Apollo's Slok warns slower AI payoff poses 3 broad economic risks

Apollo Global Management's chief economist warned that a delayed return on artificial intelligence investments could trigger a recession and a broad market selloff, as AI has been the primary driver propping up the economy and global markets.

Wall Street consensus expects hyperscalers' free cash flow to more than double in the coming years, fueled by surging AI demand. But Apollo's Torsten Slok cautioned that declining token prices and surging Chinese AI competition threaten those projections, creating risks that extend well beyond the technology sector.

"If these trends persist, expected cash flow gains may prove overly optimistic," Slok, chief economist at Apollo Global Management, said in a note published Thursday.

Hyperscalers could see earnings and free cash flow disappoint as heavy capital expenditures and depreciation charges materialize on schedule, compressing margins. With the Magnificent 7 stocks — Apple Inc., Amazon.com Inc., Alphabet Inc., Meta Platforms Inc., Microsoft Corp., Nvidia Corp. and Tesla Inc. — underpinning much of the recent market gains and index performance, a repricing could trigger a broader sell-off extending to semiconductors, power infrastructure, data centers and the S&P 500.

A strain on balance sheets may force greater reliance on debt, elevating leverage and credit risks, Slok said. As AI has become a key pillar supporting both economic growth and market sentiment, any significant slowdown in payoffs risks tipping the broader economy toward recession and sending equities into correction territory.

Chinese AI models have rapidly expanded their footprint, now leading U.S. counterparts in token usage among the world's top 20 models and capturing growing shares of global adoption, Slok noted. Token prices have continued to decline, adding pressure on monetization timelines for companies that have committed tens of billions of dollars to infrastructure buildouts.

The warning comes as Apollo itself is deeply embedded in AI infrastructure financing. The firm last month led a $35 billion capital solution as part of Broadcom Inc.'s new AI XPV Platform, in partnership with Blackstone Inc. and leading global banks. The platform is designed to enable over 20 gigawatts in compute capacity for leading frontier AI labs through 2028, with the initial transaction facilitating Anthropic's previously announced capacity expansion of more than 1 gigawatt of compute infrastructure starting in mid-2026.

For investors, the risk is threefold. If AI monetization takes longer than expected, the heavy CapEx cycle — already reflected in hyperscaler balance sheets — will hit margins before revenue materializes. Because the Magnificent 7 have driven a disproportionate share of S&P 500 returns, any earnings disappointment among those names could cascade through index-linked portfolios and sector-adjacent industries. And a broader economic slowdown would compound the pressure, as rising leverage and credit risks feed back into corporate balance sheets.

The last comparable period of infrastructure-led exuberance was the dot-com era, when telecommunications companies spent heavily on fiber-optic networks only to see capacity outstrip demand. While the current AI investment cycle is backed by real revenue growth from cloud computing and enterprise adoption, the gap between CapEx and monetization has widened as spending accelerates. The Magnificent 7 collectively allocated more than $250 billion to capital expenditures in the past year, with the majority directed toward AI infrastructure.

S&P 500 funds with significant exposure to the Magnificent 7 include the SPDR S&P 500 ETF Trust, the Vanguard 500 Index Fund and the iShares Core S&P 500 ETF.

This article is for informational purposes only and does not constitute investment advice.