Oracle reported a record fiscal year in June. Revenue hit $67.36 billion, up 17% year over year. Its cloud infrastructure business — the OCI division going toe-to-toe with AWS, Azure, and Google Cloud — surged 93% in its fiscal fourth quarter to $5.8 billion. On paper, these are the numbers of a company executing a successful turnaround. So why is the stock trading roughly two-thirds below its September 2025 peak?

Because Oracle is not executing a turnaround. It is executing a bet-the-company gamble on AI infrastructure, and the market is pricing the shares not as a growth story but as a distressed restructuring candidate. The numbers that explain Oracle's stock price are not found on the revenue line. They live on the balance sheet and the cash flow statement, and they are genuinely alarming.

Oracle spent $55.66 billion on capital expenditures in fiscal 2026 — overshooting its own $50 billion target by more than 10%. For the current fiscal year, management has guided for net capex of $70 billion, with gross spending landing between $90 billion and $95 billion. To put that in perspective: Oracle's entire annual revenue is $67 billion. The company is planning to spend more than it earns on data center construction alone, year after year. This is not capital allocation in any conventional sense. It is a forced march.

The funding mechanism for this buildout is leverage. Total debt stands at $167.4 billion, with long-term debt of $122.3 billion — up 43% year over year. Free cash flow is deeply negative at negative $23.7 billion. Debt now stands at nearly four times shareholders' equity. In July, S&P downgraded Oracle's credit rating to BBB-minus, one notch above junk territory, citing the widening gap between the company's spending commitments and its ability to generate cash. Oracle now plans to raise roughly $40 billion more in fiscal 2027 through a combination of at-the-market equity issuance — which will dilute existing shareholders — and additional debt.

On the surface, there is a bull case. Oracle's remaining performance obligations backlog hit $638 billion in the most recent quarter, up an eye-popping 363% year over year. That number suggests an enormous pipeline of committed future revenue. The problem is concentration. Bank of America estimates that more than 50% of that backlog comes from a single customer: OpenAI. The same OpenAI that is burning through cash at a rate that makes Oracle look disciplined by comparison. If that relationship sours, or if OpenAI's own financial situation deteriorates, a very large chunk of that $638 billion vanishes overnight.

Oracle's customer list is otherwise impressive — Nvidia, Meta, AMD, TikTok, and xAI are all running workloads on OCI. But the company's dependency on hyperscaler-style cluster deals means it must keep winning these massive contracts to justify the debt load. The problem is structural: Oracle is investing at a scale that assumes its customers' AI demand will grow uninterrupted, and that it can capture enough of that demand to eventually generate positive free cash flow. If either assumption proves wrong — and both are unproven — the debt becomes a trap that could take years to escape.

The market's skepticism is visible in the valuation. Oracle trades at roughly 15 times forward earnings. Most analysts maintain Buy ratings with an average price target around $248, implying more than 100% upside from current levels. But the short sellers are circling too. Michael Burry — the investor made famous by The Big Short — has taken a short position, betting that the debt-fueled expansion ends badly. When a company spends more than its entire revenue on capex, borrows billions to do it, and relies on a single customer for half its backlog, the risk of permanent capital destruction is real.

This is not obviously a distressed restructuring story. Oracle still generates significant gross profit, its cloud business is growing, and forward earnings multiples are low by historical standards. But the math on the capital structure is unforgiving. Every dollar spent on interest payments is a dollar not available for innovation or shareholder returns. Every share issued in the ATM program dilutes the equity base. And every quarter that passes with negative free cash flow inches the company closer to the credit event that investors fear. At 15 times earnings, the stock is priced for a recovery that the balance sheet may not survive long enough to see.

Disclosure: The Signal holds no position in ORCL. Positions may change. This is not financial advice.