Financebotresearch desk研究台

ORCL › analyze

ORCL · Analyze

Technology

A portfolio-specific passage was removed from the public build.


Snapshot

Field Value
Price (Jul 22, 2026 close) $126.29
Shares out / Market cap ~2.88B / ~$364B
Net debt $98.3B (total debt $156.2B)
Enterprise value ~$462B
Sector / mapping Technology (enterprise software + cloud infra)
FY (ends May 31) FY2026 reported Jun 10, 2026; Q1 FY2027 due Sep 10, 2026
FY26 revenue / growth $67.36B / +17%
FY26 GAAP / non-GAAP EPS $5.94 / $7.63
Fwd P/E (non-GAAP FY27 $8.05) ~15.7x
EV/Rev (FY26 / fwd FY27 ~$90B) 6.9x / ~5.1x
Dividend / yield ~$2.00/yr / ~1.6%
FY26 free cash flow –$23.7B (the crux)
RPO $638B (+363% y/y)

Fundamentals

Metric FY2024 FY2025 FY2026 Read
Revenue ~$53B $57.4B $67.36B Accelerating (+17.4%)
Net income (GAAP) ~$10.5B $12.44B $17.09B +37%
Operating margin ~30% 31.5% 33.3% Holding (efficiency offsets GM decline)
Gross margin ~71% 70.5% 65.8% –4.7pts — datacenter ramp; guided lower FY27
Operating cash flow ~$18B $20.8B $32.0B +54% — genuinely strong
Capex (gross reported) ~$7B $21.2B ~$55.7B Explosive; net outlay ~$48B after prepayments
Free cash flow ~+$11B ~–$0.4B –$23.7B Collapsed into deep deficit
FCF / share ~+$4 ~$0 –$8.29
Total debt ~$88B $104.1B $156.2B +$52B in one year
Net debt / EBITDA ~2.8x 3.4x 3.1x (Total debt/EBITDA 4.9x)
Interest coverage ~5.5x 5.0x 4.9x Eroding as debt compounds
Debt / assets ~55% 61.8% 59.7% High
Shares outstanding ~2.76B 2.81B 2.88B Rising — buyback era over
Tangible book / share neg –$14.79 –$8.45 Negative equity ex-goodwill ($62B)
ROIC ~15% 14.0% 12.1% Declining as capital base balloons

The capex/FCF trajectory is the whole story. OCF is excellent and rising (+54% to $32B), but Oracle spends it — and far more — into the ground: gross capex ~$56B FY26, guided ~$90–95B reported ($70B net cash outlay) FY27. FCF was –$23.7B FY26 and, against ~$40–45B FY27 OCF vs $70B net capex, likely deepens to roughly –$45B to –$50B FY27. Currently converts none of its earnings to owner cash — a capital-intensity bet dressed as a software stock.

Debt is the second story. Total debt +$52B to $156.2B. Net debt $98.3B, net-debt/EBITDA 3.1x — still IG-ish, but interest coverage thinning (4.9x, interest expense $4.6B and climbing). Guided ~$40B additional FY27 financing (debt + a $20B ATM equity program); no further debt in CY2026. The $20B ATM equity is dilutive — with shares already drifting up (2.81B→2.88B), the buyback engine is thrown in reverse.

Data gaps: (1) roic FCF field unreliable (see note); (2) no clean debt-maturity schedule via these tools — pull from 10-K; (3) segment OCI margin/ROIC is management-asserted ("high-20s project ROIC," "30–40% OCI margins"), not yet independently verifiable — consolidated ROIC is 12% and falling.


Moat

Quant base: ROIC 12.1% (down from 14%), gross margin 65.8% (down from 70.5%), op margin 33.3% (up). Margin compression is deliberate — high-margin database/apps annuity diluted by lower-margin, capital-heavy OCI. Blended returns keep sliding until (unless) OCI reaches promised steady-state economics.

Revenue-stream map — three different businesses: 1. Legacy database + apps (the annuity): Cloud Apps $4.1B/qtr (+10%), database +29%, multi-cloud +404%. Durable, wide-moat core — ERP/database switching costs are evergreen; multi-cloud tie-ups (Azure/AWS/GCP hosting Oracle DB) extend the annuity. Justifies a large fraction of EV alone. 2. OCI / AI infra (the growth + the risk): +93% to $5.8B/qtr, 97.5% GPU utilization, $67B AI-infra contracts signed in Q4. A capacity-and-execution moat, not a durable one — rests on being fast/cheap/competent at datacenters in a market flooded with capital (AWS, MSFT, Google, CoreWeave, neoclouds). "Massively higher demand than supply" describes a current shortage, not structural advantage. 3. Cerner / health: VA EHR rollout progressing (14 medical centers), OPM Fusion HCM win — steady, not the driver.

