
Oracle's stock has crashed nearly 60% in nine months despite strong earnings growth, leaving it trading at a historically cheap valuation. The sell-off reflects fears that Oracle has borrowed too much—nearly $130 billion(約21兆円) in debt—to build data centers for OpenAI, which is hemorrhaging money and recently delayed its IPO. While Oracle's backlog of contracted revenue hit $638 billion(約100兆円) (up 363% year over year), roughly $300 billion(約48兆円) depends on OpenAI's ability to pay, making the stock a risky bet despite its apparent bargain price.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Oracle shares have fallen nearly 60% in nine months despite 21% year-over-year revenue growth to $19.2 billion(約3.1兆円) and 93% growth in its cloud infrastructure division (OCI) to $5.8 billion(約9300億円). The stock's price-to-earnings ratio collapsed from 76 in September to about 22 today, and Wall Street analysts rate 84% buy, yet the decline persists.
Why it matters
Oracle has amassed nearly $130 billion(約21兆円) in debt against $43 billion(約6.9兆円) in stockholders' equity (a 300% debt-to-equity ratio) to fund AI data centers, with free cash flow negative $24.5 billion(約3.9兆円) through 2026. Roughly $300 billion(約48兆円) of Oracle's $638 billion(約100兆円) backlog depends on OpenAI, which posted $21 billion(約3.4兆円) in losses against $13 billion(約2.1兆円) in sales last year and recently delayed its IPO—triggering fears that Oracle's largest customer may not be able to pay.
What to watch
Oracle management plans to raise another $40 billion(約6.4兆円) to $50 billion(約8兆円) this year to continue building capacity. The cost of insuring Oracle's debt against default recently hit record levels, and bonds are trading more like junk debt even though they retain investment-grade status from major rating agencies.
Oracle's nine-month descent from AI darling to apparent bargain basement offers a cautionary lesson in how quickly market narratives can reverse. Nine months ago, Oracle (ORCL) was hotter than Nvidia, Alphabet, or Micron Technology as an AI play. Shares have since fallen nearly 60%, a move that looks counterintuitive given the company's strong operational performance: revenue rose 21% year over year to $19.2 billion(約3.1兆円) in its most recent earnings, and Oracle Cloud Infrastructure (OCI), the division that rents computing power for AI workloads, grew 93% to $5.8 billion(約9300億円). Even more eye-catching is the backlog. Oracle's remaining performance obligations (RPO)—contracted revenue not yet delivered—hit $638 billion(約100兆円) as of last quarter, up 363% from a year ago. A single quarter added $85 billion(約14兆円) to that backlog. Customers include Nvidia, Meta Platforms, and SpaceX alongside OpenAI, the largest of them all. The valuation has become attractive: the stock's price-to-earnings ratio crashed from 76 in September to about 22 today, well below the five-year average of 34. This combination—explosive earnings growth and a cheap multiple—has convinced 84% of Wall Street analysts to rate the stock a buy.
The problem lies in how Oracle is financing its AI build-out and, critically, who is supposed to pay for it. To construct the data center capacity behind those projections, Oracle has borrowed nearly $130 billion(約21兆円) against just $43 billion(約6.9兆円) in stockholders' equity, producing a debt-to-equity ratio of roughly 300%. For context, Amazon carries a 50% ratio and Microsoft a 30%. Capital expenditures hit $55.7 billion(約8.9兆円) in fiscal 2026, leaving free cash flow at negative $24.5 billion(約3.9兆円). Management plans to raise another $40 billion(約6.4兆円) to $50 billion(約8兆円) this year to continue the build. The debt market has grown visibly nervous: the cost of insuring Oracle's debt against default recently hit record levels, and while its bonds technically retain investment-grade status, they are trading more like junk.
The defining vulnerability is this: roughly $300 billion(約48兆円) of that $638 billion(約100兆円) backlog—the economic heart of the bull case—depends on OpenAI. And OpenAI's finances are in crisis. Leaked figures, confirmed by the Financial Times, show the company posted $21 billion(約3.4兆円) in losses last year against $13 billion(約2.1兆円) in sales. OpenAI has committed itself to roughly $1 trillion(約160兆円) in compute expenditures, a figure that is 77 times its current annual revenue. OpenAI recently delayed its IPO, a move that sent Oracle stock lower and triggered its worst week in years. The core risk is stark: if OpenAI cannot raise the capital to fulfill those compute commitments, or if it negotiates its way out of them, Oracle's backlog—and the debt Oracle took on to service it—becomes a cascading liability.
Oracle's stock decline reflects a widening gap between the company's operational strength and the financial risk it has incurred to chase AI growth. Nine months ago, Oracle was among the hottest AI stocks, but the narrative has shifted sharply. The company's most recent earnings showed solid fundamentals: revenue up 21% year over year and OCI up 93%, while its remaining performance obligations (RPO) backlog surged 363% to hit $638 billion(約100兆円), with $85 billion(約14兆円) added in a single quarter alone. This backlog was the catalyst for the stock's September surge and remains the bull case for a cheap re-entry—the P/E ratio has compressed from 76 to 22, now trading well below the five-year average of 34, and 84% of Wall Street analysts still rate the stock a buy.
Yet the bear case has proven more compelling to markets. To deliver on that massive backlog, Oracle has built a balance sheet that looks increasingly precarious: $130 billion(約21兆円) in debt against only $43 billion(約6.9兆円) in equity (a 300% debt-to-equity ratio), free cash flow negative $24.5 billion(約3.9兆円) through fiscal 2026, and a plan to borrow another $40 billion(約6.4兆円) to $50 billion(約8兆円) more this year. The debt market is signaling alarm—the cost to insure Oracle's debt against default hit record levels, and its bonds, while still investment-grade, are trading like junk. The defining vulnerability is that roughly $300 billion(約48兆円) of the backlog (nearly half) depends on a single customer: OpenAI. OpenAI's finances are in severe distress—$21 billion(約3.4兆円) in losses against $13 billion(約2.1兆円) in sales, with compute commitments of roughly $1 trillion(約160兆円) against annual sales of $13 billion(約2.1兆円) (77 times current revenue). OpenAI's delayed IPO has reignited doubts about its ability to fund those commitments, and Oracle's stock suffered its worst week in years in response. The core question investors face is whether Oracle's operational growth and cheap valuation offset the concentration risk and debt burden tied to a financially unstable anchor customer.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
No comments yet. Be the first to share your thoughts!
Log in to join the discussion





Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime