Oracle Is Down 67%. The $638B Backlog Is Real.

Every now and then, the market hands you a genuinely difficult problem. Not a simple decision disguised as a hard one, but a real fork in the road where the bull case and the bear case are both grounded in verifiable facts. Oracle right now is exactly that kind of situation.

The stock touched $114.75 on July 24, its lowest price since April 2025 and about 67% below its 52-week high of $345.72. The company this crash belongs to posted record fiscal 2026 results: revenue up 17% to $67.4 billion, cloud infrastructure revenue up 93% in the fourth quarter alone to $5.8 billion, operating cash flow up 54% to $32 billion, and a contract backlog that grew 363% in a single year.

Those two facts, the collapsing stock and the accelerating business, are not a contradiction. They are the question.

What the Market Is Afraid Of

The fear is structural, not imaginary. Fiscal 2026 capital expenditures ballooned to $55.7 billion, more than doubling the prior year’s $21.2 billion outlay. Those investments produced a severe cash-flow deficit. Free cash flow for fiscal 2026 turned negative at $23.7 billion, even as operating cash flow rose 54% to $32 billion. The disparity leaves Oracle heavily dependent on external capital markets to keep its infrastructure program on track.

The funding plan makes equity holders uncomfortable. Oracle raised $43 billion in debt financing and $5 billion in equity in the fiscal year just ended, and expects to raise about $40 billion in debt and equity in fiscal 2027, including $20 billion through a previously announced at-the-market share sale program. That ATM facility, sitting above the current stock price, creates a structural headwind every time the stock attempts to recover.

Then there is the credit rating. S&P Global Ratings affirmed Oracle’s BBB credit rating and revised its outlook to negative, citing uncertainty around the magnitude of future capital spending and its impact on credit metrics. The balance-sheet math needs to hold: free cash flow must turn positive, and leverage must stay inside the range that keeps the company comfortably investment grade.

And then there is Larry Ellison himself. Larry Ellison’s personal guarantee ties Oracle’s stock performance directly to the financing of Paramount Skydance’s bid for Warner Bros. Discovery, creating an unusually interconnected financial risk. Ellison made his $40.4 billion guarantee near Oracle’s highs. With the stock down roughly two-thirds from its peak, the trust’s 1.16 billion Oracle shares are worth far less than they were at the time of the guarantee, and Forbes has questioned whether Ellison has enough cash to honor it without selling Oracle stock or borrowing. The governance discount, the haircut investors apply when one person holds about 41.6% of the outstanding shares, is now very much part of the conversation.

The Business Has Not Broken

Strip away the balance sheet fear and the governance concerns, and the underlying demand picture is extraordinary.

Remaining performance obligations ended the year at $638 billion, up 363% year over year and up $85 billion from the prior quarter alone. A year earlier, the figure was about $138 billion. To put that in plain terms: this is not a pipeline of vague sales interest. It is contracted, signed revenue. Oracle’s CFO disclosed on the fiscal Q4 2026 call that the company expects 12% of the $638 billion RPO, roughly $77 billion, to convert to revenue in the next 12 months, with another 34% converting in the following two years, providing the clearest visibility into forward revenue in Oracle’s history.

Most of the RPO increase came from large-scale AI contracts where the customer prepaid Oracle for the purchase of GPUs, or the customer bought and supplied the GPUs to Oracle. The prepaid and customer-supplied hardware portions of those large AI contracts now total $75 billion. This substantially reduces the amount of capital Oracle must raise to build out its AI data centers. That detail matters more than it has been given credit for. The capex burden is large, but a meaningful portion is customer-funded.

Management confirmed full fiscal 2027 total revenue guidance of $90 billion and raised non-GAAP adjusted earnings per share guidance to $8.05. For Q1 fiscal 2027, Oracle guided total revenue growth of 27% to 29% year over year and cloud revenue growth of 58% to 64%. These are not aspirational projections toward an uncertain future. They are management’s translation of an existing backlog into a revenue schedule. Full-year net income under GAAP rose 37% to $17.1 billion. The company is profitable, growing, and collecting more contracted revenue each quarter than it had ever seen the year before.

The Pentagon deal announced last week is worth noting in the right proportion. The agreement carries a base value of $3.31 billion for the first five years and could expand to $6.99 billion if all five option years are exercised. Under the deal, Oracle will provide software for on-premises data centers used across the U.S. military, the Coast Guard, and the intelligence community. Against $67 billion in annual revenue, this is not a market-moving contract. But it adds another durable strand to an already deep relationship with the federal government, and it arrived at the exact moment the stock was setting new 52-week lows.

The Valuation Gap Is Wide

The multiple compression has been violent. Management guided for about $90 billion of revenue this fiscal year and $8.05 in non-GAAP adjusted earnings per share. At the 52-week low close of $114.99, that puts the stock at about 14 times the earnings management says it will produce this year. Against the 52-week high of $345.72, that same guidance implied more than 40 times.

