Oracle Corporation
Consolidated Instrument Panel
Source: Morningstar, CNN Markets, stockanalysis.com, fullratio.com, investing.com consensus (data as of Aug 28–29, 2026). EV/Revenue derived from market cap plus total debt less cash, tagged (est.).
The forward multiple in the panel above sits well below Oracle's own ten-year average near 28.8x, a discount that reflects financing risk rather than a re-rating of the underlying database and applications franchise. At a PEG below 1, the stock prices in materially less growth than consensus estimates currently project, even after accounting for the earnings quality questions raised in Block C.
Quarterly Table
| Quarter | Revenue vs. cons. | GAAP EPS vs. cons. | Adj. EPS vs. cons. | Op. margin (rep./adj.) | FCF | Next-session reaction |
|---|---|---|---|---|---|---|
| Q4 FY26 ended May 31, 2026 |
$19.18B / $19.10B BEAT | $1.45 / (data not found) | $2.11 / $1.96 BEAT | 31.8% (est.) / 44.8% (est.) | (data not found) | -7.0% |
| Q3 FY26 ended Feb 28, 2026 |
$17.20B / ~$16.93B BEAT | $1.27 / (data not found) | $1.79 / $1.70 BEAT | 32.0% / 43.0% | (data not found) | +10.0% |
| Q2 FY26 ended Nov 30, 2025 |
$16.06B / $16.21B MISS | $2.10 / (data not found) | $2.26 / $1.64 BEAT (est.-boosted) | 29.2% / 41.6% | -$10.0B | -11.0% |
| Q1 FY26 ended Aug 31, 2025 |
$14.93B / $15.04B MISS | $1.01 / (data not found) | $1.47 / $1.48 MISS | 28.9% / 41.6% | (data not found) | +30.0% |
Source: Oracle 8-K exhibits (SEC EDGAR), Oracle Investor Relations press releases, LSEG/CNBC consensus, FactSet (via Forbes), Motley Fool earnings call transcript. Q4 operating income derived from FY2026 total less Q1–Q3, tagged (est.). Next-quarter (Q1 FY27) results were not yet published as of this report's publication date.
Reading the Quarters
Oracle closed fiscal 2026 with revenue and adjusted earnings both ahead of Street estimates, extending the beat streak visible in the table above. Cloud infrastructure revenue nearly doubled year over year, and remaining performance obligations jumped $85 billion in a single quarter to $638 billion, propelled by AI contracts in which customers increasingly prepaid for GPU capacity rather than waiting for Oracle to fund it outright. That financing detail mattered more to the market than the headline beat. The after-hours decline recorded in the reaction column above followed management's disclosure that full year capital expenditure had reached $55.7 billion, well above the $50 billion guided only a quarter earlier, and that fiscal 2027 would require another $40 billion in debt and equity financing. Gross margin compression from the data center ramp, flagged explicitly by management, added to the unease. The quarter illustrated a pattern that would recur across the year: operating results kept beating estimates while financing requirements kept growing faster than investors could comfortably underwrite, leaving the print itself almost secondary to the capital plan buried in the transcript.
Q3 marked the first quarter in over fifteen years in which Oracle's organic revenue and non-GAAP earnings per share both grew above twenty percent in the same period, and both figures cleared consensus as shown in the table above. Cloud revenue accelerated to 44 percent growth, and remaining performance obligations climbed to $553 billion, a figure management attributed heavily to large AI contracts that, in this instance, did not require incremental capital because customers largely financed the hardware themselves. The company also disclosed a 15 percent equity stake in the newly independent TikTok US, an unusual source of nonoperating income unrelated to the core cloud story. The extended-trading rally captured in the reaction column above was the strongest positive move of the four quarters under review, as investors credited Oracle with proving its AI backlog could convert into reported growth rather than remain a promise. The quarter stands as the sentiment high water mark of the trailing year, a useful contrast against the capital anxiety that resurfaced three months later.
