I.Executive Summary
Oracle is turning from a mature database and applications company into one of the largest builders of AI data-center capacity. Remaining performance obligations, meaning contracted revenue not yet recognised, rose from $137.8 billion in May 2025 to $664 billion in August 2026, mostly from multi-year AI compute contracts. In the quarter to 31 August 2026 revenue grew 30% to $19.3 billion and Oracle Cloud Infrastructure (OCI) revenue grew 121%. Management guides fiscal 2027 (to May 2027) to at least $90 billion of revenue and $8.10 of non-GAAP EPS.
The cost is cash. Capital spending was $55.7 billion in fiscal 2026 and is guided to $90–95 billion this year. Free cash flow was −$23.7 billion last year, debt has risen to about $130 billion of notes, and Oracle sold $20 billion of new shares in the August quarter. The shares have halved over the past year, to $137.30, and trade at 16.9× fiscal 2027 consensus EPS.
Oracle is the cheapest of the AI names on earnings and the most stretched on cash. Analysts expect the contracted backlog to turn into 26% annual EPS growth; the balance sheet leaves less room for error than at any time in Oracle's history.
Our illustrative five-year P/E projection holds today's 16.9× P/E on fiscal 2027 consensus EPS and takes the fiscal 2027–fiscal 2028 consensus as Years 1–2. Because growth is backed by $664B of contracted revenue, we treat Oracle as non-cyclical: EPS then grows at analysts' long-term rate rounded down (26% a year in the base case, 3 points lower or higher in the bear and bull cases, with exit multiples 15% lower or higher). That gives Year-5 prices of $293 (bear), $371 (base) and $458 (bull), against today's $137.30: about +22.0% a year in the base case (+16.4% to +27.2%). Discounted at 10% a year, the base case is worth $230 today, +68% versus the price. Capped like the chip companies (EPS held at fiscal 2028 consensus), the base case would be about $186 (+6.2% a year). *All values are Pathos's illustrative estimates, not price targets or recommendations.
II.Business Overview
Cloud services and license support is the core. In Q1 fiscal 2027, cloud revenue (infrastructure plus applications) was $11.6 billion, up 62%. OCI infrastructure was about $7.4 billion, up 121%; cloud applications (SaaS) grew 10%. Software license and support revenue outside the cloud fell 3% to $5.5 billion, and services ($1.4 billion) and hardware ($0.8 billion) make up the rest.
The AI build-out. Oracle builds and leases data centers that it fills with GPUs and rents to AI model developers and other large customers. In the quarter it delivered 850 MW of capacity and more than 300,000 GPUs since the prior quarter, and booked more than $30 billion of new AI contracts. Management says about half of the $664 billion RPO will convert to revenue within 36 months (earnings call).
Why it matters for value. The legacy business (database, applications, support) produced steady 26–31% operating margins and cash that funded buybacks. The infrastructure business needs heavy capital up front, earns on long contracts, and concentrates risk in a few very large customers.
III.Share Price & Valuation Multiples
Price as of NYSE close, 30 Sep 2026 · Fundamentals: FY2026 10-K and Q1 FY2027 Form 10-Q| Metric | Value | Basis |
|---|---|---|
| Share price | $137.30 | Close, 30 Sep 2026 |
| 52-week range | $114.50 – $322.54 | Finviz; shares down 51% over 52 weeks (StockAnalysis) |
| Market capitalisation | $415B | 3,023.7M shares (10-Q cover, 7 Sep 2026) × $137.30 |
| Enterprise value | $548B | StockAnalysis; debt incl. leases $169.1B, cash $37.1B |
| P/E, trailing 12 months (GAAP) | 21.5× | StockAnalysis; TTM GAAP EPS $6.38 |
| P/E, FY2027 consensus | 16.9× | $8.14, 37 analysts (Yahoo Finance) |
| P/E, FY2028 consensus | 12.5× | $11.00, 39 analysts |
| EV / EBITDA (TTM) | 15.9× | TTM EBITDA $34.4B |
| Price / sales (TTM) | 5.8× | TTM revenue $71.8B |
| Price / book | 6.2× | Equity $66.8B at 31 Aug 2026 |
| Dividend | $2.00 a year (1.46%) | $0.50 quarterly; ex-date 9 Oct 2026 |
On earnings, Oracle is the cheapest stock in our coverage. On cash flow it has no positive multiple at all: trailing free cash flow is −$28.7 billion. The FY2028 consensus range is unusually wide, from $9.00 to $13.30, which reflects real uncertainty about how quickly the new capacity becomes profitable. For context only, the mean analyst price target compiled by StockAnalysis is $237.97 (43 analysts).