Adversarial stress-test: On OCI, a rival needn't out-innovate — just add capacity. When GPU supply loosens, OCI's pricing power and "snapped up on renewal" dynamic soften, testing the 30–40% margin promise. The cost-pass-through-clause + "cloud is more than a rack" defense is credible but a services moat, not a technology one. On the core database, the generational threat: if AI-native data platforms become where enterprises inference against data, Oracle's "data already in Oracle DB" edge erodes over 5–10yr. Management is defending (agent-memory, data-security, vector features).

Concentration as moat-vs-risk: OpenAI (~$300B/5yr, 4.5GW) is both why RPO is $638B and the single largest risk. Not a moat — a customer-financing arrangement with a cash-burning counterparty mid-pivot on self-build. Mitigant (real & important): majority of new AI contracts are now bring-your-own-hardware or prepaid ($75B BYO/prepaid), shifting capital + residual-value risk to customers at equal/better margins — materially de-risks the model vs six months ago. Does not de-risk the demand concentration.

Evergreen rating: Core database/apps = HIGH (8/10); OCI/AI-infra = MEDIUM-LOW, unproven (5/10), high-variance. Blended, the moat is widening in breadth but thinning in quality — trading a fortress annuity's returns for a capital-hungry growth call.


Valuation

Model / multiple Input Output Note
Fwd P/E (non-GAAP) $126.29 / $8.05 FY27 15.7x Cheap for +34% rev / +18% EPS guide
Trailing P/E (GAAP) $126.29 / $5.94 21.3x
EV / Revenue (fwd) $462B / ~$90B 5.1x Below hyperscaler-adjacent peers on growth
EV / EBITDA (fwd) $462B / ~$40B ~11.5x
EV / FCF n/m (FCF –$24B) negative Useless here
P/B $364B / $43B equity 8.4x Equity thin, tangible book negative
Bogle expected return ~1.6% yield + ~18% EPS growth ± P/E ~15–19%/yr if multiple holds Growth-dominated
DDM / DYT ~1.6% yield, ~15% div growth Not an anchor Dividend is a rounding error here

Fair-value range: ~$95 – $155, midpoint ~$125 → trading essentially at fair value. - Bear (~$75–95): OCI margins disappoint, OpenAI restructures/delays, dilution proceeds — ~10–11x FY27 EPS. - Base (~$113–135): Backlog converts on schedule, OCI holds 30%+ margins — 14–16x FY27 EPS. - Bull (~$150–170 near-term, optionality to $250+): If FY2030 targets (31% rev CAGR → ~$200B rev, 28% EPS CAGR → ~$20 EPS) become visible, 18–20x supports the high analyst PTs.

Reverse-DCF / "what must happen": At $462B EV the market discounts a large share of the $638B RPO converting at target margins. But the backlog is back-end-loaded — only 12% recognizes in 12 months (~$77B), 34% in months 13–36, ~54% (~$345B) beyond 3 years, much OpenAI-linked and contingent on ~1GW/quarter delivery. The price isn't demanding heroics on multiple (15.7x fwd is reasonable) — it demands heroics on execution and counterparty solvency, financed with $40B+/yr fresh debt + equity, for cash flows that don't turn positive until the capex wave crests (~FY2028+).


