Oracle
The Database Mega-Cap Reborn as an AI-Infrastructure Hyperscaler

Executive Summary

Oracle is the single most consequential repricing event in mega-cap infrastructure software since the early Azure inflection: a forty-year database compounder whose Q1 FY26 RPO disclosure converted the equity story overnight from "legacy on-prem franchise + ~$10B IaaS sideline" to "AI-infrastructure hyperscaler with a $638B order book."¹ The Q4 FY26 print on June 10, 2026 confirmed the trajectory: FY26 revenue of $67.4B (+17% YoY), OCI revenue of $18.1B (+77%), Q4 OCI growth of +93% YoY, and an RPO balance of $638B: up +363% year over year and a book-to-bill ratio of approximately 9.5x on full-year revenue.² The thesis collapses to a single execution question: can Oracle convert that backlog into recognized revenue at the cadence implied by the multi-year guide (Catz: $18B FY26 → $32B FY27 → $73B → $114B → $144B FY30), and can it absorb the capex required to deliver the contracted compute without permanently impairing the balance sheet?

We rate Oracle Buy with a 12-month price target of $235, derived from a segment-level DCF that underwrites OCI scaling to ~$90B+ by FY30 at a peak non-GAAP operating margin of ~45%, cross-checked against a NTM P/E comps screen versus hyperscaler-adjacent peers. The base case lands at $235 on a 7.75% WACC and 3.0% terminal growth; bull $310 underwrites faster Stargate / OpenAI ramp and a multiple re-rate to ~30x forward; bear $140 reflects OpenAI ramp slippage, capex creep into FY27E, and equity dilution from the $20B ATM + convertible-preferred program. Probability weights of 35 / 45 / 20 (Bull / Base / Bear) blend to a fair value of ~$242, sitting slightly above the base case and consistent with the modest bull skew the Stargate optionality warrants. Crucially, our $235 PT sits below the sell-side consensus median of ~$260, reflecting our greater discount for capex execution risk and the leverage tail (long-term debt $122.3B, net debt ~$90B, FCF FY26 of ($23.7B)).

Tactical: ORCL at $186.80 trades ~26% below our $235 12-month PT and ~46% below the late-2025 Stargate-driven peak of $345. Blended FV ~$242 on 35/45/20 weights. Rating Buy: the position pays you to own the only mega-cap with a contracted AI-compute book of this magnitude at ~23x forward, a meaningful discount to hyperscaler peers (MSFT ~32x, AMZN ~36x). Add into RPO conversion milestones and FY27 capex re-guide clarity; trim if OpenAI ramp materially slips or if the next financing tranche dilutes EPS beyond ~5%.

Investment Thesis Summary

Overall Rating: Buy · 12-month PT $235 (Blended FV ~$242)

We initiate Buy on Oracle with a $235 12-month price target. The base case underwrites OCI revenue of $32B in FY27 (+77% growth holding) on the way to Catz's $144B FY30 target, RPO conversion at a cadence consistent with the multi-year guide ($90B FY27 revenue contribution → $144B FY30), and non-GAAP operating margin held at ~43% through the capex absorption phase. At ~23x NTM EPS on $8.05 FY27 guide, Oracle trades at a structural discount to hyperscaler-adjacent peers, MSFT at ~32x, AMZN at ~36x, even CRM at ~27x, that we view as warranted by three real concerns the market is right to discount: (i) single-customer concentration in the RPO add (a majority of the $500B sequential RPO jump in FY26 came from the OpenAI / Stargate contract per secondary estimates), (ii) FCF burn through FY27 driven by capex outrunning operating cash flow, and (iii) the financing tail (ATM equity + convertible preferred) that compresses per-share metrics through the build cycle. The rating is not "ORCL deserves a hyperscaler multiple"; it is "the $90 discount per share already in the stock more than compensates for the execution risk."

