Oracle Corporation

▲ Bull Case

• ◆RPO Conversion Is Real and Large: $553B in signed contracts (33% due within 12 months) is contracted demand. At $89B FY2027 management guidance, Oracle guides 55% of near-term RPO in one fiscal year. OCI at $33B+ annualized by FY2027 would be the world's fastest cloud infrastructure scaling event.
• ◆FCF Recovery = Re-Rating: When CapEx normalizes in FY2028 ($15-18B vs. $45B peak), Oracle generates $25-30B+ in FCF on $100B+ revenue. That's $625-750B FCF value—17-41% above current market cap before debt paydown. The market is not pricing this today.
• ◆Database Moat Funds the Wait: The database business generates $18-20B/year in operating income at near-zero incremental capex. This cash flow cushion prevents the bear case from becoming severe, even if OCI normalization takes 2 extra years.

▼ Bear Case

• ◆CapEx May Not Normalize: Every large hyperscaler (AWS, Azure, GCP) that undertook major buildouts found reasons to keep CapEx high. Oracle's thesis requires 60% CapEx drop within 2 years—without precedent. If CapEx stays $30-40B through FY2028, FCF remains negative and thesis defers 2+ years.
• ◆OCI is Still #4 at 3% Share: Despite 84% growth, OCI processes a fraction of AWS/Azure AI workloads. If NVIDIA Rubin architecture resets GPU cluster designs in 2026-2027, Oracle's Blackwell-optimized infrastructure could require expensive re-tooling before current data centers earn back their cost.
• ◆$114B Debt + No FCF = Limited Margin for Error: Any meaningful revenue shortfall vs. $90B FY2027 guidance leaves Oracle with elevated CapEx, negative FCF, $5-6B annual interest expense, and multiple de-rating. Math from $185 allows only modest execution risk before downside materializes.

“The central debate is not whether OCI is growing—it's whether the $553B RPO is real and whether FCF inflects on schedule. Consensus at $185-$260 is divided: Bulls (Wedbush, ~$225-260 targets) treat RPO as a floor—$553B in signed hyperscale AI contracts is unprecedented demand confirmation, and OCI's GPU architecture advantage is defensible. At $185 with 32% revenue growth, the stock is cheap. Bears (~$160-180) argue CapEx normalization is unproven; Oracle has never built data centers at $40-50B/year pace, and commissioning delays could manifest in 2026-2027. Cerner is a $28B distraction. The debate resolves Q4 FY2026 (June 9, 2026): OCI must post >$5B quarterly and management must guide FY2027 at $90B+ to maintain credibility.”

• ◆Q4 FY2026 earnings (June 9): OCI >$5B quarterly; FY2027 guidance ≥$90B; CapEx FY2027 <$35B guided → +15-25% if confirmed; -15-20% if miss
• ◆RPO >$600B print (June-August 2026): Confirms AI demand expanding beyond current signed base → +5-10%
• ◆FY2027 CapEx guidance (June 2026): First explicit guidance on normalization path critical for FCF thesis → +10-20% if guided $25-35B
• ◆Oracle Health major win or no Cerner impairment (FY2027): De-risks $15-20B goodwill on books → +5-8%
• ◆First FCF-positive quarter (Q1-Q2 FY2027): Confirms inflection thesis materializing → +10-15%
• ◆New hyperscale AI customer beyond OpenAI/Meta (ongoing): Expands RPO addressable base → +3-8% per announcement

• ◆OCI commissioning delays causing Q4 miss (20-25% probability): -15-25% stock impact; signals data center ramp slower than guided
• ◆CapEx stays elevated through FY2028 (25-30% probability): −10-15%; FCF recovery delayed 2+ years; thesis timeline breaks
• ◆Cerner goodwill impairment $5-15B (15-20% probability): -8-15%; management credibility impaired; potential GAAP EPS shock
• ◆Leverage distress / ratings downgrade (5-8% probability): -20-35%; refinancing costs surge; equity dilution risk
• ◆Ellison health event or succession without clear technical replacement (low acute, high long-term): Indeterminate; OCI strategy could lose direction
• ◆NVIDIA Rubin GPU architecture shift (10-15% probability): -10-20%; Oracle's Blackwell-optimized data centers could require costly re-tooling