Reported revenue fell short of Street expectations, and shares sold off sharply in extended trading, Oracle's worst single-session reaction across the four quarters, as shown in the table above. The headline adjusted earnings figure, well above consensus, flattered the result: a $2.7 billion pretax gain from the sale of Oracle's stake in Ampere Computing inflated both GAAP and non-GAAP earnings per share, and stripping it out would have left profitability much closer to expectations. Free cash flow for the quarter also missed the Street's own negative estimate by a wide margin, a gap management addressed directly on the call by lifting full year capital expenditure guidance to roughly $50 billion from $35 billion in September. Remaining performance obligations rose $68 billion sequentially to $523 billion, with new commitments from Meta and NVIDIA cited by name, yet the market's attention stayed fixed on financing math rather than backlog growth. The stock's 23 percent decline for the month of November, its worst since 2001, dated from this release.
Oracle's fiscal 2026 opened with revenue and adjusted earnings both narrowly below Street estimates, as shown in the table above. Ordinarily a double miss produces a selloff. Instead, the stock spiked sharply the next session, per the reaction column above, the strongest positive move Oracle has produced from an earnings release in years, because remaining performance obligations jumped 359 percent to $455 billion on four multibillion dollar contracts signed with three customers, and management laid out a five year cloud infrastructure revenue path running from $18 billion in fiscal 2026 to $144 billion by fiscal 2030. Then chief executive Safra Catz called it an astonishing quarter, and traders agreed enough to push options activity to unusual levels in the print's aftermath. The quarter set the template that would define the rest of the fiscal year: near term operating figures became secondary to the trajectory implied by the backlog, and the market's tolerance for that framing would prove considerably more fragile by the time Q2 arrived two months later.
Patterns & Watch List
Structural Patterns
- Backlog growth has outpaced revenue conversion every quarter. As Block C's quarterly progression shows, remaining performance obligations expanded roughly 40 percent in aggregate across the year while trailing revenue growth stayed in the high teens.
- Stock reactions have tracked financing disclosures more closely than headline beats or misses. Q1 and Q3 both beat or nearly matched estimates and produced double digit rallies tied to backlog commentary, while Q2 and Q4 beat on earnings yet sold off once capex and debt guidance came into view.
- Non-GAAP earnings have been repeatedly inflated by one-time items. The Ampere sale added materially to Q2 adjusted EPS, and Q4 included separately disclosed net investment gains, meaning reported growth rates overstate underlying operating leverage.
- Non-GAAP operating margin held a narrow band across the year, opening at 41.6 percent in Q1 and reaching an estimated 44.8 percent in Q4, even as management guided that fiscal 2027 margins would step down as new data center capacity ramps ahead of the revenue that fills it.
Watch List · Next Print
- ↑Q1 FY27 revenue and cloud growth against the guided 27–29% band. Confirmed if growth lands inside guidance; broken if it undershoots amid OpenAI financing headlines.
- →Capital expenditure discipline against the FY27 net capex guidance detailed in Block F. Confirmed if quarterly capex tracks roughly one quarter of that figure; broken by another mid-year upward revision.
- ↓Customer concentration disclosure. Confirmed (negatively) if Oracle continues withholding OpenAI's share of RPO; a positive break would be voluntary disclosure that lowers perceived tail risk.
- ↓Free cash flow trajectory. Confirmed if FY27 free cash flow stays negative through the first half; broken by an inflection toward positive free cash flow as prepaid AI contracts convert to cash.
Strategic Architecture
E.1 · Business Model & Revenue Architecture
| Segment | FY26 revenue | % of total | YoY | Margin profile |
|---|---|---|---|---|
| Cloud (IaaS + SaaS) | $34.0B | 50.4% | +39% | Improving with scale, currently thinnest |
| Software (license & support) | $24.5B | 36.4% | -1% | Highest in the portfolio |
| Services | $5.7B | 8.5% | +10% | Modest, labor-intensive |
| Hardware | $3.1B | 4.6% | +5% | Lowest, commoditized |
Source: Oracle FY2026 Q4 and full-year 8-K (SEC EDGAR). Segment sum of $67.3B reconciles approximately to the reported $67.4B total revenue after rounding.
