Apollo warns: The credit logic of hyperscale cloud providers hinges on a $2 trillion cash flow "big bet"

Apollo Global Management's chief economist, Torsten Slok, recently published a report warning about the credit outlook for hyperscale cloud providers. He pointed out that the credit narrative of these five tech giantsAlphabet, Meta, Amazon, Microsoft, and Oracleis built on a forecast widely accepted by the market: that their combined operating cash flow will triple from about $600 billion to $2 trillion. Whether this forecast can be delivered, however, is becoming the core question hanging over global credit markets. The hyperscale cloud provider groupAlphabet, Meta, Amazon, Microsoft, and Oraclehas mainly relied on debt financing to support an unprecedented capital expenditure cycle at a time when free cash flow has turned negative. The market accepts this leverage because it expects AI-related revenue to eventually drive a sharp rebound in cash generation. In the past, tech giants relied on abundant cash flow generated by software, advertising, and cloud services to maintain an asset-light image of strong cash flow, low leverage, and high margins, and long enjoyed extremely low financing costs. But competition in AI infrastructure has completely changed this landscape. Data centers, high-end GPU chips, power systems, and network facilities each mean massive capital expenditures, longer payback periods, and rapid depreciation pressure. The expansion of capital expenditures has far outpaced the growth of operating cash flow. The market expects that the combined capital expenditures of Alphabet, Microsoft, Amazon, Meta, and Oracle will reach $769.2 billion in 2026, nearly double that of 2025, and may climb further to about $1 trillion in 2027. Earnings data confirm this pressure: in the second quarter of 2026, Alphabet generated about $39.1 billion in cash flow from operating activities, while capital expenditures reached about $44.9 billion, pushing free cash flow down to about -$5.9 billionthe first time Alphabet has posted negative quarterly free cash flow since going public. Slok believes that if this rebound fails to materialize, the consequences will not be limited to the balance sheets of a few tech companies. Investor interest in the broader AI trade could fade, credit spreads on hyperscaler debt could widen, and companies could be forced to scale back spending plans. Because this spending has become an important contributor to U.S. economic growth, any pullback would ultimately drag on GDP. The report frames this assumption as the only point of consensus sustaining the credit story. It does not predict that it will fail, but it clearly lays out the downside path: weaker cash conversion would first hit valuations and financing costs, then spill over into investment, and ultimately transmit to the real economy. S&P Global Ratings also assumes a 2028 inflection point in its modelaccelerating revenue growth and slower capital expenditure growthbut whether this assumption holds depends on whether AI commercialization can outrun the pace of swelling capital expenditures. For credit investors, the implication is this: hyperscaler bonds are less a traditional investment-grade story than a concentrated bet on the timing and scale of AI monetization.