The 35 / 45 / 20 weight reflects our view that the bull case is genuinely available, the contracted RPO is real, the OpenAI partnership is documented, the Stargate sites are under construction, but the bear is also live: OpenAI's own financing posture remains a non-trivial overhang, the FY27 capex envelope has not been re-guided, and the $20B ATM has been authorized but not fully drawn. The blended fair value of ~$242 sits modestly above the base PT of $235, capturing the Stargate optionality without overweighting it. We frame Oracle as the highest-conviction asymmetric setup in mega-cap infrastructure software: a 2.2x reward-to-risk (+65.9% bull vs –25.0% bear from spot) at the current multiple, with execution milestones (Q1 FY27 print, OpenAI 8-K filings, multicloud Database@AWS GA) all clustered in the next two quarters.

1 · Business Overview

What Oracle Is

Oracle reports four segments that map cleanly to a "legacy cash-cow + emerging hyperscaler" frame: Cloud Services & License Support (the largest line and the one that contains OCI plus the recurring license-support annuity from the on-prem installed base), Cloud License & On-Premise License (new-license sales, structurally in secular decline), Hardware (Exadata, Sun servers, networking, modest, stable, integrated with the database stack), and Services (consulting tied to ERP / database implementations). The FY26 mix is Cloud $34.0B (+39%), Software $24.5B (1%), Services $5.7B (+10%), and Hardware $3.1B (+5%)². The Cloud line, which now includes both OCI (IaaS / AI compute) and Oracle's Cloud Applications (Fusion ERP, NetSuite, HCM), is what re-rates the equity; the Software line is what funds the build until the cloud business earns its keep.

Q4 FY26 Segment Breakdown

Segment Growth Profile (FY26)

Inside Cloud Services & License Support

This is the franchise that anchors the thesis and the multiple. The segment contains two structurally different revenue streams. The first is OCI: Oracle Cloud Infrastructure, the IaaS / AI-compute business that reported $18.1B in FY26 (+77% YoY) with a quarterly acceleration that ran +55% (Q1) → +68% (Q2) → +84% (Q3) → +93% (Q4). OCI is what enterprise customers now buy when they want to run Oracle Database against contracted GPU capacity, and what OpenAI / Stargate is buying in size for inference and training. The second is Cloud Applications + license-support: Fusion ERP, NetSuite, Oracle HCM, plus the multi-decade license-support annuity on the legacy on-prem Oracle Database installed base. License support is the cleanest cash-flow stream in the portfolio: a sub-10% growth, ~90% gross-margin renewals book that compounds quietly and that funds the entire OCI build. The mix inside the segment is shifting, OCI was <20% of segment revenue two years ago and is now ~35%, but the license-support floor is what makes the equity story bankable.

Inside Cloud License & On-Premise License

The "Software" line at $24.5B FY26 (1%) is the structurally challenged piece of the portfolio and the part of the income statement most easily misread. The decline is not collapse; it is the predictable consequence of workload migration to OCI plus the multicloud Database@AWS / @Azure / @Google footprint. New perpetual license sales decline year over year; what offsets them inside the line is the residual license-support attached to those legacy contracts (~80%+ attach rate, materially sticky). The line tells you nothing about the AI thesis. It tells you something about the cash-flow profile: every dollar that comes out of this segment via license retirement is being replaced, at higher growth and eventually higher absolute revenue, inside the OCI book. The market will look through the segment-level decline as long as Cloud Services & License Support continues to grow at +30%+; a sudden acceleration of the on-prem decline would be the kind of customer-flight signal that would force a reassessment.

Inside Services & Hardware

The two smallest segments are best read as enablers of the larger ones. Services at $5.7B (+10%) is the consulting and implementation arm tied to Fusion ERP, NetSuite, and Oracle Health (the Cerner business acquired in 2022, now integrated as Oracle Health), growth tracks the Cloud Apps attach curve and the EHR rollouts inside Oracle Health. Services margin is structurally low and the franchise is not the multiple driver, but the segment is the proof-point that Cloud Apps revenue is being implemented at a customer-success cadence that justifies the recurring book. Hardware at $3.1B (+5%) is the Exadata / Sun / networking franchise, high-margin, engineered, and bundled with the database. The Exadata appliance specifically is the highest-performance Oracle Database deployment configuration; selling Exadata is effectively selling the database moat in physical form. The segment will not be the growth story, but it generates real margin and it reinforces the architectural lock-in that funds OCI adoption inside the existing customer base. We are also tracking the ~30,000-headcount reduction Oracle executed around March 2026 (~18% of staff), which redirects ~$8-10B of annual cash flow to AI data center investment and is the cleanest signal that management views OCI as the priority over the legacy services-and-people cost base.