Oracle now operates as two businesses layered inside one income statement: a legacy license and support franchise that still throws off revenue near flat year over year, and a cloud infrastructure and applications business growing well into double and triple digits. The legacy piece, roughly 36 percent of fiscal 2026 revenue, carries the highest margin in the portfolio and effectively subsidizes the cash burn on the newer piece while it scales. Cloud revenue, just over half of the total, splits between infrastructure, now growing near triple digits off a low base, and applications, growing at a comparatively modest low double digit pace as Fusion ERP and NetSuite mature into large, slower-growing franchises in their own right. Services and hardware round out the remainder, largely low margin and strategically peripheral. Operating leverage in this structure does not come from the traditional software playbook of incremental license sales against fixed infrastructure. It comes instead from utilization of the AI data center fleet Oracle is building, where gross margin depends heavily on how quickly newly energized capacity fills with paying workloads relative to the depreciation and power costs that begin accruing the day a facility goes live. That dependency explains why management has guided fiscal 2027 gross margins lower even as revenue accelerates: the company is intentionally running ahead of demand on capacity to win contracts like the OpenAI Stargate agreement, financing the gap with debt and customer prepayments rather than internally generated cash. The multicloud database business, letting Oracle's database run natively inside AWS, Azure, and Google Cloud, is the one segment growing off both a low base and clean unit economics, since it monetizes data gravity Oracle already owns rather than infrastructure it must build. That segment, still small in absolute dollars, is the clearest evidence that Oracle's competitive advantages are not confined to the capital-intensive parts of the story.
E.2 · Competitive Moat
| Pillar | Strength | Erosion risk | Rating |
|---|---|---|---|
| Database switching costs & multicloud distribution | Very high | Low | ★★★★★ |
| Enterprise applications (Fusion ERP, NetSuite) | Moderate | Moderate (Workday, SAP) | ★★★ |
| AI cloud infrastructure (OCI) scale | Emerging | High (hyperscalers, chip cycle) | ★★ |
Source: PGS internal analysis based on segment disclosures, earnings call commentary and Morningstar/GuruFocus moat framework references.
Oracle's most durable moat sits in database switching costs accumulated over four decades of mission critical enterprise deployments, reinforced now by the multicloud database initiative that lets customers keep Oracle's database engine while moving compute to a hyperscaler of their choice. That combination is difficult to replicate because it requires both decades of engineering trust and commercial relationships with three competing hyperscalers simultaneously, something no pure infrastructure entrant can offer. The enterprise applications suite carries a real but narrower moat built on implementation cost and workflow lock-in rather than technical differentiation, competing directly against Workday and SAP on comparable terms. The newest and least proven pillar is AI cloud infrastructure itself. Oracle has captured genuine scale, evidenced by a remaining performance obligation figure that quadrupled inside a year, but the underlying mechanism, being a fast-building landlord for GPU capacity, is far easier for Amazon, Microsoft, Google, or well-capitalized specialists like CoreWeave to contest than a four-decade-old database franchise. Pricing power in GPU rental is inherently thinner than pricing power in mission critical database licensing, and the chip generation turnover investors have flagged as a risk applies equally to every infrastructure provider in this category, Oracle included. Investors underwriting the current valuation are, in effect, buying the database moat at a premium multiple and receiving the AI infrastructure bet largely for free, or paying up for the infrastructure bet and receiving the database moat as ballast, depending on which side of that debate they occupy.
E.3 · Long-Term Growth Architecture
| Driver | Timeline | Current state | Potential | Status |
|---|---|---|---|---|
| OCI / Stargate AI infrastructure buildout | FY26–FY30 | Scaling rapidly, backlog at a record level | Very large | at risk |
| Multicloud database (AWS/Azure/GCP) | FY26–FY28 | Fastest-growing line, +404% in Q4 | Large | on track |
| Cloud applications (Fusion ERP, NetSuite) | Ongoing | Steady double-digit growth | Moderate | on track |
| Oracle Health AI (Cerner) | FY27+ | Early rollout | Moderate | emerging |
Source: Oracle Q4 FY2026 earnings call and press release commentary (Investing.com transcript, June 10, 2026).