IV.Five-Year Financial Record
Source: Oracle Forms 10-K; FY2027 = management guidance of $90–95B (Q1 FY2027 earnings call), drawn at the $92.5B midpoint (Pathos). OCF = operating cash flow for the year.
| US$ millions, FY to 31 May | FY2022 | FY2023 | FY2024 | FY2025 | FY2026 |
|---|---|---|---|---|---|
| Revenue | 42,440 | 49,954 | 52,961 | 57,399 | 67,357 |
| Growth | 4.8% | 17.7% | 6.0% | 8.4% | 17.3% |
| Operating income | 10,926 | 13,093 | 15,353 | 17,678 | 20,606 |
| Operating margin | 25.7% | 26.2% | 29.0% | 30.8% | 30.6% |
| Net income | 6,717 | 8,503 | 10,467 | 12,443 | 17,087 |
| Diluted EPS (US$) | 2.41 | 3.07 | 3.71 | 4.34 | 5.83 |
| Operating cash flow | 9,539 | 17,165 | 18,673 | 20,821 | 31,977 |
| Capital expenditure | (4,511) | (8,695) | (6,866) | (21,215) | (55,663) |
| Free cash flow | 5,028 | 8,470 | 11,807 | (394) | (23,686) |
| Remaining performance obligations (US$B) | 46.6 | 67.9 | 97.9 | 137.8 | 638.0 |
Until fiscal 2025 Oracle was a slow-growing, highly profitable company: revenue grew 5–8% a year apart from fiscal 2023, which included the Cerner acquisition, and operating margin widened from 26% to 31%. Since then capital spending has gone vertical, from $6.9 billion in fiscal 2024 to $55.7 billion in fiscal 2026. Revenue growth reaccelerated to 17% last year and 30% in the latest quarter.
Bars: quarterly revenue and net income, US$ billions. Line: trailing P/E = closing share price at quarter-end ÷ trailing-12-month diluted EPS (sum of the last four quarters; fiscal Q4 derived as full-year minus nine-month figures). Both axes share zero; the P/E scale extends below 0× only to make room for net-loss bars (P/E is never plotted below zero). Sources: Form 10-Q/10-K data via SEC XBRL company facts; Yahoo Finance price history. Hover over a quarter for exact values.
Total tangible book value, US$ billions = shareholders’ equity minus goodwill and other intangible assets, at quarter-end. From February 2026 equity includes $5.0bn of preferred stock, which the total includes and the per-share figure in the tooltip deducts. Oracle’s August 2026 10-Q no longer presents intangible assets separately ($3.2bn at 31 May 2026), so only goodwill is deducted for that quarter, which is why the latest total is positive. Source: balance sheets in Forms 10-Q/10-K via SEC XBRL company facts. Hover over a quarter for the components and the value per share.
V.Cash Conversion & Earnings Quality
Accounting earnings still look healthy, with fiscal 2026 GAAP net income of $17.1 billion. Cash does not: free cash flow has been negative for two years and was −$28.7 billion over the trailing twelve months (StockAnalysis). In Q1 fiscal 2027 operating cash flow of $23.1 billion was flattered by customer prepayments, while capex was $28.5 billion.
- Depreciation is coming. Trailing depreciation and amortization was $9.8B, small next to $75.7B of trailing capex. As the GPUs and buildings enter service, depreciation will rise sharply, and short GPU useful lives would weigh on reported margins.
- Prepayments and leases. Some of the build is funded by customer prepayments and by leases (operating lease liabilities of $30.6B and finance leases of $8.4B, non-current, at 31 Aug 2026). These are obligations even when they are not counted as debt.
- Stock compensation was $4.8B in fiscal 2026 (7.1% of revenue), excluded from the non-GAAP EPS that consensus uses.
VI.Balance Sheet Strength
| US$ millions | May 2024 | May 2025 | May 2026 | Aug 2026 |
|---|---|---|---|---|
| Cash and equivalents | 10,454 | 10,786 | 31,289 | 36,369 |
| Notes payable and borrowings | 86,869 | 92,568 | 129,541 | n/a |
| Goodwill | 62,230 | 62,207 | 62,261 | – |
| Shareholders' equity | 8,704 | 20,451 | 42,508 | 66,772 |
Oracle has always run with high leverage: shareholders' equity was negative in fiscal 2022 after years of buybacks. The difference now is scale and direction. Borrowings rose about $37 billion in fiscal 2026. StockAnalysis puts debt/EBITDA at 4.5×, interest coverage at 4.8× and the Altman Z-score at 1.93, which is below the conventional 'safe' threshold. The $20 billion at-the-market equity sale in the August quarter (net proceeds $19.9 billion) explains most of the jump in equity, and shows management chose to dilute shareholders rather than add more debt.