2 · OCI & AI Infrastructure: Stargate, OpenAI, and the $638B RPO

The Oracle thesis collapses to a single question: does OCI convert the $638B remaining performance obligations book into recognized revenue at the cadence Catz disclosed on the Q1 FY26 call ($18B FY26 → $32B FY27 → $73B FY28 → $114B FY29 → $144B FY30)³, and does it absorb the capex required to deliver the contracted compute without permanently impairing operating margin or the balance sheet? OCI grew +77% in FY26 on a base that was $10.2B the prior year, and the quarterly cadence shows accelerating, not decelerating, growth: Q1 +55% → Q2 +68% → Q3 +84% → Q4 +93% YoY: a remarkable trajectory at this scale and the cleanest evidence that Oracle is not just absorbing AI demand but is structurally taking share from AWS / Azure / GCP at the contracted-compute layer. OCI exited FY26 with a run-rate well above $20B annualized; we model an FY27 exit closer to $32B in the base case.

The OpenAI Partnership & Stargate Economics

The single biggest piece of news in the FY26 catalyst path was the disclosure of the ~$300B / 5-year direct compute contract between Oracle and OpenAI⁴, which secondary estimates suggest contributed the majority of the $500B sequential RPO jump in Q1 FY26. The deal economics: roughly $60B/year peak compute, ~$30B/year cadence, with prepayments and customer-supplied GPU mechanics on a meaningful portion of the contract that insulate Oracle's FCF from a pure-spot customer slowdown. Stargate, the broader Oracle / OpenAI / SoftBank / G42 infrastructure consortium, anchors the supply side of that contract: 4.5 GW of Oracle-OpenAI partnership capacity, 7 GW total planned, with five new sites added in September 2025 and a >$400B three-year investment envelope announced. The Abilene, Texas site is the lead deployment; additional Stargate sites are under permitting / construction across the US, EU, and APAC. The combined effect of the OpenAI contract plus the broader Stargate book is what reprices Oracle from "infrastructure software" to "AI hyperscaler with contracted compute."

Multicloud: Database@AWS / @Azure / @Google

The second leg of the OCI story is the multicloud Database initiative, running Oracle Database natively inside AWS, Azure, and GCP data centers, with the customer billing relationship flowing through the cloud of origin. Oracle exited Q1 FY26 with 71 multicloud regions live across the three hyperscalers; multicloud Database revenue grew +1,529% YoY in Q1 FY26 and +404% in Q4 FY26: by far the fastest-growing business Oracle has ever reported. The strategic logic: enterprise customers who already have an AWS-first or Azure-first posture do not want to migrate workloads to OCI just to keep running Oracle Database, but they will pay Oracle for the Database running inside their existing hyperscaler. Multicloud monetizes the database moat without forcing the workload migration, a structurally cleaner motion than the OCI-only stance Oracle held until 2023. Database@AWS specifically launches in volume in late FY26 / early FY27 and is the cleanest swing factor on the FY28 OCI base case.

The $55.7B → ~$70B CapEx Framing

Oracle spent $55.7B on capex in FY26: up from $21.2B in FY25 and $7B in FY24, and materially above the $35B initial guide the company carried into the year⁵. The FY27E framing is ~$70B net, with gross capex meaningfully higher when customer prepayments ($20-25B) are added back through the cash-flow statement. This is the single biggest piece of execution risk in the equity story. Operating cash flow set a $32B FY26 record but does not come close to funding the build, the resulting FCF of ($23.7B) is the negative number the equity market has been pricing since the Q4 print. The funding plan: $43B of debt and $5B of equity raised in FY26, with management framing $45-50B more in calendar 2026 at roughly 50/50 debt / equity, anchored by a $20B at-the-market equity program and a convertible-preferred issuance. The capex absorbs starting late FY27 / FY28 in the base case as contracted RPO converts and depreciation on the FY26 layers matures; the bear case is that it does not, and the equity is diluted further to fund the gap.

