Oracle Corporation: A Long-Term Business Owner's Perspective

Moat breakdown. Oracle's strongest moat is not network effects but switching costs. Once databases, ERP, finance, HR, and supply-chain systems are deeply embedded in enterprise processes, migration risk is extremely high, implementation cycles are extremely long, and the cost of failure is enormous. Next is scale and installed base: a large number of customers already run within Oracle's database, Fusion, NetSuite, Java, industry suites, and support-contract ecosystem, and Oracle can bundle applications, databases, and cloud infrastructure into a single sale. Brand strength is significant in enterprise IT but nearly meaningless to ordinary consumers; network effects are relatively weak; the channel and partner ecosystem is moderate; regulation and patents are not core moats; the data advantage exists in vertical settings such as Oracle Health but is not the company's foremost advantage overall. On balance, I view the moat as stable and slightly widening: the legacy-business moat is stable, the new OCI moat is strengthening, but it has not yet reached the unassailable level of Azure or AWS.

Pricing power and resilience to cycles. Oracle has some ability to raise prices in database maintenance, ERP, and mission-critical applications, and in an inflationary environment it can usually pass through part of its costs via renewals, version upgrades, and portfolio selling; but in OCI, its pricing power is markedly weaker than in maintenance support. In economic downturns, Oracle's mission-critical software revenue is typically more resilient than advertising or consumer internet, and in fiscal 2022, despite the impact of consolidating Cerner, it still achieved profitability with positive operating cash flow, evidence of strong defensiveness in the legacy business. High margins overall reflect a structural advantage rather than a mere cyclical dividend.

Management, Capital Allocation, and Governance

Fact. Oracle's current management underwent a major change in September 2025: Clay Magouyrk and Mike Sicilia were promoted to co-CEO, Safra Catz moved to Executive Vice Chair, and Larry Ellison remains Chairman and CTO. The arrangement is itself very Oracle: Ellison still tightly controls technology and strategic direction, while the new CEO pairing represents the two main lines of OCI and application software.

Fact. Ownership is highly concentrated. According to the 2025 proxy filing, Larry Ellison holds roughly 1.158 billion shares, about 40.6% of the company; all current directors and officers together hold about 40.9%. This means management and shareholder interests are tightly aligned, but it also means governance is not "diffuse and democratic" but clearly a founder-led company. For long-term investors this is double-edged: incentives are aligned on one hand, but the decision-making style may be more aggressive on the other.

Fact. Oracle's governance framework is not loose. The proxy filing discloses executive stock-ownership requirements, a clawback policy, and an anti-hedging policy, and notes that as of 2025-09-19, non-employee directors and senior officers had met or were on track to meet ownership requirements within the allotted time. In fiscal 2025, the compensation committee also decided not to award annual cash bonuses to Ellison and Catz, stating explicitly that this was to preserve cash and support AI strategic projects, not because of poor performance. This move shows the board at least formally understands capital scarcity and prioritization.

Fact. The capital-allocation track record is "strong earlier, riskier later." Oracle has long been very good at buybacks and dividends, with cash dividends growing from $3.67 billion in 2021 to $4.74 billion in 2025; but buybacks slowed markedly, and the company stated in its 2024 annual report that future buyback levels may be below the past and will not be raised until total debt falls below a certain threshold. By fiscal 2026, reality is even more aggressive: in the first nine months through 2026-02-28, capex reached $39.17 billion, net debt issuance was $44.54 billion, and the company issued roughly $4.95 billion of mandatory convertible preferred stock.

M&A assessment. In 2022 Oracle completed its acquisition of Cerner for a transaction value of about $28.3 billion, the largest acquisition in company history. The deal gave Oracle a healthcare IT platform and a large, regulation-heavy slice of the industry software market, but it also significantly raised leverage and depressed buyback capacity in subsequent years. Whether it ultimately created enough value still cannot be considered "proven" today. From a long-term owner's perspective, I would view Cerner/Oracle Health as a capital-allocation case that is "potentially valuable but with an as-yet-unproven return."

Assessment. Management has two clear strengths: it is no stranger to shareholder returns, and it is decisive enough on strategy; but the biggest question today is not "do they understand capital allocation" but "are they allocating capital in a higher-risk way." When Oracle was a high-cash-flow, low-capex software company, this approach was excellent; now that Oracle has become a heavy-investment player in AI data centers, it demands a higher evidentiary bar. My score: management and capital allocation 3/5.