VII.Capital Allocation & Dilution
| US$ millions | FY2022 | FY2023 | FY2024 | FY2025 | FY2026 |
|---|---|---|---|---|---|
| Share repurchases | 16,248 | 1,300 | 1,202 | 600 | 95 |
| Dividends paid | 3,457 | 3,668 | 4,391 | 4,743 | 5,787 |
| Stock-based compensation | 2,613 | 3,547 | 3,974 | 4,674 | 4,811 |
| Diluted shares (M) | 2,786 | 2,766 | 2,823 | 2,866 | 2,914 |
- Buybacks, $16.2B in fiscal 2022, have essentially stopped ($95M in fiscal 2026). The dividend has kept rising (12 years of increases, StockAnalysis).
- The share count is now rising. Shares outstanding went from 2,880.5M (12 Jun 2026) to 3,023.7M (7 Sep 2026), up 5.0% in under three months, after the equity sale. Further issuance is possible if capex stays above cash flow.
- Insiders, led by co-founder Larry Ellison, own about 38% of the shares (StockAnalysis), which aligns management with shareholders but concentrates control.
VIII.Five-Year P/E Projection (Illustrative)
Method. Today's price divided by consensus EPS for fiscal 2027 (year to May 2027) gives the current P/E of 16.9×. We take fiscal 2027 and fiscal 2028 consensus EPS as Years 1 and 2 in every scenario, grow EPS at a stated annual rate in Years 3–5, and apply an exit P/E to Year-5 (fiscal 2031) EPS. The base case holds today's P/E; the bear and bull cases use multiples 15% lower and higher and growth 3 points lower and higher. We compare the resulting Year-5 price with today's price as an annualized return, and discount it at a 10% required return to a present value. With the multiple held, the base-case return is simply EPS growth from Year 1 to Year 5.
| Input | Value | Type and basis |
|---|---|---|
| Share price | $137.30 | Source: NYSE close, 30 Sep 2026 |
| Year 0: FY2026 EPS | $7.63 non-GAAP $5.83 GAAP | Source: company results (Section XIV). Non-GAAP shown because consensus is on that basis |
| Trailing P/E | 18.0× 23.6× GAAP | Calculation: price ÷ FY2026 EPS. Shown for reference, not applied (see below) |
| Year 1: FY2027 consensus EPS | $8.14 | Source: Yahoo Finance, 37 analysts (range $8.07–$8.37), 30 Sep 2026 |
| Year 2: FY2028 consensus EPS | $11.00 | Source: Yahoo Finance, 39 analysts (range $9.00–$13.30); +35.1% on Year 1 |
| Current P/E (applied) | 16.9× | Calculation: price ÷ Year-1 consensus EPS (the fiscal year already under way) |
| Analysts' 5-year EPS growth | 26.6% a year | Source: Finviz "EPS next 5Y", 30 Sep 2026 |
| EPS growth, past 5 years | 5.1% a year | Source: Finviz "EPS past 5Y". Shown for contrast: the base case assumes a sharp break from history |
| EPS growth, Years 3–5 | Bear +23% · Base +26% · Bull +29% | Assumption. Base: analysts' 5-year estimate (26.55%) rounded down; bear and bull ±3 points |
| Exit P/E on Year-5 EPS | Bear 14.3× · Base 16.9× · Bull 19.4× | Assumption: current P/E −15% / held / +15% |
| Required return (present value) | 10% a year | Assumption |
Why today's forward P/E, not the trailing one. Fiscal 2027 began on 1 June 2026 and one quarter is reported (non-GAAP EPS $1.92). Unusually, the trailing P/E on fiscal 2026 non-GAAP EPS (18.0×) is only a little above the forward one, so the choice matters less here: holding it would give a base Year-5 price of about $396. We hold the 16.9× the market pays on fiscal 2027, consistent with our other notes.