Sequencing matters more than breadth here. The Stargate-anchored infrastructure buildout must convert booked RPO into recognized, cash generative revenue before any other growth driver can meaningfully move the multiple, since it is both the largest dollar opportunity and the one consuming essentially all of the company's incremental capital. Multicloud database expansion and the Oracle Health AI rollout are lower capital intensity and can proceed largely independent of the infrastructure buildout's outcome, which is why they are rated on track rather than at risk. The infrastructure driver is rated at risk not because underlying demand is illusory, but because its economics depend on a financing structure, prepayments, customer-supplied hardware, and debt issuance, that has not yet been tested through a full capacity build and utilization cycle. If OCI utilization lags the aggressive capacity additions Oracle has guided, the entire growth stack reprices downward together, since the database and applications businesses, while healthy, are not large enough on their own to justify the current enterprise value. The infrastructure driver is the gating item; everything else in the growth story is real but secondary until it clears.
Risks
Oracle's RPO backlog is dominated by a small number of very large AI infrastructure contracts, most prominently the roughly $300 billion, five year Stargate compute agreement tied to OpenAI. Bank of America analysts have estimated that more than half of Oracle's remaining performance obligations trace to this single relationship. The mechanism is straightforward: Oracle is building and financing data center capacity years ahead of when contracted revenue will be recognized, and if OpenAI's own financing or demand assumptions falter, if a competing provider wins expansion capacity as has already occurred at specific sites, or if the venture's structure shifts, a large share of Oracle's forward backlog could reprice or unwind. Reporting has documented lenders declining to finance Stargate-adjacent developments over Oracle counterparty concentration concerns. The observable signal to watch is any disclosure of customer-level RPO concentration, currently withheld, or any renegotiation of Stargate contract terms in a future filing.
Fiscal 2026 free cash flow was negative $23.7 billion, and Oracle has guided to roughly $70 billion in net capital expenditure for fiscal 2027, funded through an already announced $20 billion at-the-market equity program alongside further debt issuance on top of more than $167 billion in existing debt. The mechanism is a timing mismatch inherent to infrastructure buildouts: capital goes out immediately for chips, power and construction, while contracted revenue converts over five to ten year terms. The risk is that financing conditions tighten, credit spreads widen further as they have already begun to do, or equity issuance dilutes shareholders faster than the backlog converts to cash generative revenue. The observable signal is credit market pricing itself: further widening of Oracle's credit default swap spreads or a ratings action would confirm the risk is moving from theoretical to structural, while a shift toward positive quarterly free cash flow would indicate the buildout is beginning to self-fund.
Oracle Cloud Infrastructure competes directly against Amazon, Microsoft and Google, each carrying a larger balance sheet and, in Microsoft and Google's cases, in-house chip design that Oracle lacks following the sale of its Ampere stake. The mechanism is straightforward margin compression: as more capacity comes online industry-wide, GPU rental pricing faces downward pressure, and Oracle's comparatively higher leverage leaves less room to compete on price than peers funding buildouts from operating cash flow. A related hazard is generation mismatch: AI accelerator chips refresh faster than data centers can be built, raising the possibility that facilities designed around one chip generation become less competitive before their construction cost is recovered. The observable signal is Oracle's own disclosed cloud infrastructure pricing and utilization commentary on future earnings calls, alongside customer commentary on comparative GPU rental pricing across the major providers.
Scenarios
Assumes RPO converts to recognized revenue roughly on the schedule management has guided, OCI utilization climbs fast enough that non-GAAP operating margin resumes expansion rather than the step-down currently guided for fiscal 2027, and the OpenAI Stargate relationship proceeds without renegotiation or a customer-side financing disruption. Under these conditions the multicloud database and enterprise applications franchises keep compounding at their current double-digit pace largely independent of the infrastructure outcome, providing a profitability floor beneath the higher-growth, lower-margin cloud infrastructure business. Multiple expansion back toward Oracle's own historical range would follow evidence that the financing structure is working as designed, meaning prepayments and customer-supplied hardware keep reducing Oracle's own capital requirement as they did in Q3 and Q4 FY2026.
