Oracle & the Build-Out

Oracle has, in four quarters, converted itself from a steady cloud-share gainer into the owner of the largest contracted backlog in enterprise software — $638 billion of remaining performance obligations, up $85 billion in a single quarter. That backlog is real and largely pre-sold. What the market repriced on 10 June was not its existence but its cost: fiscal-2026 free cash flow of −$23.7 billion on $55.7 billion of capex, a guide to ~$70 billion of net-cash capex in fiscal-2027, and ~$40 billion of fresh debt and equity — including roughly $20 billion of new equity — to fund it. Strip the build-out to its parts and the two lenses disagree: on revenue the stock looks undemanding (the price implies roughly 18% annual growth against management’s +31% CAGR), but on earnings and cash it sits at fair value, because an AI-infrastructure dollar carries 30–40% gross margins and heavy capital intensity. After a 48% derating from $345, Faircurve reads Oracle as fairly valued — base case ~$180, a touch above spot — with genuine two-sided risk: a bull case of $245 if the backlog converts on plan, a bear of $115 if it does not.

Three signals will settle the debate. First, RPO conversion velocity — only 12% of the $638 billion backlog converts within twelve months and 34% in 13–36 months; an acceleration in those percentages is the cleanest evidence the backlog is becoming revenue. Second, the free-cash-flow inflection — whether capital intensity peaks in fiscal-2027 as guided and cash generation turns the corner toward fiscal-2029. Third, the financing path — the size, mix and dilution of the ~$40 billion fiscal-2027 raise, and whether Oracle holds its investment-grade rating through the build. A faster conversion and an earlier cash turn push the read toward the bull case; a capex overrun or a financing strain confirms the bear.

The transcript arc is a step-change in two acts. Through fiscal-2025 the story is steady cloud-share gain — OCI growth in the 50s, backlog drifting from $99B to $138B. Then, in the first quarter of fiscal-2026, remaining performance obligations quadruple to $455B on the Stargate-era frontier-AI contracts, and the call’s centre of gravity shifts from demand to delivery: ‘capex / data-centre build-out’ and ‘power & capacity’ mentions rise while management begins to concede, explicitly, that gross margin steps down during the ramp and that AI-infrastructure margins settle at 30–40%. The same eight quarters that produced the largest backlog in software also reveal its cost: a build-out large enough to turn the cash-flow statement negative, and a leadership change — Safra Catz to Executive Vice Chair, co-CEOs Clay Magouyrk and Mike Sicilia, a new CFO — that visibly shifted the call from visionary monologue to capital discipline.

Across eight quarters the backlog went from a $138 billion rounding error against the hyperscalers to $638 billion — the single largest contracted order book in enterprise software — and the disclosed OCI ramp hardened into specific multi-year figures ($18B → $144B over fiscal-2026 to fiscal-2030). That cadence is the bull case in one data series. The same record carries the bear’s warning: capex guidance was raised at every call, free cash flow went deeply negative, and management moved from ‘margins improve with scale’ to an explicit admission that infrastructure gross margin steps down during the build and settles at 30–40%. The arc is one of accelerating conviction financed by accelerating capital intensity.

At 22x forward earnings and 18x EV/EBITDA Oracle screens at a premium to the peer average (20x and 15x) while growing revenue at roughly twice the cohort’s pace (+32% versus +15%). On growth-adjusted terms the premium is modest — and on fiscal-2028, as the backlog converts, the multiple compresses to 16.5x earnings for ~46% growth. The peers do not form a clean growth-to-multiple line (Salesforce 12x on +11%, Alphabet 24x on +19%, Amazon’s margin distorted by retail), so Faircurve anchors Oracle by judgement rather than extrapolation: a deserved premium to the mature large-caps for its growth and contracted backlog, disciplined by the lower margins, negative free cash flow and balance-sheet leverage the higher-quality peers do not carry. The premium is to the cohort — but it is a steep discount to where Oracle traded eight months ago.

Oracle’s discounted-cash-flow value is unusually sensitive, because free cash flow is deeply negative through the build-out and inflects only in fiscal-2029 on Faircurve’s model. A seven-year ramp — crediting the full consensus revenue path and capital intensity normalising from ~80% of revenue toward ~26% — returns roughly $167 at a 10.8% cost of capital and 3.0% terminal growth, close to spot. FMP’s standard levered-DCF, anchored to today’s negative cash flow, returns just $31 and is not meaningful for a company mid-build. The DCF therefore carries only a 20% weight; the case rests on the multiple the market assigns the contracted backlog.

At a $637 billion enterprise value, a 10% annual return over five years requires enterprise value to compound to roughly $1.03 trillion by fiscal-2032. Held to a normalised ~5x EV/Sales — below today’s 7x, reflecting the lower-margin, capital-intensive infrastructure mix — that implies about $205 billion of revenue, a ~18% compound annual growth rate off the fiscal-2027 base (about 20% off fiscal-2026). Consensus has revenue growing +32% in fiscal-2027 and +46% in fiscal-2028, and management guides a +31% revenue CAGR through fiscal-2030 — all well above the ~18% the price requires. On the revenue lens, in other words, Oracle looks undemanding. The catch is that revenue is a poor proxy for value here: an AI-infrastructure dollar carries 30–40% gross margins and heavy capital intensity, so it is worth less than a legacy-software dollar — which is why the earnings-and-cash lens lands the blend at fair value rather than cheap. The market is not disbelieving the backlog; it is pricing the cost of building it.

Faircurve Equities · Independent Single-Name Research · ORCL · 11 Jun 2026 · Issue No. 018

Sources: Financial Modeling Prep (quote, profile, statements, financial-estimates, key-metrics, treasury rates, DCF). Oracle Q4 FY2026 press release & Form 8-K (10 Jun 2026). Last 8 quarterly earnings-call transcripts. Peer data: MSFT, GOOGL, AMZN, CRM (FMP).

Methodology: Forward FY+1 (calendarised ~2027) peer comparison. Consensus-anchored base case. 40/40/20 blend of Forward P/E, Forward EV/EBITDA and a Faircurve DCF. CAPM cost of capital. Reverse-DCF on implied revenue growth. EBITDA = operating income + D&A.

Disclaimer: Research, not investment advice. Not a recommendation to buy or sell any security. The author held no position in ORCL at publication.