The Magnificent 7 tech stocks are listed on the Nasdaq.
Adam Jeffery | CNBC
The race to build artificial intelligence infrastructure at an annual clip of billions of dollars is destroying free cash flow and increasing balance sheet risk in so-called hyperscales, warns Moody’s ratings.
In a research note released this week, Moody’s said rising costs are straining even the world’s richest companies. Alphabets again Microsoft relying heavily on debt, stock sales and off-balance-sheet moves to fund their AI ambitions.
“In the past, these companies have relied on asset-light structures focused on software, intellectual property, and risky cloud services that require little capital investment,” Moody’s said in a Wednesday paper. “The transition from asset-light to heavy-duty models requires unprecedented levels of investment and leverage.”
Measures that “threaten credit quality” at six companies tracked by Moody’s, including Microsoft, Amazonalphabet, Meta, The Oracle again CoreWeave, according to the report.
The firm projects that capital expenditures — or capex, which is investment in physical assets like data centers — will reach $785 billion by 2026 before reaching nearly $1 trillion next year.
This change breaks the Silicon Valley formula that took decades to create the most valuable companies in the world. The software is less expensive to replicate, producing fatter profit margins and declining balance sheets. Generative AI, by contrast, requires a large footprint: warehouses crowded with expensive and power-hungry servers and chips.
To fund this expansion, tech giants are increasingly turning to Wall Street, leading to growing profits in the financial industry.
Direct debt across all six hyperscalers has reached nearly $460 billion, according to Moody’s. Tech companies are also tapping the public markets to raise money, including Google-parent Alphabet, which last month announced an $85 billion equity sale.
Leasing of data centers
The rating firm noted that because AI hardware and infrastructure require large upfront investments while revenue is seen over the long term, free cash flow across the sector is under pressure.
To keep the debt straight on their balance sheets, hyperscalers rely on cheap financing, particularly through long-term data center leases, the report explains.
Moody’s said the group’s lease obligations totaled $1.2 trillion. More than $820 billion of that amount comes from leases that have not yet begun, meaning data centers are still under construction.
Although these bonds do not appear as regular debt, Moody’s says they consider them debt-equity bonds that will bind companies to large rent payments.
Despite the warning, Moody’s noted that Microsoft, Alphabet, Amazon and Meta remain among the strongest corporate balance sheets in the world, making it unlikely that investment grade ratings are under imminent threat.
The immediate pressure is on low-cost businesses such as Oracle and specialist AI cloud provider CoreWeave. Oracle maintains a negative Baa2 rating, putting it just two notches above junk status.
Meanwhile, CoreWeave operates within a high-yield market with a Ba3 rating, relying on complex private debt structures to finance its GPU hardware fleet.
A circular ecosystem
Moody’s also pointed to structural volatility within the AI boom. Some of the billion-dollar backlog reported by hyperscalers comes from strategic deals with pre-IPO artificial intelligence labs including OpenAI and Anthropic, Moody’s noted.
Firms that have invested billions in AI labs are, in turn, spending heavily on cloud computing from those same companies, creating what Moody described as a circular AI ecosystem.
Overlapping relationships increase risk because many of the industry’s largest companies increasingly rely on the same AI customers and the same assumptions about future demand, Moody’s said.
However, tech giants have the potential to help offset those risks.
Demand for AI computing remains strong, cloud businesses continue to grow and hyperscalers have signed hundreds of billions of dollars in long-term customer contracts that should provide predictable revenue. Those agreements support strong credit profiles in the industry, even amid rising costs.
However, investors should realize that the financial profile of the technology industry is undergoing a structural change unlike anything seen in the cloud era, according to Moody’s.
“Investors will focus more on the ability of these companies to earn an adequate return on investment,” the rating firm said.