3 · Financial Health

Quarterly & Annual Summary

Note on FY27E: our base case underwrites ~$79B in FY27 revenue (+17% YoY), ~12% below management's ~$90B guide. The haircut reflects our discount for RPO conversion timing and OCI capacity-stand-up cadence. The bull scenario reconciles to the $90B guide; see Bear vs Base reconciliation in the scenario section.

Q4 FY26 Print Recap & FY27 Guidance Posture

The Q4 FY26 print on June 10, 2026 delivered $19.2B in revenue (+21% YoY), with OCI growing +93% YoY and the headline RPO balance landing at $638B: broadly in line with the high end of the implied consensus range that built coming out of the Q3 call. Non-GAAP operating margin held at ~43%, broadly flat sequentially despite the AI mix and capex absorption. Non-GAAP EPS of $2.11 (+24% YoY) brought the FY26 print to $7.63 (+27% YoY) and supported the FY27 guide of $8.05 non-GAAP EPS. The notable disclosures inside the print: operating cash flow of $32B (a record), capex of $55.7B (vs $35B initial guide, a meaningful overrun), and an FY27 capex framing of ~$70B net with the gross figure higher when customer prepayments are added back. Catz reiterated the FY30 multi-year OCI guide of $144B, anchored by the contracted RPO. The market sold the print initially on the FCF print of ($23.7B) and the implied calendar 2026 financing tranche, and that sell-off is what produced the current $186.80 entry into a multi-year compounder setup.

Quarterly Seasonality Framework

Oracle's revenue has a recognizable but mild seasonal pattern tied to the legacy on-prem license sales cycle. Q4 (calendar Q2) is the largest revenue quarter, enterprise customers true up annual contracts and on-prem renewals concentrate, typically running 2-3 percentage points above the trailing-four-quarter run-rate. Q2 (calendar Q4) is the second-largest, also tied to enterprise budget true-ups. Q1 and Q3 are the cleanest reads on underlying OCI momentum because the legacy license noise is lowest in those periods; FY26's Q1 was the print that disclosed the headline $138B sequential RPO jump and effectively kicked off the re-rating, and FY26's Q3 was the print where OCI's +84% growth confirmed the trajectory was accelerating. The implication for our model: a beat on Q1 FY27 carries more signal weight than the same-magnitude beat at Q4 because the read-through is cleaner on OCI versus license mix.

FCF Bridge & CapEx Intensity

The free-cash-flow trajectory, $11.8B (FY24) → $5.8B (FY25) → ($23.7B) (FY26) → ~($15B) (FY27E) → ~$8B (FY28E): is the part of the financial profile that requires the most explanation, and the part the market is most actively pricing. The mechanical bridge: FY26 operating cash flow set a $32B record against $67.4B of revenue (a strong 47% conversion rate), but capex stepped from $21.2B (FY25) to $55.7B (FY26), a +163% capex increase against an OpCF increase of +53%. The arithmetic produces the headline FCF of ($23.7B): $32B OCF − $55.7B capex = ($23.7B) cash burn. FY27E carries a similar dynamic, OCF tracks toward ~$40B while capex lands ~$70B, producing FCF of approximately ($15B) and a second consecutive year of cash burn. The crossover comes in FY28-FY29 as RPO conversion lifts OCF materially (we model ~$55B) while gross capex begins to plateau (~$63B in FY28E), and FCF turns positive at roughly $8B. Bears argue the depreciation tail compresses OCF in FY28-FY29 just as gross capex plateaus, keeping FCF anchored near breakeven; bulls argue prepaid-customer-supplied-GPU mechanics on a portion of the OpenAI contract effectively shift capex off Oracle's balance sheet and accelerate the FCF inflection by a year.