The market has cut what it will pay for Oracle’s earnings by roughly two-thirds in a year while the earnings themselves keep rising. That is a profound compression, and it rarely happens without a reason. The reason here is the balance sheet: investors are discounting the probability that free cash flow turns negative for a sustained period, that debt continues piling up, and that dilution from the ATM program erodes per-share value even as the business grows.

Thirty-six analysts carry Buy ratings on the stock. Seven have Holds. One has a Sell. JPMorgan maintained its Overweight rating and described the risk-reward as positively biased, pointing to diversified growth across cloud applications, database, and AI infrastructure as evidence that the IaaS thesis is not the only engine running. Guggenheim reiterated its Buy with a $400 price target after conversations with Oracle management about data center timelines.

Cheap vs. Broken: The Honest Assessment

Oracle passes The Cheap Investor’s business quality test without difficulty. It has durable competitive advantages in database software, a cloud infrastructure business growing faster than any hyperscaler competitor, pricing power rooted in deep enterprise integration, and an RPO backlog that effectively pre-funds years of revenue. These are not soft claims. They are reflected in 37% GAAP net income growth and 54% operating cash flow growth during the same year the stock fell about 67% from its high to its low.

The financial risk is real, not imagined. The combination of deeply negative free cash flow, rising debt near the investment-grade threshold, planned equity dilution at depressed prices, and a controlling shareholder whose personal guarantee on an unrelated media deal is now worth far less than it was when he made it creates genuine uncertainty about capital allocation. These are not temporary problems that will resolve in a quarter or two.

That said, the operational case has not collapsed. By every measure relevant to cloud investors, Oracle continues to operate profitably and grows rapidly. In FY2026, cloud infrastructure (IaaS) revenue increased 77% to $18.1 billion. The backlog is real, the customers are real, and the conversion schedule is the most specific Oracle has ever disclosed. What the market is pricing today is not a broken business. It is a business that mortgaged its future for a buildout that cannot yet be judged by the cash it returns.

For a patient investor, the key question is whether the fear has traveled further than the facts warrant. The market is now treating Oracle’s AI opportunity as a balance-sheet problem. But with the stock trading at about 14 times forward earnings management has guided to, it is worth asking whether the fear has traveled further than the facts warrant.

The answer is: probably yes on the business, and probably not yet resolved on the governance. The Ellison guarantee, the ATM dilution program, and the credit rating trajectory need to stabilize before the stock can be valued on its fundamentals alone. That does not make Oracle a value trap. It makes it a situation where the margin of safety needs to be earned by watching two or three specific things rather than bought with a single trade.

The Cheap Investor Scorecard

• Business Quality: High. Cloud infrastructure growing 93%, database installed base sticky, enterprise integration deep.
• Financial Strength: Under pressure. Free cash flow was negative in fiscal 2026. Leverage is a central investor concern.
• Valuation: Genuinely cheap on earnings. Roughly 14 times fiscal 2027 guided EPS, a level historically associated with mature, slow-growth businesses, not companies growing cloud revenue at 58-64%.
• Competitive Position: Strong. OCI is growing far faster than larger hyperscaler competitors on reported growth rates.
• Balance Sheet: The primary risk. About $122.3 billion in long-term debt as of May 31, 2026, rising.
• Cash Flow: Operating cash flow record of $32 billion. Free cash flow deeply negative due to capex program.
• Management Execution: Cloud growth metrics are exceptional. Governance questions around Ellison’s external commitments are a real discount factor.
• Catalyst Strength: Q1 fiscal 2027 results will be the first test of whether the roughly $77 billion in near-term RPO converts as disclosed. That is the catalyst to watch.
• Margin of Safety: Meaningful at about 14 times forward earnings if the backlog converts. Thin if the credit situation deteriorates and equity dilution accelerates.
• Long-Term Potential: Very high if the AI infrastructure buildout generates the returns the backlog implies. The $638 billion figure is one of the largest contracted revenue bases in enterprise technology.

What Needs to Be True

Investors waiting for this story to resolve need three things to go right. First, RPO conversion must track the disclosed schedule. Management guided Q1 fiscal 2027 total revenue growth of 27% to 29% year over year and cloud revenue growth of 58% to 64%. If those numbers land, the story shifts from capex panic to backlog confirmation. Second, the Ellison guarantee and the Warner Bros. Discovery deal need to reach a conclusion, one way or the other, because the legal cloud over the controlling shareholder’s personal finances is currently making Oracle’s stock a proxy for a media merger outcome rather than a cloud infrastructure one. Third, the credit rating needs to hold investment grade. A downgrade would mechanically raise Oracle’s borrowing costs at exactly the moment it is trying to raise about $40 billion.

None of those things require Oracle’s cloud business to outperform. They require the business to perform as management has already guided. That is a lower bar than the stock’s current price implies, and that gap is where the potential opportunity lives for investors willing to hold through the noise.

The question is not whether Oracle is a great business. It demonstrably is. The question is whether a great business carrying a heavy balance sheet and a founder whose personal financial commitments are entangled with a contested media deal can still be bought with a meaningful margin of safety at about 14 times forward earnings. Reasonable investors can disagree. The discipline is making sure you know which bet you are actually making before you make it.