Oracle Corporation

01 / The central investment debate

Oracle combines enterprise databases and applications with a rapidly expanding cloud-infrastructure business. The bull thesis is that contracted demand and software relationships support years of profitable capacity expansion. The bear thesis is that too much of the resulting value accrues to lenders, landlords, hardware suppliers and newly issued shares before existing shareholders receive cash. This is a capital-return question as much as a technology question.

Q1 FY2027 revenue reached $19.35bn, up 30%. Infrastructure-as-a-service revenue was $7.39bn, up 121%, while cloud applications grew 10%. Current FY2027 guidance is at least $90bn revenue and $8.10 adjusted EPS; accessible consensus is $90.5bn and $8.14 respectively. Near-term estimates already require strong growth.[S1][S7]

Remaining performance obligations, or RPO, total $664bn. Only 13% is expected to become revenue within twelve months: approximately $86.3bn. Backlog is neither cash in the bank nor a profit guarantee; delivery costs, timing and customer credit still matter.[S3]

Oracle’s October 2025 analyst-day aspiration was $225bn revenue and $21 adjusted EPS in FY2030. Our base FY2030 estimates are lower—$183.9bn and $13.66. The old long-term targets are a useful bull-case reference, not a substitute for updated financing assumptions or fully refreshed September 2026 guidance.[S15]

02 / Competitive position

Oracle does not have to displace every hyperscaler to grow. Database availability across multiple clouds can monetize its installed base even when customers use another provider’s infrastructure. Enterprise applications also provide a route into cloud workloads. Conversely, supplying large GPU clusters is not automatically a software-like moat: price, power, utilization, hardware refresh and access to capital can determine the return.[S1][S5]

Sources: Oracle, Microsoft and Amazon releases; Alphabet results reporting; current consensus snapshots. Cloud definitions, fiscal periods and EPS adjustments differ. Oracle’s infrastructure revenue starts from a much smaller base than AWS’s $42.2bn quarter.[S1][S8][S9][S10][S11][S12][S14]

*The Amazon and Alphabet multiples are not clean normalized valuation comparisons. Their earnings include large investment-revaluation gains; Amazon alone disclosed $53.4bn of Q2 pretax other income, primarily from Anthropic. A low screen P/E can be misleading for competitors too. Microsoft is a cleaner direction-of-travel comparison, but still has a different mix and funding capacity.[S9][S14]

My assessment: Oracle has credible distribution and database advantages. What is not yet established is whether its marginal AI-capacity investments earn superior full-life cash returns. Revenue growth ranks do not answer that question.

03 / Earnings are not distributable cash

Q1 operating cash flow was $23.103bn, including $11.363bn of financing-like customer prepayments. Subtracting $28.499bn gross cash capex gives −$5.396bn free cash flow. Excluding those prepayments first gives a more demanding core-cash screen of −$16.759bn. That screen is not normalized annual FCF: ordinary working capital is also seasonal.[S2]

Avoid the double-count: Oracle’s $17.966bn “net cash outlay” for capex already reflects customer funding and capex-financing movements. Deducting that from reported OCF would count the prepayment benefit twice. On a trailing-twelve-month basis, conventional OCF less gross capex is approximately −$28.7bn.[S2][S4]

The quote implies approximately $415.9bn common market value. Financial debt, including finance leases, is $134.5bn; unrestricted cash and securities are $37.1bn, leaving $97.4bn net financial debt. Our current fully diluted enterprise-value proxy is $522.4bn after allowing for award dilution and mandatory preferred conversion. The completed $20bn gross ATM share sale is already reflected; it is not unused financing capacity.[S2][S3][S6][M1]

Beyond $34.6bn of existing operating-lease liabilities, there are $288bn of uncommenced nominal lease commitments, with 15–19-year terms. Spreading that over seventeen years suggests roughly $17bn of additional annual rent once operating—an illustration, not the actual payment schedule. The model ramps incremental rent explicitly; it does not also deduct the entire nominal commitment from enterprise value.[S3]

The forward adjusted earnings yield is 5.9%, versus a 1.45% indicated common dividend yield. Q1 buybacks were zero. Neither the earnings yield nor borrowed-funded distributions are additional independent return components. Our FY27 after-SBC earnings proxy is about $6.83 per share, making the same price approximately 20.1× earnings after recurring stock compensation, rather than 17.0× adjusted earnings.[S1][S3]

Margin quality also matters: Q1 GAAP operating margin was 34.8%, adjusted operating margin 42.1%, and a direct-cost gross-margin proxy about 60.0%, versus 67.3% a year earlier. The gross-margin proxy excludes separately reported intangible amortization. A mix shift and lower overhead can support operating margin while gross margin falls; neither metric proves cash returns after capex.[S2]

04 / Our five-year operating scenarios

The exit is approximately September 2031, valued on estimated FY2032 forward earnings. This preserves the same forward-earnings convention as today’s FY2027 P/E. Dividends accumulate as cash at zero interest. Scenarios include new borrowing, additional common issuance, net award dilution, preferred conversion and zero buybacks; they do not assume today’s share count stays fixed.

The base case grows IaaS from an assumed $40bn in FY2027 to $169.4bn in FY2032; SaaS grows roughly 8–10% annually while legacy software is broadly flat. Cash capex peaks at $100bn in FY2028 and declines to $65bn by FY2031–32. All operating rent is included before projected profit. These are substantial execution assumptions, not mechanical extrapolation of today’s 121% growth rate.

After the completed ATM, the base financing model still requires $28.5bn of additional common equity, with financial debt peaking near $192bn and ending diluted shares around 3.463bn. This is our funding-gap estimate, not an announced Oracle issuance plan. Existing maturities must remain refinanceable. The bear case is not a worst-case loss bound; the bull case requires both better economics and a higher valuation multiple.

