The AI Capex Gamble: Cloud Debt Risk on the Rise
The Warning Flashing in the Credit Market
If you think the AI revolution is all upside, the debt markets are whispering a different story. In a sharp note this week, Apollo Global Management cut through the hype with a stark warning: the corporate debt of the hyperscale cloud giants—the very players bankrolling the AI boom—is getting riskier.
The smoking gun? Credit default swaps (CDS)—the insurance contracts that protect against a bond issuer defaulting—for bonds issued by these tech titans are getting pricier. Apollo’s chief economist, Torsten Slok, argues this isn't just noise from banks hedging their exposure as new bonds flood the market. No, this is a fundamental repricing of risk.
"What the market is repricing is hyperscaler credit fundamentals, namely a debt-financed AI capex cycle with rising leverage, negative free cash flow and uncertain payback on depreciating assets," Slok wrote.
Translation: The market is starting to sweat the massive, debt-fueled capital expenditure (capex) these companies are throwing at AI infrastructure. They're borrowing more, their free cash flow is deep in the red, and there’s no guarantee these expensive data centers and chips will pay off before they become obsolete.
Follow the Spread: The Gap Tells the Tale
Here's where it gets technical, and telling. If banks were simply hedging new bond issuance, the widening risk premiums would show up in *their* CDS prices. But Apollo’s research shows that’s not happening. Instead, the gap between hyperscaler CDS and bank CDS has blown out to around 60 basis points from virtually zero last October.
That 60-point chasm is a market signal you can't ignore. It implies the credit risk of the hyperscalers—companies like AMZN (Amazon), MSFT (Microsoft), META (Meta), and GOOGL (Alphabet)—is being judged on its own deteriorating merits, separate from the financial system's health.
The Two-Front Storm for AI Investors
This credit market tremor hits as the AI narrative faces pressure from another angle: its own creators. Over the weekend, leaders of frontier large language model (LLM) companies called for a slowdown in AI advancement, citing safety concerns.
Think that’s just altruism? Many on the Street see a savvy business play. A deliberate slowdown, or regulatory intervention, could protect incumbents from hungry startups and shield them from the legal minefield of autonomous agents run amok.
"What I believe they're really shooting for is the Communications Act treatment that protected the social media guys," said Dan Alpert, founding managing partner of Westwood Capital. "They want some legislation that 'regulates' them, but what it really does is absolves them."
He's pointing to Section 230, the liability shield for online platforms. For cloud providers, the financial consequence is clear: any throttling of the AI development pipeline they host could directly hit their growth projections—the very projections used to justify their massive spending spree.
The Bulls' Counter: Too Early to Panic?
Not everyone is hitting the sell button. The tech bull case argues this is simply the cost of building the next digital empire, and margins are already beginning to justify the spend.
"[Hyperscalers] don't really add their capacity in a significant enough way into 2027 and 2028 to know how this is going to turn out," said Paul Meeks, head of technology research at Freedom Capital Markets. "We're starting to see a turn to the positive in their margins, and if we continue to see [this], there will be less concerns."
He has a point. The payoff horizon for this AI infrastructure is long-term. But the bond market is notoriously impatient. It's asking: can these companies service their growing debt load while cash flows are negative? The numbers show the strain:
GOOGL(Alphabet): Forward debt-to-equity of 13%, negative free cash flow of $25.7B.AMZN(Amazon): Debt-to-equity of 23%, negative free cash flow of $30B.META(Meta Platforms): Debt-to-equity of 34%, negative free cash flow of $25.7B.MSFT(Microsoft): Debt-to-equity of 7.34%, *positive* free cash flow of $33.4B.
Notice the outlier? Microsoft’s fortress balance sheet and positive cash flow stand in stark contrast to its peers, which may explain its relative resilience in both equity and debt markets.
What This Means for Your Portfolio
Forget the daily stock price gyrations. The real action for discerning investors is in the credit market. When the smart money in CDS—described by Dean Baker of the Center for Economic and Policy Research as "the most sophisticated investors anywhere"—starts demanding higher premiums to insure tech debt, you need to pay attention.
"They obviously think there is a substantial risk that the AI companies cannot make good on their commitments," Baker noted.
So, what’s the trade here? It’s about differentiation. The market is no longer treating "Big Tech" as a monolithic bloc. It's starting to pick winners and losers based on financial discipline. Companies funding the AI arms race with a clear path to profitability and strong balance sheets will be rewarded. Those burning cash with reckless abandon will see their cost of capital rise, squeezing future growth.
This also reshapes the risk landscape. As Alpert noted, banks today have "fortress balance sheets," making a 2008-style systemic crisis from tech debt unlikely. The risk is concentrated, not contagious. That means potential volatility and repricing for specific companies, not a broad market meltdown.