Moody's warns AI spending eroding hyperscaler balance sheets

Moody's Ratings released a research note this week warning that the trillion-dollar annual race to build artificial intelligence infrastructure is eroding free cash flow and increasing balance-sheet risk at so-called hyperscalers—the world's largest cloud and tech companies. The spending surge is forcing even cash-rich corporations like Alphabet and Microsoft to lean heavily on debt, stock sales, and off-balance-sheet financing moves to fund their AI ambitions.

The transition represents a fundamental break from a decades-long Silicon Valley formula. Previously, dominant tech firms relied on asset-light structures centered on software, intellectual property, and scalable cloud services that required modest capital investment and generated fortress balance sheets with fat profit margins. Generative AI demands the opposite: asset-heavy physical infrastructure. Moody's projected that capital expenditures across six tracked hyperscalers—Microsoft, Amazon, Alphabet, Meta, Oracle, and CoreWeave—will hit $785 billion（約130兆円） in 2026 before reaching about $1 trillion（約160兆円） in 2027. This vast spending is required to build warehouses crammed with expensive servers and chips.

To finance the expansion, tech giants are increasingly turning to Wall Street. Direct debt across the six hyperscalers has reached approximately $460 billion（約74兆円）, while companies like Alphabet have launched major equity offerings; Alphabet announced an $85 billion（約14兆円） equity sale last month. However, the most striking financing innovation is off-balance-sheet leasing. Because AI hardware requires massive upfront investment while revenue materializes over a longer horizon, companies are leasing rather than owning data centers to keep direct debt off their balance sheets. Moody's reported that lease commitments across the group have ballooned to $1.2 trillion（約190兆円）, with more than $820 billion（約130兆円） from leases that haven't started yet—meaning those data centers are still being built. While these obligations don't appear as traditional debt, Moody's considers them debt-equivalent liabilities that will bind companies to significant rent payments going forward.

The credit risk, however, is not evenly distributed. Moody's noted that Microsoft, Alphabet, Amazon, and Meta retain among the strongest corporate balance sheets in the world, making it unlikely their investment-grade ratings face imminent threat. The immediate pressure is concentrated on lower-rated entities. Oracle carries a rating of Baa2 with a negative outlook, placing it just two notches above junk status. CoreWeave, a specialized AI cloud provider, operates within the high-yield market with a Ba3 rating and relies on complex private debt structures to finance its GPU hardware fleets.

Moody's also flagged a structural risk in what it called a circular AI ecosystem. Hyperscalers have invested billions into pre-IPO artificial intelligence labs including OpenAI and Anthropic; those labs in turn spend heavily on cloud computing from the same companies, creating mutual dependence. The overlapping relationships heighten risk because many of the industry's biggest companies are increasingly dependent on the same AI customers and the same assumptions about future demand. Still, Moody's acknowledged that hyperscalers have significant strengths: demand for AI computing remains robust, cloud businesses continue to grow, and companies have signed hundreds of billions in long-term customer contracts that should provide predictable revenue. These factors support the industry's largely strong credit profiles even amid the spending boom. Nevertheless, Moody's concluded that the tech industry's financial profile is undergoing a structural change unlike anything seen in the cloud era, and investors should focus on these companies' ability to realize an adequate return on investment.

The AI infrastructure boom marks a historic break from the software-dominated business model that defined Silicon Valley for decades. For thirty years, companies like Alphabet and Microsoft built fortress balance sheets by selling software and cloud services that required modest capital investment—high margins, light balance sheets, strong free cash flow. Generative AI demands the opposite: vast warehouses of expensive, energy-hungry hardware that must be built before revenue materializes. Moody's projects spending to reach about $1 trillion（約160兆円） by 2027, forcing even the world's most cash-rich corporations to turn to debt and equity markets on an unprecedented scale.

The use of off-balance-sheet financing—particularly long-term data center leases totaling $1.2 trillion（約190兆円） across the six tracked firms—reflects how structurally different this model is. These lease obligations do not appear as traditional debt but Moody's treats them as debt-equivalent liabilities that will bind companies to significant rent payments for years. The risk is concentrated on lower-rated entities like Oracle and CoreWeave, which lack the balance-sheet cushion of Microsoft or Alphabet; however, all six companies now operate within a structural circularity that Moody's flagged as a concern. Hyperscalers have invested billions into pre-IPO AI labs like OpenAI and Anthropic, which in turn spend heavily on cloud computing from those same companies, creating mutual dependence on the same customer relationships and demand assumptions.

Despite the warning, Moody's acknowledged that the strongest hyperscalers retain sufficient financial strength to absorb this transition without immediate rating pressure. The real question, as Moody's concluded, is whether these companies can realize an adequate return on investment at this scale—a structural question investors are now forced to assess.