Operating Leverage at ~43% Non-GAAP Margin

Non-GAAP operating margin has held at ~43-44% across FY24, FY25, and FY26E despite the AI mix shift toward OCI (lower segment margin than license support) and despite the $55.7B FY26 capex feeding the depreciation line. The architecture: license-support revenue continues to generate ~90% gross margin and provides a meaningful margin floor; OCI gross margin is structurally lower than license support (driven by GPU procurement and underutilized newly-stood-up capacity), but rises sharply with utilization; Cloud Apps (Fusion / NetSuite) carries SaaS-style margins in the 80s once revenue scales. The net is the ~43% blended non-GAAP operating margin that the company reports, with a modest dip to ~42% in FY27E as the largest tranche of capex hits depreciation, then recovery to ~43-45% by FY28-FY30 as OCI scale absorbs the fixed-cost base. Bears underwrite a more pronounced dip, closer to 40% in FY27E, on the view that under-utilized GPU capacity drags GM by 200-300bps; bulls underwrite a milder dip and an earlier recovery on the view that contracted RPO conversion fills capacity ahead of schedule.

Balance Sheet

Oracle ends Q4 FY26 with ~$31.9B in cash and marketable securities against ~$122.3B in long-term debt: a net debt position of ~$90B, the most levered balance sheet of any mega-cap in the comp set and the single biggest structural difference between Oracle and the hyperscaler peers. Credit ratings stand at BBB+ / Baa2, multiple notches below the AAA / Aaa stack at AAPL and MSFT and one notch below the A-rated paper at AMZN. The capex cycle materially stresses the credit profile, and not by accident: the funding plan announced in FY26 was $43B of new debt plus $5B of equity, and the calendar 2026 framing adds another $45-50B at roughly 50/50 debt / equity weighting. The $20B at-the-market equity program plus the convertible preferred issuance are the named tranches; the convertible structure is what limits near-term EPS dilution at the cost of optionality that resets at higher share counts. The cost-of-capital ladder matters here: every 25bps the front end of the curve sits at maturity adds roughly $250-300M of annualized interest expense to the marginal debt issuance, and a higher-for-longer regime would force a rebalance toward more equity. We model the base-case financing plan executing on the disclosed mix; the bear case includes meaningful dilution beyond 5% of share count from incremental ATM draws if RPO conversion runs slower than guided.

4 · Capital Allocation

Oracle's capital allocation in the current cycle is unambiguous and unapologetic: CapEx is the priority, capital return is the rounding error. The ~$55.7B FY26 capex commits the entirety of operating cash flow plus ~$24B of incremental external financing to AI infrastructure build-out. The dividend was held flat at $0.50/quarter ($2.00 annualized, ~1.1% yield) for the first time in over a decade, a deliberate signaling decision to communicate that management views the FY26-FY28 build cycle as worth pausing dividend growth to fund. Buybacks at ~$95M in FY26 are functionally zero (down from $13B in FY22). The framing on capital return is that Oracle is treating this build cycle the way Amazon treated its 2010-2015 AWS investment phase: maximum reinvestment, minimum return-of-capital, until the franchise crosses the FCF inflection. We model dividend held flat through FY28; buybacks remain <$1B/yr through the same window.

The Financing Tail: ATM & Convertible Preferred

The most consequential capital-allocation development of FY26 was the authorization of a $20B at-the-market equity program and a separately structured convertible preferred issuance, jointly framed as the equity component of the calendar 2026 funding plan. The ATM is the cleaner mechanic, Oracle can drip-feed equity into the market at prevailing prices, raising capital without a single dilutive event, but it does mark the equity at spot, meaning Oracle effectively underwrites the capex with the equity at $186 (instead of the $345 the stock saw at the Stargate-driven peak in late 2025). The convertible preferred is more interesting: it carries a fixed coupon (we model 4.5-5.5% based on current Baa2 paper pricing), converts above a strike that we expect to land 25-35% above spot, and limits near-term EPS dilution at the cost of a future conversion event. The total equity component sits in the $25-30B range across the ATM and converts; on a fully-converted basis we model 3-5% share-count creep over the next 18 months, the bear case carries that to 7-9% if RPO conversion underperforms and additional draws are required.