Year-by-year EPS
| US$ per share | Y0 FY2026A | Y1 FY2027E | Y2 FY2028E | Y3 FY2029E | Y4 FY2030E | Y5 FY2031E |
|---|---|---|---|---|---|---|
| Bear EPS | $7.63 | $8.14 | $11.00 | $13.53 | $16.64 | $20.47 |
| Base EPS | $7.63 | $8.14 | $11.00 | $13.86 | $17.46 | $22.00 |
| Bull EPS | $7.63 | $8.14 | $11.00 | $14.19 | $18.31 | $23.61 |
| Growth, Years 1–2 (consensus) | +7% | +35% | ||||
| Growth, Years 3–5 | Bear +23% · Base +26% · Bull +29% | |||||
Year-5 price and implied return
| US$ | Bear | Base | Bull |
|---|---|---|---|
| EPS growth, Years 3–5 | +23% | +26% | +29% |
| Year-5 EPS (FY2031) | $20.47 | $22.00 | $23.61 |
| Exit P/E | 14.3× | 16.9× | 19.4× |
| Year-5 price | $293 | $371 | $458 |
| vs. today's price of $137.30 | +114% | +170% | +234% |
| Implied annual return (5-yr CAGR) | +16.4% | +22.0% | +27.2% |
| Present value at 10% a year | $182 | $230 | $284 |
| Present value vs. today's price | +33% | +68% | +107% |
IX.Return vs. Today's Price
Pathos illustrative Year-5 (fiscal 2031) price range, with each scenario's implied annual return. Red line: today's price. Dashed green line: the base-case Year-5 price discounted at 10% a year.
At $137.30 the base case implies about +22.0% a year over five years, with +16.4% in the bear case and +27.2% in the bull case, before a dividend yield of about 1.46%. Discounted at a 10% required return, the base-case Year-5 price is worth $230 today, +68% versus the price. Put the other way: holding today's P/E, a 10% annual return needs Year-5 EPS of about $13.11, which means EPS growth of about 6.0% a year in Years 3–5 after the consensus years. The cushion is large on paper, but it rests on three assumptions at once: that contracted AI revenue turns into profit at consensus margins, that growth of 23–29% continues for three more years after fiscal 2028, and that per-share earnings are not diluted by further equity raises.
X.Superinvestor Holders (Dataroma)
Which of the portfolio managers that Dataroma tracks own ORCL, according to their latest 13F filings? The figures below come only from Dataroma’s ORCL ownership page and activity page.
Source: Dataroma · retrieved 30 Sep 2026 · 13F period Q2 2026 (holdings at 30 Jun 2026)Dataroma shows 14 tracked superinvestors with a position at 30 Jun 2026. The largest weight belongs to Polen Capital Management, at 4.51% of the manager’s reported portfolio.
XI.Key Risks for Value Investors
Customer concentration and counterparty risk
Much of the $664B backlog comes from a small number of AI developers whose own funding depends on capital markets. If a large customer cannot pay, Oracle is left with specialised capacity and the debt that financed it.
Funding and leverage
Free cash flow is deeply negative and guided capex is $90–95B. Further borrowing or equity issuance is likely. Debt/EBITDA of 4.5× and an Altman Z of 1.93 leave limited margin for error.
Returns on AI infrastructure
Renting GPU capacity is a lower-margin business than database software. If pricing for AI compute falls as supply catches up, contracted revenue may convert into thinner profits than consensus expects.
Depreciation and accounting
Non-GAAP EPS excludes stock compensation and amortization. Depreciation on new capacity will rise steeply, and the assumed useful lives of GPUs materially affect reported earnings.
Dilution
The share count rose 5% in one quarter. Our projection uses consensus EPS, which may not fully reflect future issuance.
Legacy erosion
License and support revenue outside the cloud is shrinking (−3%), and the applications business grows only about 10%. The old cash engine is not getting stronger.
XII.Advice to Management
Pathos's long-term owner principles judge management by the per-share value it builds over many years, not by quarterly results. Applied to Oracle's figures in this note, we would offer the following recommendations.
Set a firm leverage ceiling and finance ahead of need
Oracle's notes and borrowings rose about $37 billion in fiscal 2026 to roughly $130 billion, trailing free cash flow is −$28.7 billion, and StockAnalysis puts debt/EBITDA at 4.5× and interest coverage at 4.8×. Fiscal 2027 capex is guided to $90–95 billion. We would publish a maximum leverage ratio that the company will not exceed in a bad year, and secure long-dated funding before committing capex rather than after.
Match long-lived capex to firm, creditworthy contracts
Remaining performance obligations of $664 billion are concentrated in a few very large AI customers, and about half convert to revenue within 36 months. Data centres and GPUs are long-lived; contracts can be renegotiated. We would pace spending to signed contracts with strong counterparties and report the expected return on capital for each cohort of capacity.
Issue shares only when the company receives full value
Oracle sold $20 billion of new shares in the August 2026 quarter, lifting the share count 5.0% in under three months, after the price had halved over the year. Selling stock at a depressed price transfers value from existing owners to new ones. We would explain to shareholders why this was cheaper than the alternatives and commit to issuing equity only when the price reflects intrinsic value.