Financial Quality and Owner Earnings

The table below summarizes Oracle's key financials for the most recent five complete fiscal years. The FY2021–FY2025 figures for revenue, operating income, net income, operating cash flow, capex, free cash flow, share count, and borrowings come, respectively, from Oracle's FY2022, FY2024, and FY2025 annual reports.

Units: billions of USD; share count is the diluted weighted-average share count; shareholders' equity is total shareholders' equity.

First, quality. On the revenue side, 2021–2025 revenue rose from $40.48 billion to $57.40 billion, a four-year compound growth rate of about 9.1%; on the profit side, 2022 margins were depressed by Cerner consolidation and integration, after which operating margin recovered from 25.7% to 30.8%; on the cash side, operating cash flow grew from $15.89 billion in 2021 to $20.82 billion in 2025, and remains strong over the long run. Looking only at the period before 2025, this is a very typical enterprise software company that "turns profit into cash."

But the key change is in capex. FY2024 capex was just $6.87 billion; FY2025 suddenly jumped to $21.22 billion, and free cash flow swung from $11.81 billion to -$0.39 billion; by the first nine months of fiscal 2026, operating cash flow was $17.36 billion, but capex reached $39.17 billion, and free cash flow was about -$21.81 billion. Management simultaneously guided FY2026 capex to $50 billion. The conclusion is direct: profit still looks good, but real cash flow has been dragged into a period of heavy reinvestment by the AI data-center expansion. This is not accounting-profit fabrication; on the contrary, it shows precisely that rising capital intensity is genuinely happening.

On the balance sheet, Oracle has always preferred to optimize its capital structure with debt plus buybacks, which has long distorted ROE. More worth watching is the scale of borrowings: as of 2025-05-31, total borrowings were about $92.57 billion; and as of 2026-02-28, current and non-current borrowings combined had risen to about $134.61 billion, with shareholders' equity rising to $39.05 billion. Debt rose markedly over the first nine months, corresponding mainly to enormous capex and financing arrangements. FY2025 interest expense was $3.58 billion; interest expense in the first nine months of fiscal 2026 was $3.16 billion, higher than the $2.60 billion in the prior-year period. For a company once known for software cash flow, this leverage level has entered a range that must be taken seriously.

On working capital, Oracle's traditional strength is prepayments and renewals, so cash conversion was high in the past. But in the first nine months of fiscal 2026, trade receivables rose $2.20 billion, accounts payable and other liabilities fell $0.82 billion, income taxes payable fell $1.65 billion, while deferred revenue rose only $0.25 billion. Simply put, growth is now consuming cash rather than releasing large amounts of cash as in the past. For valuation, this matters more than the growth rate of profit.

Accounting-quality assessment. I see no clear signs of financial fabrication or large-scale aggressive accounting. The 2024 annual report's cover page discloses no prior financial-statement error corrections and no resulting executive compensation clawback analysis; the company's internal-control audit opinion is issued by EY. What truly warrants caution is not "whether the profits are fake," but that stock-based compensation is a real cost, growth capex is a real cost, and the fulfillment of future large deals carries execution risk.

Owner earnings estimate. Here we must strictly separate "fact" from "inference." Fact: In the third quarter of fiscal 2026, Oracle disclosed trailing-twelve-month operating cash flow of about $23.5 billion; rolling estimates based on the FY2025 10-K and the 2026 Q3 10-Q put TTM capex at about $48.25 billion, so reported FCF is about -$24.7 billion, consistent with third-party data. Inference: This -$24.7 billion cannot be taken directly as "permanent earning power," because most of it is clearly growth-oriented AI infrastructure investment rather than the maintenance capex needed to keep the old Oracle business running. Conservative estimate: Starting from TTM operating cash flow of $23.5 billion, deducting real stock-based compensation cost of about $4.9 billion, and then assuming maintenance capex of about $9–10 billion, I give Oracle a conservative owner-earnings range of about $8.5–9.5 billion, with a midpoint around $9 billion. This basis is clearly closer to "cash distributable without harming the moat" than GAAP net income, and clearly better reflects the legacy business's earning power than reported FCF; but it remains an inference, because maintenance capex cannot be mechanically separated from the financial statements.

At the current market cap of about $528.4 billion, Oracle corresponds to about 56–62x conservative owner earnings; on reported TTM FCF, P/FCF is in fact meaningless, because free cash flow is negative. For a "Buffett-style" investor, this means: buying Oracle today is essentially prepaying for several years of organic growth and improvement in returns.

