Apollo Global Management’s president warned that financing the artificial intelligence buildout by major cloud providers will require capital from every corner of the markets — and that the debt markets appear ready to supply it. Jim Zelter told Yahoo Finance at the Milken Institute conference Monday that debt funding for AI-related capital expenditures should remain healthy “at least until 2028,” and that net origination of investment-grade debt this year will be “north of $1 trillion.”
Zelter framed the AI funding picture as a multi-legged stool. Of the roughly $800 billion in projected hyperscaler capex tied to AI infrastructure, he said a portion will come from operating cash flow, “a few hundred billion” from public investment‑grade markets, and private investment‑grade capital will be “the third leg” that helps close the gap. “It’s going to take all the markets, it’s going to take the equity market, operating cash flow, the public [investment‑grade] market, and the private [investment‑grade] market for those kinds of numbers to come through,” he said.
Market forecasts have shifted aggressively higher in recent months. Goldman Sachs strategists noted that consensus estimates for hyperscaler capex in 2026 have climbed by nearly $80 billion since the start of the earnings season. The five largest hyperscalers — Amazon, Google (Alphabet), Meta, Microsoft and Oracle — are now expected to spend about $751 billion in capex in 2026, up 83% from 2025 and substantially above earlier projections of $673 billion at the start of the earnings season and $546 billion at the beginning of the year.
Related Recipes
Ed Yardeni Becomes Central Figure in 2026 S&P 500 Earnings Surge
Middle East Inflation Pushes 2027 Social Security COLA to 3.9%, TSCL Forecast
Gold slides after hotter US PPI cements hawkish Fed outlook; silver climbs on China demand and India’s tariff move
Micron Pushes Chip Rally Higher as AI Demand and Memory Shortages Persist
That jump helps explain why large technology companies are increasingly turning to longer‑dated debt rather than relying solely on internal cash flow or equity to finance sprawling data‑center and networking projects. Recent transactions underline the shift: Oracle unveiled plans in February to raise $45 billion to $50 billion through a mix of a one‑time bond issuance and an “at‑the‑market” equity program to expand cloud infrastructure for clients including Nvidia and OpenAI. Meta followed in May with a $25 billion, six‑tranche bond sale aimed at financing a data center in Louisiana and broader AI expansion, while Alphabet issued $31.1 billion in senior unsecured notes in the first quarter.
Zelter argued the willingness of debt investors to back such large offerings at reasonable rates is a positive sign, noting that a jittery fixed‑income market would be the last thing the sector needs while equity investors chafe at heavy capex. Equity markets have shown signs of impatience as some hyperscalers boost AI spending even after posting strong quarterly results, but so far debt buyers have not been spooked.
Trending Now
The evolving financing mix represents a fundamental change in how Big Tech will underwrite the next wave of infrastructure. Where past cycles relied more on cash generation and share issuance, the coming years look set to be dominated by a blend of cash flow, public debt, and a rising role for private investment‑grade capital — potentially extending the runway for even more ambitious AI deployments through 2028 and beyond.
Continue Reading
