AI in their existing products, hosting third-party models, charging customers by the token and leasing computing capacity to AI laboratories.
The spending, however, is advancing faster than operating cash generation. S&P expects all six companies to produce negative free operating cash flow in both 2026 and 2027, with recovery not projected until 2029.
The rating agency generally assumes 2028 will mark an inflection point, when AI-related revenue accelerates and capital expenditure growth begins to moderate. That outlook depends on demand remaining durable and the companies successfully converting computing investment into revenue.
“As AI infrastructure investment accelerates, the focus is expanding beyond the scale of spending to the funding models, financial commitments and long-term implications that accompany it,” said Naveen Sarma, managing director and sector lead at S&P Global Ratings.
Hyperscalers are increasingly supplementing operating cash flow with corporate bonds, equity issuance, leases, joint ventures and special-purpose vehicles. Some arrangements include residual value guarantees, under which a technology company may agree to cover part of an asset’s value if it falls below a specified level.
Those structures can reduce the amount of debt reported directly on a company’s balance sheet, but they do not necessarily eliminate economic risk. Long-term leases, purchase obligations and guarantees may function like debt and could constrain financial flexibility if AI demand disappoints.
Corporate bond issuance by U.S. technology companies tied to AI investment reached roughly $220 billion in 2026, compared with $12.5 billion in 2025. Technology-sector bond spreads widened to about 89 basis points, nine basis points above the broader investment-grade market, signaling investors are beginning to demand more compensation for the expanding supply.
The financing wave is also reaching private credit and structured markets. Infrastructure managers, insurers and private lenders are funding data centers, power generation, chips and computing equipment through project loans, asset-backed securities and specialized joint ventures.
That broader funding base helps distribute risk, but it also makes the ultimate allocation of obligations more difficult to assess. S&P said it will monitor contractual commitments alongside traditional leverage when evaluating credit quality.
The principal risk is overcapacity. Hyperscalers are building ahead of expected demand, often making commitments based on assumptions about model adoption, computing intensity and future pricing. Faster technological obsolescence or slower monetization could leave companies with underused facilities and long-dated financing costs.
The more favorable outcome is that AI revenue scales rapidly enough to support the investment while improving productivity across other industries. S&P’s base case leans toward eventual monetization, but the timing remains crucial.
The S&P report did not result in any rating actions. Its central message is that strong current credit quality does not remove the need to scrutinize how the AI buildout is financed, particularly as borrowing and contractual commitments replace internally generated cash.