Valuation and Margin of Safety

At the current price, Oracle's valuation can no longer be judged by P/E alone, because P/E still rests on the income statement, whereas long-term returns are truly determined by: whether RPO can become revenue, whether revenue can become cash, and whether cash can exceed reinvestment and financing costs. Oracle's crux today is the third step, not the first.

Discounted Owner Earnings

The valuation below is an inference based on three owner-earnings scenarios from conservative to optimistic, not a fact. To avoid false precision, I give ranges rather than single points:

The logic behind these scenarios: the conservative case assumes Oracle merely runs the "old software cash cow" steadily while AI/OCI returns are mediocre; the neutral case assumes OCI and multi-cloud database growth continues for several years and capex falls back markedly after FY2027; the optimistic case assumes AI contracts are fulfilled excellently after FY2027, capex as a share of revenue declines significantly, and new businesses still earn high returns. At the current price of $181.46, the stock is already near the top of my optimistic scenario. Therefore, from a long-term owner-earnings perspective, the margin of safety is not evident.

Relative Valuation

The table below is an approximate valuation snapshot around mid-May 2026. Because data providers, accounting conventions, and fiscal years differ, the table is better suited to directional comparison than to precise pricing to the decimal.

The relative-valuation conclusion is clear: Oracle is by no means the "obviously cheap" name among its peers. Versus Microsoft, it has lower ROIC and greater financial pressure yet no clear valuation discount; versus Salesforce and SAP, it is not cheap on EV/EBITDA either, and its TTM free cash flow is even negative. The only thing that can defend the current pricing is the market's belief that Oracle's AI infrastructure investment will rapidly convert into very high-quality revenue and cash flow over the next two to three years. That assumption is not impossible, but it is by no means a "margin of safety."

Asset and Liquidation Value

For Oracle, the asset approach offers very weak support. As of 2025-05-31, the company carried goodwill of about $62.2 billion and net intangibles of about $4.6 billion; by 2026-02-28, total borrowings were about $134.6 billion, far above liquid assets. In other words, Oracle's value comes almost entirely from its going-concern earning power, not from liquidation value. Book value and true value are severely distorted, and P/B can serve only as a supplementary indicator.

Price Range Assessment

This is my range assessment based on owner earnings and capital-return risk. Through this lens, $181.46 carries a clear premium to the fair intrinsic value range and is only marginally defensible under the optimistic scenario. For a balanced long-term investor, I consider the current price not cheap enough and not offering a sufficient margin of safety.

Risks, the Bear Case, and Opportunity Comparison

The most important risk is not short-term volatility but permanent loss of capital. Each of the risks below could change the conclusion:

Most of the above risks have clear support: in its 2026 Q3 earnings release, Oracle directly lists data-center capacity, GPU procurement, government contracts, and legal/regulatory changes as risks; Synergy's data show the OCI market is still dominated by the three giants; and the 2026-02-28 10-Q shows that enormous capex and added leverage are reality, not imagination.

The strongest bear case is, in fact, straightforward: Oracle may not be a "cheap legacy software leader in transition" at all, but rather a case where "the market overestimates the quality of AI mega-deals and underestimates capital intensity and financing costs." In other words, if over the next two to three years RPO cannot convert into high-margin revenue at high quality, or capex cannot decline meaningfully, then Oracle's valuation basis will fall from "growth cloud platform" back to "mature software plus mid-tier cloud platform," at which point a large part of the optimism embedded in the share price will be stripped away.

Which facts would overturn this report? If, over the next 6–8 quarters, the following facts emerge, I would admit I was too conservative: Oracle continues to grow revenue rapidly while capex/revenue falls quickly; OCI growth stays above the industry and markedly lifts margins; net debt/EBITDA falls clearly; RPO converts to revenue at high speed with very few bad debts or renegotiations; and free cash flow per share returns to double-digit growth. Conversely, if OCI growth drops quickly below 30%, RPO growth slows markedly, capex still exceeds 20%–25% of revenue long after FY2027, debt keeps rising, or financing again dilutes common shareholders, then the investment thesis needs to be re-examined.

Comparison with other opportunities. Versus the S&P 500, Oracle offers no firm conviction of being "significantly better than buying the index." Owning it, you bear single-company execution risk, leverage risk, and AI-infrastructure cycle risk; buying the index, you get diversification. Versus the U.S. 10-year Treasury, Oracle's yield on a trailing-P/E basis is only about 3.1%, below the contemporaneous 10-year Treasury at about 4.61%, and reported FCF is still negative, meaning what you get is not a present cash return but a bet on future growth. Unless you have a clear informational edge over the market on the quality of Oracle's OCI and AI contracts, it is not inherently superior to the index or the risk-free rate.

