Oracle's AI compute opportunity is clearer, but cost and funding constraints still limit a rating upgrade

Overview

This report updates its assessment of the benefits and costs of Oracle's large-scale transition from traditional software and cloud database businesses to AI computing infrastructure/GPUaaS. 3Q26 showed improvements in demand, cRPO, and capacity delivery, but GPUaaS expansion requires enormous capex, lease commitments, and diversified financing, making gross margin, free cash flow, leverage, and credit risk the key constraints.

Core views

Morgan Stanley is more optimistic on Oracle's execution capabilities and AI compute demand, believing that higher cRPO coverage, accelerated cloud database growth, and improved GPUaaS capacity delivery support higher long-term earnings assumptions. But at the stock level, risk/reward remains balanced: upside comes from AI compute demand, OCI capacity ramp, and expense discipline; downside comes from gross margin pressure, data center delivery delays, customer concentration, negative free cash flow, and funding uncertainty.

Analysis framework

The report splits Oracle into core business earnings power and GPUaaS earnings power, and uses base, bull, and bear scenarios for SOTP valuation; it also combines cRPO coverage, capex/MW, leases not yet commenced, funding mix, free cash flow path, and credit spreads to assess equity and credit implications.

Methodology notes

• Valuation methodsSOTP EPS multiple

  sum-of-the-parts valuation

  The report separately estimates FY30 earnings power for the core business and GPUaaS, then applies multiples to NGAAP EPS and discounts back to FY28 to derive bull, base, and bear target prices.

• scenario_analysisBull/Base/Bear risk reward

  risk-reward scenario analysis

  The base-case target price is $207, the bull-case scenario is $304, and the bear-case scenario is $75, used to assess the asymmetric impact on the stock price of successful AI compute expansion, execution deviations, and lease cost pressure.

• credit_analysisFunding and leverage model

  funding and leverage stress test

  The report estimates cash needs before FY28, bond issuance, chip asset financing, equity financing, and lease obligations, and translates them into total adjusted debt, leverage, and CDS/bond investment views.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• US.ORCL equity

  primary covered asset

  Strengths

  Strong AI compute demand, significant improvement in cRPO, better execution in OCI and cloud database, and solid expense discipline.

  Weaknesses

  GPUaaS gross margin remains low and volatile, free cash flow is negative before FY28, and funding and lease obligations are increasing significantly.

  Comparison

  The report believes that if Oracle's execution suffers further setbacks, the market may trade it more in line with hyperscaler peers on a GAAP basis.

  Risks

  Data center delivery delays, customer concentration, gross margin decline, rising funding costs, and insufficient achievability of EPS targets.

• ORCL 5Y CDS / credit

  credit trading view

  Strengths

  The CY26 funding plan may keep bond spreads range-bound in the short term.

  Weaknesses

  Long-term funding transparency is limited, and total adjusted debt and leverage may rise significantly.

  Comparison

  The report believes the valuation comparison should be more aligned with lower-quality IG telecom, cable, and media companies rather than HY BB tech companies.

  Risks

  CDS could widen to or beyond about 200bp, and Moody's rating still carries residual risk.

• US.CRWV

  reference company for AI compute infrastructure

  Strengths

  A comparable observation target for compute expansion and facility delivery risk.

  Weaknesses

  Its latest results show that facility delivery delays may pressure gross margin and capacity ramp.

  Comparison

  Used to map execution and gross margin risks during Oracle's GPUaaS buildout cycle.

  Risks

  Industry-wide data center delivery delays or gross margin pressure could affect Oracle's valuation narrative.

Key data

• RatingEqual-weightRating maintained, with current risk/reward considered balanced.
• Price target$207.00Cut from $213.00, based on 25x FY28e base-case NGAAP EPS of $8.28.
• Current price$187.50Closing price on April 22, 2026.
• Implied upside10.4%Based on the $207 price target versus the current price of $187.50.
• Bull-case scenario$304.00Based on 30x FY30 SOTP NGAAP EPS of $11.97, discounted at 8.7% WACC.
• Bear-case scenario$75.00Based on 18x FY28 core business GAAP EPS of $4.16, including part of the lease expense pressure.
• FY30 MS Case SOTP EPS$11.97Of which core earnings power is $9.07 and GPUaaS contributes $2.90.
• cRPO coverage81%3Q26 cRPO rose to about $66bn, accounting for about 81% of estimated NTM revenue.
• 10GW capex estimate约$315bn-$325bnRoughly estimated at about $32.5mn/MW.
• Leases not yet commenced$261bnUsed to assess data center shell capacity lock-in and future lease burden.
• Peak total adjusted debt at end-FY28约$380bnUnder the funding scenario, total adjusted leverage is about 5.0x.
• Free cash flowExpected to be negative before FY28Under the MS Case, about -$64.2bn in FY27 and about -$65.9bn in FY28, turning positive in FY29.

Impact & implications

For equity investors, Oracle's AI compute transformation offers greater long-term earnings upside, but valuation support will be weakened by negative free cash flow before FY28 and higher funding needs; for credit investors, insufficient funding transparency, rising leverage, and expanding lease obligations make ORCL credit risk more prominent, and the report recommends maintaining a cautious credit view.

Risks

• GPUaaS capacity ramp may fall short of expectations or data center shell delivery may be delayed.
• GPUaaS and OCI gross margins may be weaker than expected, dragging on FY27-FY28 EPS.
• More than 50% of RPO comes from one large AI customer, creating elevated customer concentration and contract renegotiation risk.
• Free cash flow remains negative before FY28, weakening valuation support.
• Funding needs are enormous, and the mix of debt, equity, and asset-backed financing may bring dilution, interest expense, and rating pressure.
• Leases not yet commenced and long-term capacity commitments may create excess capacity risk if demand slows.
• If layoffs or restructuring affect sales execution, growth in the non-AI core business may weaken.
• Open-source models and customers building their own compute capacity may reduce long-term external GPUaaS demand.

What to watch

• Whether the delivery timetable for Abilene Phase 2 and other data center shells continues to be pushed back.
• Oracle's subsequent lease disclosures, especially how the $261bn of leases not yet commenced translates into capacity and cost.
• Whether EPS guidance improves in step with the FY27 revenue increase, to validate GPUaaS margins.
• Whether actual GPUaaS gross margin can remain stable within the guided 30%-40% range.
• Whether the improvement in cRPO coverage reflects true incremental demand or changes in contract timing with a large AI customer.
• The actual size and cost of bonds, chip asset financing, and equity financing.
• Whether FY27-FY28 free cash flow burn is close to the MS Case assumptions.
• Whether the core cloud database, multi-cloud database, and SaaS businesses continue to accelerate.