Sentiment

  • Consensus: Buy (≈37 buy / 5 hold / 1 sell). Avg PT ~$248–262, high $400, low $110 — ~50% above spot; the Street is emphatically bullish, signaling how much execution the market is pricing out. Barclays → $202.
  • Q4 FY26 (Jun 10): Record — revenue $19.2B (+21%), OCI +93% to $5.8B, Cloud Apps +10%, non-GAAP EPS $2.11 (+24%). RPO $638B (+363%), +$85B in the quarter; $67B AI-infra bookings; 4 customers >$8B each (some diversification). FY27 guide: revenue +34% cc, non-GAAP EPS $8.05 (+18%); Q1 FY27 revenue +27–29%. LT targets reconfirmed: 31% rev CAGR / 28% EPS CAGR through FY2030.
  • Leadership change (material): Safra Catz → Executive Vice Chair (Sep 2025); Clay Magouyrk + Mike Sicilia now co-CEOs; Hilary Maxson CFO (Apr 6 2026). Ellison remains Chairman/CTO, 40.3% / 1.16B shares — still controls the company. Tone aggressively confident; new CFO deliberately disciplined ("investment-grade credit rating," new "net cash outlay for capex" disclosure) — pre-empting the FCF/leverage criticism.
  • Stargate/OpenAI: Mixed but net-intact. Broader Stargate expanded (5 new sites; OpenAI reconfirmed the 4.5GW deal). But a ~600MW Abilene expansion scrapped (Mar 2026), and OpenAI reportedly pivoting toward third-party compute leasing over self-build — genuine edge-wobble. Oracle site progress real (Abilene 42% delivered). Net: relationship bigger, concentration bigger, counterparty visibly straining.
  • Debt/dilution: $30B raised early CY26; $40B more planned FY27 incl a $20B ATM equity (dilutive).
  • Class action: Active — Kessler Topaz securities-fraud suit; leads appointed Apr 27, 2026; class period Jun 12 – Dec 16, 2025, alleging Oracle overstated capex-to-revenue conversion speed. Ongoing overhang.
  • Insider/institutional: Insiders ~40.7% (Ellison the bulk — no sell-down signal); institutions ~41.9% (Vanguard 4.5%, BlackRock 3.8%); retail ~17%. No red-flag insider exodus.

Tensions / Debate

  • Fundamentals vs Sentiment/Moat: Fundamentals flashing red (negative/deepening FCF, +$52B debt, dilution, ROIC falling); Sentiment/Moat point to a once-in-a-decade demand capture with $638B contracted visibility. Both true — a timing disagreement: bear owns FY26–FY28 (cash bleed, leverage, execution), bull owns FY2028+ (margin harvest). Fair value reflects an honest standoff.
  • Valuation vs Valuation: On earnings/growth (15.7x fwd for 34% growth) it's cheap; on cash (negative FCF, negative tangible book) it's uninvestable by classic value standards.
  • Real de-risking: The BYO/prepaid shift ($75B) genuinely lowers capital-at-risk vs the prior thesis — the single biggest improvement since 2026-07-14.
  • Trigger partly fired: Prior thesis named "dilutive raise" as a break trigger. A $20B ATM equity issuance is now official guidance — planned, not distressed, but it counts. Flag, don't panic.

Verdict — Conviction [7]/10 (unchanged; a high-variance 7)

The core thesis strengthened on bull metrics — RPO $638B (vs $523B), OCI +93% (vs +68%), both blowing past the "<40% for 2 quarters" break floor — and the BYO/prepaid shift materially de-risks the capital model. Simultaneously the bear case strengthened: FCF –$23.7B heading toward ~–$50B, debt $156B, a $20B dilutive equity program now guided, ROIC falling to 12%, an active fraud class action. They offset. At ~$126 the stock sits at the midpoint of a wide $95–155 fair-value range — priced honestly as a coin-flip between flawless execution and a capex/leverage trap.

Key risks: (1) Customer concentration/counterparty — OpenAI dominates the backlog and is itself cash-burning, mid-pivot; a restructuring is the biggest single thesis-killer. (2) FCF/leverage spiral — FY27 burns ~$50B; sub-30–40% OCI margins or delivery slips make the funding treadmill expensive, threatening the IG rating. (3) Dilution — $20B ATM real, more could follow. (4) Margin/moat quality — ROIC and gross margin both falling; high-return OCI economics asserted, not proven. (5) Litigation overhang.

A portfolio-specific passage was removed from the public build.

Thesis-break (updated): OCI/cloud-infra growth <40% for 2 consecutive quarters (now +93%, miles away); OpenAI materially restructures/delays/defaults on Stargate (partial wobble already visible — monitor); a distressed or larger-than-guided equity raise / debt-market shutout; loss of investment-grade credit rating (new — the line the CFO is defending); BYO/prepaid mix reverses, pushing capital risk back onto Oracle.

Bottom line: a great annuity business making a heroic, debt-and-dilution-financed bet on a single AI counterparty. At fair value, held small, a defensible [7] — earned on backlog and execution promise, not on the cash it generates today, which is deeply negative and getting worse before it gets better.