Pause dividend growth while free cash flow is negative
Dividends were $5.8 billion in fiscal 2026 and have risen for 12 years, while free cash flow was −$23.7 billion and the gap was filled with debt and new shares. A dividend paid from borrowed or newly raised money is not a return of surplus capital. We would hold the dividend flat until free cash flow turns positive.
Report depreciation, leases and stock pay conservatively
Trailing depreciation of $9.8 billion is small next to $75.7 billion of trailing capex, operating and finance lease liabilities total about $39 billion, and stock compensation of $4.8 billion is excluded from the $8.10 non-GAAP EPS guidance. We would use conservative GPU useful lives, present leases next to debt and lead with earnings that include stock pay, so owners see the full cost of the build-out.
Opinion, not advice. These recommendations are Pathos's opinion, based on long-term owner principles and the figures in this note; they are not investment advice.
XIII.Conclusion
Oracle combines the most attractive earnings multiple in our coverage with the weakest cash position. The base case rests heavily on analysts' 26.6% long-term EPS growth estimate, which, unlike our semiconductor projections, we have not capped, because the growth is backed by contracted revenue rather than cyclical demand. That judgment drives the result, and we show the capped alternative below.
Pathos view (illustrative). On consensus and analysts' long-term growth, Oracle shows the largest projected return in our coverage: about +22.0% a year in the base case (Year-5 range $293–$458), worth about $230 today at 10%. With growth capped as for the chip companies, the base case falls to about $186 (+6.2% a year). The gap between those two numbers is the investment debate. We would gain confidence if (a) free cash flow turned positive as capacity comes on line, (b) depreciation on new capacity proved consistent with consensus margins, or (c) the customer base behind the backlog broadened. This is not a recommendation to buy, sell or hold.
XIV.Sources & Notes
As-of dates. Share price, market capitalisation and price-based multiples use the NYSE close of $137.30 on 30 September 2026. Consensus, the 52-week range and analysts' long-term growth estimate were retrieved on 30 September 2026. Annual figures are from Forms 10-K via SEC XBRL data. The latest quarter is from the 10 September 2026 results release, the earnings call and the 10-Q for the quarter ended 31 August 2026. Ratios and free cash flow are Pathos calculations unless attributed.
- Oracle Corporation, Forms 10-K and 10-Q, via SEC XBRL company facts and the Form 10-K for fiscal 2026.
- Oracle Form 8-K, 10 Sep 2026: first quarter fiscal 2027 results (revenue by line, OCI growth, EPS, RPO, cash flow, equity sale, dividend, Q2 and FY2027 guidance).
- Oracle Form 10-Q for the quarter ended 31 Aug 2026 (filed 11 Sep 2026): balance sheet, leases, share count.
- Oracle Q1 FY2027 earnings call, 10 Sep 2026 (highlights via Yahoo Finance): capex guidance of $90–95B, OCI revenue, RPO conversion within 36 months.
- Yahoo Finance, ORCL analysis (retrieved 30 Sep 2026): FY2027–FY2028 consensus EPS and revenue, ranges, analyst counts; FY2026 non-GAAP EPS.
- Finviz, ORCL (retrieved 30 Sep 2026): analysts' five-year EPS growth estimate (26.55%), 52-week range.
- StockAnalysis, ORCL statistics (retrieved 30 Sep 2026): close, enterprise value, trailing ratios, TTM cash flow, debt incl. leases, Altman Z, interest coverage, insider ownership, mean price target.
- SEC XBRL company facts (SEC EDGAR XBRL API, retrieved 30 Sep 2026): quarterly revenue, net income and diluted EPS from Forms 10-Q and 10-K (fiscal Q4 derived as full year minus nine months), and quarter-end equity, goodwill, intangible assets, preferred stock and shares outstanding. Used for the quarterly charts in Section IV.
- Yahoo Finance, ORCL daily price history (retrieved 30 Sep 2026): split-adjusted quarter-end closing prices for the trailing P/E line.
- Dataroma, ORCL superinvestor ownership page and activity page (retrieved 30 Sep 2026): ownership count, rank, % of all portfolios, hold price, holders, % of portfolio, shares, value and recent activity for 13F period Q2 2026 (holdings at 30 Jun 2026). 13F data are filed up to 45 days after quarter-end and cover long US-listed positions only.
Not verified or not used: the identity and share of Oracle's largest AI customers; the August 2026 debt balance from the 10-Q (we cite StockAnalysis for total debt including leases); useful-life assumptions for GPU assets; Q2 FY2027 results (due December 2026).