Torsten Slok, chief economist at Apollo Global Management, has issued a stark warning about the credit outlook for the world's largest hyperscale cloud providers. The credit narrative for these technology titans—Alphabet, Meta, Amazon, Microsoft, and Oracle—is now built entirely on a widely accepted market assumption: that their combined operating cash flow will roughly triple from about $600 billion to $2 trillion. Whether that prediction actually materializes is fast becoming the single most critical question hanging over global credit markets.

The core tension

The hyperscaler group has funded an unprecedented capital expenditure boom primarily through debt at a time when their free cash flow has already turned negative. The market has tolerated this rising leverage because it expects AI-driven revenue to eventually spark a dramatic recovery in cash generation. Historically, these tech giants enjoyed ultra-low financing costs by maintaining a pristine profile of strong cash flows, low leverage, and high margins, powered by software, advertising, and cloud services. The competitive race to build AI infrastructure has now completely upended that model.

The magnitude of the spending surge

Data centers, premium GPU chips, electrical power systems, and network equipment each bring massive capital outlays, extended payback periods, and rapid depreciation pressure. The growth of capital spending is now vastly outpacing the growth in operational cash flow. Consensus projections show that Alphabet, Microsoft, Amazon, Meta, and Oracle will collectively spend $769.2 billion on capital expenditures in 2026—nearly double the anticipated 2025 level—with that figure potentially rising to roughly $1 trillion by 2027.

Recent earnings data highlight the strain: in the second quarter of 2026, Alphabet generated approximately $39.1 billion in operating cash flow, yet its capital expenditures reached roughly $44.9 billion, pushing free cash flow to negative $5.9 billion. That marks the first time Alphabet has ever posted a negative quarterly free cash flow since becoming a public company.

The downside scenario

Slok argues that if the anticipated cash flow rebound fails to arrive, the consequences will not be confined to a handful of tech balance sheets. Investor enthusiasm for the broader AI trade could fade, credit spreads on hyperscaler debt could widen significantly, and companies might be forced to scale back their spending plans. Since these outlays have become a major contributor to US economic growth, any pullback would ultimately drag down the country's GDP.

Slok frames this assumption as the single consensus pillar holding up the entire credit story. He does not predict it will fail, but he explicitly maps out the path of downside risk: weaker-than-expected cash conversion would first hit valuations and funding costs, then curb investment, and finally transmit shockwaves through the broader real economy. Even S&P Global Ratings has built a projected inflection point into its models for 2028—featuring accelerating revenue growth alongside decelerating capital expenditure—but that scenario only holds if AI commercialization can outpace the relentless expansion of capital spending.

For credit investors, the implication is clear: hyperscaler bonds no longer resemble a traditional investment-grade narrative. They are, in essence, a concentrated wager on the timing and scale of AI monetization.