The CapEx-vs-Return-of-Capital Trade-off

The CapEx-first stance reflects where Oracle sits in its lifecycle: a 47-year-old database compounder with a once-in-a-generation reinvestment opportunity in front of it. At a $186 stock price and a ~23x NTM multiple, the buyback would be modestly accretive on a per-share basis, but every incremental $1B of capex directed at OCI capacity that gets absorbed at the segment's incremental gross margin compounds inside the business at a high-teens to low-20s unlevered return, meaningfully above Oracle's ~7-8% blended cost of capital. The math is the same logic that drives every hyperscaler-pivot capital-allocation cycle, and the proof-point is that the OpenAI deal economics (with prepayment mechanics on a portion of the contract) implicitly fund a meaningful share of the GPU capex through the customer rather than through Oracle's balance sheet. A material change in capital-allocation priority, a dividend acceleration, a buyback restart at >$5B/yr, a reduction in capex below FY26 actuals, would be the strongest signal that management views the OCI build as nearing maturity, which would be a different (and lower-multiple) equity story than the one we are underwriting.

Workforce Restructuring & Operating Cost Discipline

The ~30,000-headcount reduction Oracle executed around March 2026 (~18% of pre-reduction staff) is the operating-expense corollary of the capex priority. The action redirects an estimated $8-10B of annual cash flow from the legacy services / sales / G&A cost base toward AI data center investment and freed-up capacity in the FY27 operating-margin model. The cuts concentrated in services, legacy on-prem sales, and corporate functions, the parts of the cost structure that scale less than linearly with the OCI revenue base. The strategic logic is consistent with the Amazon and Microsoft analogues: when a mega-cap pivots toward a capital-intensive growth franchise, the operating-expense base on the legacy franchise needs to come down to fund the build without compressing operating margin. We do not model an additional headcount action of similar size in the next 18 months; a second action would be the signal that the FY27 revenue ramp is meaningfully softer than guided, and would be incrementally bearish for the equity even though it would mechanically support operating margin.

5 · Valuation & Comps

Peer Comparison

Why the Discount to Hyperscaler Peers

The headline number, ~23x NTM P/E vs MSFT at ~32x and AMZN at ~36x: is the central valuation question for Oracle. Three structural concerns explain (and arguably justify) the discount. First, customer concentration: a single OpenAI / Stargate counterparty contributes the majority of the $500B sequential RPO add in FY26, and the equity market correctly charges a concentration discount versus MSFT (where the OpenAI economics are similarly important but balanced against $295B+ of diversified non-AI revenue) or AMZN (where AWS revenue is distributed across hundreds of thousands of accounts). Second, balance sheet: Oracle's $122.3B long-term debt and ~$90B net debt vs MSFT's modest net cash and AMZN's net-neutral position represents real credit risk that the multiple should and does reflect; a 100-200bps move in the front end of the curve hits Oracle materially harder than the AAA-rated peers. Third, FCF profile: Oracle is the only mega-cap in the comp set running negative FCF (–$23.7B FY26 vs MSFT ~$70B, AMZN ~$50B+), even though the cash burn is funding genuinely contracted growth, the equity market discounts negative-FCF compounders at a structurally lower multiple than positive-FCF compounders.

Sell-Side Distribution & Our Below-Consensus PT

The sell-side dispersion on Oracle is meaningfully wider than for any other mega-cap we cover. The PT distribution: Barclays $180 · BofA $215 · Citi $235 · Morgan Stanley $245 · JPM $260 (median) · GS $280 · UBS $295 · Evercore $330. The bull-end positions ($300+) underwrite the Catz multi-year guide at full conviction and apply a hyperscaler-style multiple (28-30x forward) on the FY28E EPS; the bear-end positions ($180-215) discount the RPO conversion meaningfully and apply a software-conglomerate multiple (16-18x) on a more conservative FY27 EPS path. Our $235 PT sits ~10% below the sell-side median of $260, reflecting our greater discount for the capex execution tail and the financing dilution we model into FY27, but our blended FV of ~$242 (which carries the bull-case Stargate optionality at 35% weight) is materially closer to the sell-side median. The asymmetry that informs our Buy rating: even our below-median base PT implies +25.8% upside from spot, and the bull case at $310 (3% below UBS, modestly above GS) implies +65.9% upside if RPO conversion runs anywhere close to the multi-year guide.

