Photo via Fast Company
Oracle has become the poster child for the hidden risks embedded in the AI infrastructure boom. Since September 2025, when the company signed its landmark $300 billion contract with OpenAI, its stock has swung wildly—first soaring 36% in a single day, then losing 43% of those gains. What's most alarming to financial analysts is Oracle's balance sheet: the company now carries over $160 billion in liabilities, with $133 billion directly tied to AI data center buildout, while holding less than $40 billion in cash. This aggressive capital structure puts Oracle in a precarious position compared to peers like Meta and Amazon, whose debt-to-equity ratios hover around 80%, while Oracle's sits at an eye-watering 415%.
The core problem is Oracle's overdependence on a single customer. More than $300 billion of the company's $553 billion in remaining performance obligations come from OpenAI—a firm that is itself reportedly losing billions annually. This makes Oracle, for many investors, a proxy bet on OpenAI's unproven path to profitability rather than an independent operating company. A class-action lawsuit filed by the Ohio Carpenters' Pension Fund alleges Oracle misled bondholders about future financing needs, claiming in a September 2025 bond sale that significant additional borrowing might not be necessary, then requesting $38 billion in additional loans just seven weeks later.
What's particularly concerning for Charlotte's financial and business community is the broader private credit market enabling this risk. Over $3 trillion in private lending now circulates through the economy, with $450 billion specifically flowing into tech sector deals. When major private lenders like Blue Owl Capital withdrew from a $10 billion Oracle data center financing in Michigan, it signaled something deeper: even sophisticated financial engineers were seeing downside risks too large to manage. This retreat by private credit specialists—firms designed to structure deals that shield them from risk—raises the question of whether the entire AI infrastructure financing model has become unstable.
The cascading implications extend beyond Oracle to every company betting heavily on AI. Microsoft and Meta have similarly massive commitments to AI buildout, while firms like CoreWeave are collateralizing chips that become obsolete in years. If AI demand has been overestimated—a concern that even OpenAI's leadership has acknowledged—the value of these data centers could collapse, wiping out investors from pension funds to insurance companies. For Charlotte-area investors and corporate finance officers, Oracle's struggles underscore a critical lesson: in today's boom-or-bust tech landscape, balance sheet strength and diversified revenue streams remain the surest insurance against spectacular losses.