If I could hold only five assets, at the current price I would not put Oracle in the portfolio. If the price returns to a zone that I believe better reflects a margin of safety, or if the coming year proves that "high capex is only a short bridge, not the new normal," then it could re-enter the candidate list. This conclusion is not bearish on Oracle's business; it places weight on price and rate of return.

Checklist and Final Judgment

【Final Rating】 Watch

【One-Sentence Investment Thesis】 Oracle is a high-quality software cash cow layered with a high-potential but capital-intensive AI/OCI expansion story; the business is good, but the current price lacks a Buffett-style margin of safety.

【Core Bull Case】

• The core database, ERP, and maintenance-support businesses have high switching costs and strong recurring-revenue characteristics.

• From 2021 to 2025, revenue, operating income, and operating cash flow all trended up, with excellent historical financial quality.

• OCI and multi-cloud database growth is extremely high and RPO is growing explosively, evidence that demand is real.

• Larry Ellison's high ownership creates strong incentive alignment.

• If the AI infrastructure investment ultimately monetizes at high returns, Oracle's per-share value still has room to be revised upward.

【Core Bear Case】

• From FY2025 to FY2026, capex and financing rose sharply, and reported FCF has turned significantly negative.

• The current valuation is not cheap and sits near the top of my optimistic scenario's value range.

• The OCI market is still dominated by AWS, Microsoft, and Google, and competition is not easy.

• After the debt increase, future shareholder returns rely more on execution than on the existing moat.

• The current investment logic looks more like "paying up front before growth is verified" than "buying cash flow at a discount."

【Key Assumptions】

• AI/OCI mega-deals can convert into high-quality revenue as expected.

• Capex as a share of revenue declines markedly after FY2027.

• OCI/cloud-application growth ultimately covers higher depreciation, interest, and potential dilution.

• The return on the company's incremental investment is at least no lower than its cost of capital.

【Fair Buy Price】 $80–110. Within that, $80–95 better fits a conservative value investor's requirements; $95–110 suits investors more confident in OCI/AI fulfillment. The basis is the three-scenario owner-earnings valuation above plus a margin-of-safety requirement of roughly 25%.

【Target Holding Period】 10 years or more. The premise is that your entry price does not overpay too much of the future return.

【Expected Annualized Return】

• Conservative scenario: -5% to 0%

• Neutral scenario: 2% to 6%

• Optimistic scenario: 9% to 13% This is a rough inference based on the current price relative to the three intrinsic-value frameworks above, not a return guarantee.

【Maximum Loss Risk】 If AI contract fulfillment falls short of expectations, capex stays high for the long term, and debt keeps rising while the valuation falls back toward a peer-like EV/EBITDA range, the share price carries 35%–50% of medium-term downside; layering on demand renegotiation or financing constraints, a permanent capital loss of 50%–65% is not unimaginable in the extreme case.

【Tracking Metrics】 Track continuously: OCI revenue growth, total cloud revenue growth, RPO growth and conversion efficiency to revenue, capex/revenue, operating cash flow, reported FCF, net debt/EBITDA, interest expense, stock-based compensation and share dilution, and Oracle Health's goodwill and operating performance.

【Signals That Trigger Reassessment】

• OCI growth falls clearly below expectations for several consecutive quarters.

• RPO growth is high but revenue conversion is weak, or signs of contract renegotiation/cancellation emerge.

• Capex stays elevated after FY2027 with no FCF recovery.

• Borrowings keep rising sharply, or large-scale equity financing is needed again.

• Major governance changes, accounting issues, or a clear impairment at Oracle Health.

【Open Questions and Limitations】 This report makes a conservative inference about the "maintenance capex" within owner earnings, a component that is not a fact readable directly from the financials; in addition, some peers' P/B and ROIC figures vary substantially across databases, so they are used only for directional comparison, not as a basis for the core conclusion.

【Final Recommendation】 Calmly place Oracle on the watch list rather than treat it as one of the rare bargains in the current market. This is an excellent company, but an excellent company does not automatically warrant buying at any price. For balanced investors with a 10-year-plus horizon, the best action is not to chase the narrative but to wait for at least one of two kinds of evidence to appear: a cheaper price, or more solid cash-flow fulfillment.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.