EV / Sales & The Forward P/E Lens

Three multiples bracket the valuation argument. On NTM P/E, $186.80 ÷ $8.05 = 23.2x, the headline discount to hyperscaler peers and the most-watched single number. On EV / NTM Sales, ($544.5B + $90B net debt) ÷ $79B FY27E = 8.0x, broadly in line with the SaaS / hyperscaler comp set, and the multiple that arguably rebalances the picture once leverage is accounted for. On EV / FY30 Sales (which the bull case rests on), the same ~$635B EV against the implied $144B OCI alone (plus another ~$50B of non-OCI revenue) lands <3.5x, a hyperscaler-like figure if you accept the multi-year guide, a high figure if you do not. Our PT of $235 implies a 23-25x sustained NTM P/E on $9.50 FY28E EPS, a modest multiple expansion from spot, consistent with the structural discount and the FCF inflection coming in FY28.

Consensus

5b · DCF Model

The DCF projects each revenue stream independently across a 5-year explicit horizon, then converges to a consolidated terminal value. OCI carries the highest growth (modeling +77% FY27 → ~+128% FY28 on Stargate site activation → tapering to ~30% by FY30 as the base scales to $144B); Cloud Applications + License Support compound at low double-digits reflecting Fusion / NetSuite share gains plus the license-support annuity; On-Premise License declines low-single-digits as the secular migration to OCI continues. Margins follow segment: OCI operating margin trends from ~25% (FY26, underutilized GPU drag) toward ~45% by FY30 as scale absorbs fixed costs; License Support holds ~85%+; the consolidated non-GAAP operating margin trends from ~43% FY26 to ~45% FY30. Net debt of $90.4B (= $122.3B long-term debt − $31.9B cash) is subtracted from enterprise value; diluted shares of 2.915B are the divisor (we layer in 3% share creep from the ATM + convert in the base case, 7% in the bear).

Sensitivity: WACC × Terminal Growth

6 · Catalysts (Next 12 Months)

What the Q1 FY27 Print Actually Determines

The single most important calendar event in this 12-month catalyst path is the Q1 FY27 print in early September 2026. Three disclosures inside that print effectively reset the next-12-month framing of the stock. First, the FY27 capex re-guide: management's Q4 FY26 framing of "~$70B net" is the placeholder, but FY26 itself ran $20B+ above the initial $35B guide, the Q1 FY27 update is the cleanest tell on whether the $70B net is the ceiling (which would be incrementally bullish for FCF) or another stepping stone (which would push the FCF inflection right by a quarter or two). Second, the OCI revenue print: the FY26 Q4 cadence implied a run-rate around $25B+ exiting the year; a Q1 FY27 print materially above $7B (~+85% YoY) would confirm the acceleration thesis and could re-rate the multiple. Third, the RPO update: the $638B FY26 close is the new baseline, and Q1 typically adds 20-30% sequentially when major contracts close, the figure will reveal whether the OpenAI ramp is on schedule and whether additional Stargate contracting has materialized.

