OROne Pager
Oracle Corporation ORCL
At $146.47 Oracle is priced at 18.2x FY2027 consensus earnings, 57.6% below its high, while the street was raising its revenue estimates.
The whole question is one number: $55.7B of capital spending against $32.0B of operating cash flow, closed with $40.1B of borrowing.
Key data
ORCL · price with moving averages
Source: market data.
The business
Oracle sells three things: the database software that stores corporate records, the applications that run finance and human resources departments, and rented cloud compute. The first two are a mature annuity with decades of switching costs behind them, sold to large enterprises on multi-year licenses and subscriptions. Gross margin is 65.8% and operating margin is 30.8%, which is what a software toll booth looks like.
What changed is the third business. Oracle committed to building enormous AI compute capacity and renting it to a short list of very large customers under multi-year contracts, and it is funding that with debt. Revenue grew 17.4% to $67.4B in the year ended 31 May 2026, capital spending hit $55.7B, and free cash flow was negative $23.7B. The annuity now finances a construction project.
Business read. A 30.8% margin software annuity is underwriting a capital program it cannot yet fund.
Things you might not know
Oracle issued $5.0B of preferred stock in fiscal 2026, an instrument it had not used before, alongside $40.1B of net debt issued. Preferred ranks ahead of common stock on claims, so the cushion behind $156.2B of total debt is thinner than the debt line alone suggests.
Interest coverage is 4.5x, down from the double digits a software balance sheet normally carries, and tangible book value per share is negative $7.79 against a $146.47 share price. Fixed charges and the capital budget now draw on the same $32.0B of operating cash flow, which covered 57% of capital spending last year.
Property and equipment carries at $129.6B against $9.3B of annual depreciation and amortization, a ratio of roughly 14 to 1 by my computation, on assets whose replacement cycle is set by the next chip generation. Useful-life assumptions, not demand, decide how much of that spending reaches reported earnings each year.
Fundamentals
| Measure | QoQ | YoY |
|---|---|---|
| Revenue | not disclosed | +17.4% |
| Operating income | not disclosed | +17.5% |
| GAAP diluted EPS | not disclosed | +34.3% |
| Operating cash flow | not disclosed | +53.8% |
| Capital spending | not disclosed | +162.7% |
Growth is real and it accelerated: 6.0%, 8.4%, then 17.4% across three years, with the operating margin held at 30.8%. The problem sits one statement lower. Operating cash flow grew 53.8% and capital spending grew 162.7%, so free cash flow went from negative $0.4B to negative $23.7B in a single year.
Valuation
| Metric | Company | Peer median |
|---|---|---|
| Forward P/E, FY2027 consensus | 18.2x | not disclosed |
| Trailing P/E, FY2026 GAAP | 25.1x | not disclosed |
| Enterprise value to revenue | 8.1x | not disclosed |
| Gross margin | 65.8% | not disclosed |
| Operating margin | 30.8% | not disclosed |
Peers: Microsoft, SAP, IBM, Salesforce, Alphabet.
The source report carries no peer medians, and vendor comparables were not retrievable in this session, so those cells stay empty rather than invented. On Oracle's own basis the two halves of the page disagree: 18.2x FY2027 consensus earnings and 25.1x trailing GAAP earnings against 8.1x enterprise value to revenue, with a 65.8% gross margin intact and no free cash flow multiple to quote, because free cash flow is negative $23.7B.
Management
| Measure | Record |
|---|---|
| Capital allocation | FY2026: $55.7B capital spending against $32.0B operating cash flow, $40.1B net debt issued, $5.79B dividends paid, $206M of buybacks |
| Diluted shares | 2,766M (FY2023) to 2,915M (FY2026), up 5.4%; FY2026 stock compensation $4.81B against $206M repurchased, about 23 to 1 |
| Insider activity (12mo) | Vice Chairman Jeffrey Henley exercised 400,000 options at $40.93 on 2026-06-24 and sold 359,068 shares the same day at $156.06 to $162.99, about $57M, cutting his direct stake to 40,932 shares; routine award to CLO Stuart Levey in July 2026; no open-market purchases through 2026-07-28 |
Compensation
| Horizon | Goals | Outcome |
|---|---|---|
| Annual cash, FY2025 | Financial performance metrics set under the stockholder-approved Executive Bonus Plan; the specific metric targets are not disclosed | Metrics exceeded, but bonuses reduced to $0 for Ellison, Catz and Henley to preserve cash for AI projects; Catz salary $950,000, total compensation $1,113,417 |
| Long-term, FY2018 PSO grant, performance period ended FY2025 | 17.5M performance options each to Ellison and Catz at $51.13, in seven equal 2.5M tranches, each requiring a stock price and market capitalization goal matched to an operational goal, such as $10B of non-GAAP SaaS revenue | All six market capitalization goals satisfied by FY2024 year end; four operational goals satisfied by FY2025 year end; five of seven tranches vested and two forfeited; no new equity awards to either executive in FY2025 |
The linchpins
Win big if
Capital spending peaks and the contracted backlog converts. Consensus needs $89.6B of FY2027 revenue against $67.4B delivered, which is $22.2B of new revenue in one year, and it needs that without another step up in the $55.7B capital budget. First observable proof is the 8 September print: revenue clearing the $19.1B consensus with the FY2027 capital spending outlook held or trimmed. Confirming signal is FY2027 capital spending guided flat or lower than $55.7B while revenue still grows above 30%, which puts free cash flow break-even inside FY2028 and pays the $5.79B dividend out of the business again.
Surprised down if
Capital spending rises again and the revenue behind it does not. Oracle borrowed $40.1B in a year it paid $5.79B of dividends out of negative $23.7B of free cash flow, so every extension of the buildout lengthens the period when the equity depends on faith rather than cash. First observable proof is a September capital spending raise toward $70B without a matching backlog disclosure. Confirming signal is revenue growth decelerating below 25% year over year, which would mean the concrete is running ahead of the demand, with $124.9B of net debt and negative $7.79 of tangible book underneath it.
Last word
Eighteen times forward earnings, and negative $23.7B of free cash flow: the multiple and the cash flow are describing different companies.
Methodology
Compressed from the Back of Napkin: Oracle (ORCL), which sources financials from Oracle's Form 10-K filed 2026-06-22 for the year ended 2026-05-31, prior annual filings, Forms 4 through 2026-07-28, and vendor consensus as of 2026-08-22 on the 2026-08-21 close. Compensation is from Oracle's DEF 14A filed September 2025 covering fiscal 2025 and the fiscal 2018 DEF 14A for the original PSO grant terms. Enterprise value to revenue, trailing P/E, the property to depreciation ratio, the operating cash flow coverage of capital spending, and all growth percentages are my computations from as-filed inputs. Quarterly comparisons are marked not disclosed because the source report is annual basis; peer medians are marked not disclosed because neither the source nor official materials provide them. Not investment advice.
Bid Cap
Daily ideas, a 390-name database, and a model long/short book from an investor who mostly covers financials. $70 a month or $700 a year.
Subscribe on Substack

