Oracle Corporation reported strong second-quarter fiscal 2026 earnings after the market closed on Dec. 10, driven by heavy demand for AI-related services. Yet the company’s stock tumbled as investors increasingly weighed the risks of its rapid, debt-financed expansion to meet AI-driven capacity needs.
The surge in spending on AI infrastructure has created a broader dilemma for markets: weighing the promise of high future payoffs against the risk that these investments may fail to deliver them, especially for firms that rely heavily on borrowing.
Until recently, investors appeared unconcerned, bidding up shares of nearly any company announcing major plans to build data centers, expand cloud capacity, develop computing hardware and software, or increase the power infrastructure needed to support these systems.
Many of these companies trade at high valuations, carry substantial debt, and invest capital at returns that lag the cost of raising it. Oracle is one example.
At the same time, Oracle has continued to add debt, much of it used to fund these projects.
Companies that fail to earn returns above their cost of capital erode rather than create shareholder value.
As of 1:30 p.m. ET on Dec. 11, Oracle’s shares fell 12.46 percent.
Applied Digital, an AI infrastructure developer, illustrates the same pressures.
Warnings About Growing AI Debt
Warnings about such risks are growing. A review published by the Harvard Law School Forum on Corporate Governance notes that companies are rushing into AI investments, including experimental pilot programs, to remain competitive even as the payoff remains uncertain.The review also found that many publicly traded companies have begun formally disclosing “AI risk,” citing potential impacts on reputation, cybersecurity, regulation, supply-chain dependencies, and vulnerabilities in AI-infrastructure.
Russell Rhoads, associate clinical professor of financial management at Indiana University, told The Epoch Times that he is troubled by what Warren Buffett calls the “institutional imperative,” a tendency for corporate leaders to pursue strategies simply because competitors are doing so.
“AI capital spending is outpacing revenues by a ridiculous amount; however, this is based on a belief that AI is going to generate enormous revenues in the future—I cannot connect giant revenue opportunities with AI—I asked ChatGPT and it couldn’t either (the setup for that story is great),” Rhoads said.
Rhoads also points to the concentration of the AI boom, noting that half of Nvidia’s revenue comes from just five companies.
“I am closely watching their capital spending versus what they say publicly,” he said.
Potential Financial Fallout
Sunil Kansal, head of consulting at Shasat, said there is growing agreement among financial professionals and regulators that large-scale AI investment has clear credit-risk implications.“The issue is not AI itself, but the financial structure around it. Organizations are committing significant capital to data infrastructure, cloud capacity, and model development, while the financial benefits often arrive slowly. That timing gap can place pressure on liquidity, leverage, and overall financial resilience,” Kansal told The Epoch Times.
Roman Eloshvili, founder of XData Group, expressed concern about the mounting credit risks linked to the AI spending boom.
“It certainly looks like Wall Street is pumping money into the AI investment frenzy. The amounts going into data centers, chips, and cloud resources are unprecedented, as multi-trillion-dollar spends like the ones we are witnessing are seldom seen,” he told The Epoch Times.
Eloshvili warned that lenders may underestimate the risks.
“Banks don’t seem to see how compressed their expectations are. The ability to monetize AI models is expected to grow at an unprecedented pace, but the technology remains in its infancy and is both untested and expensive to run. Lending banks with exposure to the hyperscalers, chip makers, and tech infrastructure builders would see credit loss if revenues dwindled and growth fails,” he said.
Investors have a few ways to protect their portfolios against such outcomes.
Traditional diversification remains one option, adding assets that are not correlated with AI-linked companies.
Another is purchasing “insurance” in the form of Credit Default Swaps (CDS), in which buyers pay a premium in exchange for protection against a bond issuer’s default or a credit event. Rising CDS premiums can signal growing unease in credit markets.
Banks Buying Protection
Georgios Koimisis, associate professor of finance at Manhattan University, said major banks are playing a complex role in the current cycle.“Big banks are racing to fund a huge wave of borrowing by tech giants, such as Microsoft, Meta, and Oracle, so they can build data centers and, generally, AI infrastructure,” Koimisis told The Epoch Times.
“At the same time, those banks are buying protection in case things go wrong, using credit default swaps and new ‘AI baskets’ of bonds and derivatives that they can sell to hedge funds and private credit firms. On the surface, banks look safer. Much of the risk is simply moving to parts of the financial system that are harder to see.”
Koimisis questioned whether these hedges can truly protect the system.
“The main question is not whether AI turns out to be a bubble, but who will take the hit if expectations prove too optimistic,” he said.
“Even if the big tech companies never come close to default, a sharp drop in AI-related valuations could leave pension funds, insurers, and mutual funds with massive losses on complex credit products.
“In that sense, today’s ‘AI borrowing binge’ looks more like the latest chapter in an old playbook. Basically, it is just another case of borrowing heavily on the promise of future growth, then passing the risk along and hoping it lands with someone else,” Koimisis said.