Stargate, OpenAI, and the Multicloud Tail

The partnership milestones drive equally important signal beyond the print itself. Stargate site activations: Abilene phase-2 commissioning, the five new sites added in September 2025 progressing through permitting and construction, and any new site announcements through the AI World conference, are the operational proof points that the contracted compute can be delivered on schedule. Any single-site delay does not break the thesis, but a pattern of permitting or power-procurement issues would compress the implied FY28 OCI revenue path materially. The OpenAI ramp into FY28 is more sensitive: the contract structure implies $30B/yr cadence with peaks toward $60B; OpenAI's own financing posture (cap-profit conversion, fundraising rounds, enterprise revenue trajectory) is now an upstream variable in Oracle's revenue model, and Microsoft 8-K filings on the OpenAI partnership remain the cleanest read on whether OpenAI is meeting its broader compute commitments across hyperscalers. Database@AWS GA in late FY26 / early FY27 is the third catalyst, initial customer wins and pricing disclosures will determine whether multicloud Database contributes the $3-5B FY28 run-rate the bull case requires.

Regulatory Schedule & The Financing Tail

The regulatory backdrop is less consequential for Oracle than it is for Microsoft or Amazon, Oracle has historically maintained a quieter posture on antitrust given the legacy database moat is not a recent acquisition, but two strands matter. First, the EU AI Act enters its General-Purpose AI compliance window in August 2026 with high-risk system rules phasing in through 2027; Oracle's exposure runs through the OpenAI contract economics rather than direct Oracle compliance, but a meaningful slowdown in EU enterprise AI adoption would compress the Stargate EU site economics. Second, the Federal AI procurement framework: Oracle has historically been a meaningful US-Federal customer through Oracle Database deployments, and the US government's AI-compute procurement direction (including specific Stargate-adjacent contracting) is a smaller-but-real catalyst path. On the financing tail, the pacing of $20B ATM equity draws and the timing of the convertible-preferred issuance are the technical-overhang catalysts that will create entry points through calendar 2026, every meaningful equity-raise tranche creates a price disconnect that long-duration holders can exploit.

7 · Risk Factors

8 · Technicals

Key Levels

Technical Setup in the Context of the Thesis

The technical picture maps cleanly to the fundamental setup, and that is exactly the value. The stock has corrected (37.73%) YTD from a $300 YE25 close, and sits ~46% below the 52-week high of $345.72: a textbook over-correction setup if the underlying thesis is intact. The base-case PT of $235 implies +25.8% upside from spot, and the bull PT of $310 implies +65.9%, far higher implied returns than the ~12% / ~25% setups that comparable mega-cap compounders typically offer at a single-quarter entry. An RSI of ~42 reads approaching-oversold for a fundamental compounder of this scale, particularly in the context of an EMA stack that has rolled bearish (21 < 50 < 200) through the FY26 capex-fear correction. The technical setup is consistent with a stock that has been over-sold into a confirmable thesis, the opposite of the over-bought setups that typically mark the top of a mega-cap re-rating cycle.

Level Confluence & Position Sizing

The chart levels cluster in a way that matches the scenario PTs. The $200 level is the first piece of psychological resistance and the post-Q4 print pivot, a clean break above $200 on volume would be the confirmation that the market is willing to look through the FCF burn to the FY28 inflection, and would open the path toward the base-case $235. $235 sits below the sell-side median of $260, providing a clean technical target with sell-side support if the print path executes; $260-$280 is the second meaningful zone (the bulk of the sell-side PT range) and is where the bull-case framing of $310 becomes operationally available. On the downside, $160 is the tactical stop: a break below would imply a meaningful Q1 FY27 disappointment on capex re-guide or OCI growth deceleration, and would mark the entry into the bear-case $140 zone (where the structural floor at the 52-week low of $134.57 sits as the absolute downside). Sizing-wise, the implied tactical R/R from spot to base PT vs the $160 stop is favorable (~$48 reward vs ~$27 risk, a ~1.8x ratio); to the bull PT vs the same stop, the ratio rises to ~4.6x. The 35% probability weight on the bull case is what carries the expected-value math; long-duration positions can build at current levels with the conviction that the asymmetric setup is genuinely available rather than constructed by aggressive assumptions.

★ Interactive Tools

Ask the Thesis AI-assisted checking…

Describe an Oracle scenario in natural language; the assistant returns a structured impact analysis against this dashboard's segment-level DCF, OCI / Stargate economics, RPO conversion model, and OpenAI contract framework. Powered by Claude via a Cloudflare Worker proxy.

9 · Glossary