Assumes OpenAI's financing or demand for Stargate capacity falls short of the contracted schedule, forcing either a renegotiation of committed volumes or a write-down of related backlog, while credit markets simultaneously reprice Oracle's debt wider given its already elevated leverage relative to hyperscaler peers. Under these conditions capital expenditure continues at a pace free cash flow cannot support, forcing either dilutive equity issuance beyond the program already announced or a construction slowdown that cedes competitive ground to Amazon, Microsoft and Google. Non-GAAP operating margin compresses further as underutilized capacity depreciates against thinner revenue than guided, and the market extends the multiple discount already visible in Oracle's forward P/E relative to its own five-year average.
PGS Verdict
Oracle trades near a 56 percent discount to its September 2025 high and at a forward P/E below its own five year average, a dislocation that reflects genuine financing uncertainty rather than a deterioration in the underlying database and applications franchise, which continues compounding at healthy double digit rates largely untouched by the AI infrastructure controversy. The gap between Wall Street's own bullish consensus, summarized in the panel above, and the stock's actual price action suggests the market is pricing a wider range of financing outcomes than the sell side's models currently capture.
Beneath the AI infrastructure headlines, Oracle's core business remains fundamentally healthy. Database switching costs, an entrenched multicloud distribution advantage across three competing hyperscalers, and a maturing enterprise applications suite together generate the operating income that funds the AI buildout without yet depending on the buildout's own economics to prove out. That separation, a profitable legacy core subsidizing an unproven growth bet, is unusual among AI infrastructure names and provides a floor under the equity that pure-play infrastructure competitors lacking Oracle's database franchise do not have.
The primary risk is the customer concentration detailed in Block F, compounded by a debt-funded capital structure that leaves limited room for the buildout's economics to disappoint before credit markets or equity investors demand a higher return for continued financing. Because Oracle has not disclosed customer-level RPO concentration, the market is pricing this risk on incomplete information, itself a reason for caution independent of how the underlying contracts ultimately perform.
The next earnings release, expected September 14, 2026, will be the first opportunity to observe whether Q1 fiscal 2027 revenue and cloud growth land within the guided range and whether management offers any incremental disclosure on customer concentration or financing terms. Capital markets activity around the announced $20 billion equity program, along with any movement in Oracle's credit default swap spreads, will likely move the stock as much as the earnings print itself between now and the next report.
Notices
This report has been produced exclusively for the internal use of Polaris Global Strategies Ltd. (PGS), a company incorporated in the British Virgin Islands (BVI), operating as a private vehicle for investment research and analysis on behalf of its partners. This document does not constitute an investment recommendation for any third party, whether individuals or legal entities, and must not be interpreted as such. PGS does not provide asset management, advisory or investment consulting services to any external client, has no client base, and offers no products or services to third parties. Any eventual access to this document by unauthorised parties does not confer validity as investment advice or recommendation. The information and analysis contained herein are based on public sources considered reliable, but PGS makes no warranty as to their completeness or accuracy. Investments in variable income instruments involve risks, including the possibility of total loss of invested capital. The opinions expressed reflect internal analytical judgement as of the publication date and are subject to change without notice.
This report was produced with the assistance of artificial intelligence. Generative model: Anthropic · Claude Sonnet 5 · version claude-sonnet-5, accessed via Claude mobile app. All figures were sourced from live web search within the same session and cross-checked against the anchoring protocol in Step 1; the model's role was data synthesis, formatting and drafting, not the origination of financial estimates. Human review by the internal analyst precedes publication.
Report Reference: PGS-ORCL-202608 · Date: August 31, 2026 · Internal Analyst: Polaris Global Strategies Ltd. · Ticker: ORCL · Exchange: NYSE · AI Model: Anthropic Claude Sonnet 5