05 / The base-case cash path

FY2027 revenue and EPS are guidance anchors; the revenue split, cash flow and later years are modeled. Cash OCF is approximated as adjusted income before preferred distributions plus depreciation, less ordinary working-capital use. Financing-like customer inflows then enter once and unwind later. Core FCF excluding those flows becomes positive in FY2030 and reaches about $48.2bn in FY2032. Management has not committed to a specific FCF-positive date.[S5]

A separate owner-cash DCF produces $82.97 per share at 10.5% WACC and 3% mature growth. At 9–11% WACC with 3% growth, the range is roughly $71–130. The model includes an FY2027 residual stub, detailed FY2028–32 cash flows and a five-year fade to FY2037; about 78% of enterprise value comes from the terminal period.

This DCF treats recurring SBC as a cash-equivalent labor cost, includes new finance-lease assets as reinvestment, keeps operating rent in margins, and uses today’s fully converted diluted shares. It therefore does not charge future award dilution again. Mature adjusted operating margin is 36%, SBC 4.5% of revenue, and incremental ROIC 12%; 3% growth requires reinvesting 25% of mature after-tax operating earnings.

Why the valuations disagree: the P/E case assumes future buyers pay 20× adjusted earnings, including optimism about growth beyond FY2032. The DCF imposes a stricter mature cash-conversion regime after paying for the buildout. These are alternative valuation assumptions—not two independent confirmations of the same fair value. I would not average them to manufacture a reassuring target.

The DCF also deducts a $15.955bn proxy for opening financing-like prepayments, based on known inflows rather than an audited remaining balance. Full elimination of that conservative deduction would add only about $5.17 per share. More material uncertainties are refresh capex, annual lease payments, useful lives and sustainable capacity returns.

06 / The 20% hurdle—and the PEG trap

With $10 of cumulative cash dividends, the shares need to end near $332.29. Required earnings vary with the terminal multiple:

Base FY2032 EPS of $18.74 needs approximately 17.7× at exit to meet 20% from today’s quote. At the chosen 20× base exit, the mathematical entry ceiling is $154.68. At today’s unchanged 16.98× forward multiple, it falls to $131.95. Those are conditional return hurdles, not independently supported buying prices.

Forward PEG is approximately 0.91× on a normalized one-year basis. The calculation uses 16.98× P/E divided by 18.7% EPS growth. The growth base is inferred from June’s $8.05 guidance and management’s stated 18% growth excluding Ampere/Bloom one-offs: $8.05 ÷ 1.18 ≈ $6.82. Updated $8.10 guidance then implies about 18.7% growth.[S1][S4]

Using unnormalized FY2026 adjusted EPS of $7.63 instead yields just 6.2% growth and a 2.76× PEG. Our five-year base EPS growth of 18.3% gives approximately 0.93×. Neither is a universal “correct PEG”: one depends on an inferred one-off adjustment; the other on our forecast. PEG ignores reinvestment, long leases and funding risk, and should not override the cash model.[S4]

07 / What would confirm or break the thesis?

Delivery and unit economics. Keep FY2027 revenue at or above the $90bn guidance floor and adjusted EPS near $8.10. Monitor actual RPO conversion rather than the headline balance. Q1’s 850MW delivery is tangible progress, but future projects still require power, completed facilities and paying customers. I would require capacity-cohort returns above a roughly 12% capital-return hurdle after rent, power and refresh spending—not merely post-build EBITDA conversion.[S1]

Capital discipline. FY2027 gross capex above the $90–95bn range, net cash outlay beyond $70bn without offsetting economics, or equity funding materially above our $28.5bn additional base assumption would require a new valuation. The base also needs core FCF to turn positive by FY2030. Persistent operating margins below roughly 35% after startup capacity ramps would undermine the modeled profit recovery.[S5]

Financing and customer concentration. Long-lived lease commitments can outlast a particular hardware cycle or customer funding cycle. Reported September pressure on $18bn of external data-center project loans is relevant to ecosystem financing, but those project loans are not automatically additional Oracle corporate debt. No precise customer-default probability or contract-level recovery value is assumed here.[S13]

Near-term catalysts. The October 2026 investor day could clarify long-run margins, leases and financing. The next results were indicated for 14 December 2026 on the earnings call, subject to change. Evidence that capacity can be delivered with less Oracle-funded capital would be more useful than another large backlog announcement.[S5]

08 / Conclusion and limitations

The attraction is real: moderate earnings expectations can coexist with a very large cloud opportunity. The unresolved issue is whether growth earns enough after-financing cash return per existing share. A lower purchase price reduces the burden of proof, but does not repair weak contracts or excessive reinvestment. The research stance remains watchlist until cash economics strengthen or valuation compensates for the funding risk.

20% CAGR means a compounded five-year result, not a positive 20% every year. USD returns are not automatically Australian-dollar returns; taxes, transaction costs and exchange rates are excluded. Dividends, refinancing access and modeled exit multiples are not guaranteed.

Evidence confidence: high for filed financial statements; moderate for guidance and consensus; low-to-moderate for long-run scenarios. Underwriting status: the earnings case is modeled, but the capital-return gate is unresolved. Missing diligence includes capacity-cohort cash returns, annual uncommenced-lease schedules, maintenance capex, customer-specific credit/cancellation terms and exact prepayment runoff. Future borrowing and equity costs, interest timing and depreciation are stylized, not a full GAAP three-statement forecast. This is independent public-source research, not audited or personalized financial advice.