Oracle Says Its Expiring AI Contracts Are Renewing 20% Higher

Oracle reported fiscal Q1 revenue of $19.3 billion, up about 30% year over year, while cloud infrastructure revenue accelerated to 121% growth and $7.4 billion. GPU capacity renewing during the quarter was renewed or resold at a 20% premium, supporting the case for longer-lived AI infrastructure economics; Oracle also booked more than $30 billion in new AI contracts, lifting remaining performance obligations to $664 billion. However, quarterly capex surged to $28.5 billion, free cash flow was negative $5.4 billion, and full-year capex is projected at $90 billion-$95 billion, with gross margins still declining as new data centers ramp. Management raised adjusted EPS guidance by $0.05 to $8.10, but the article remains cautious pending evidence that higher renewal pricing translates into expanding margins.

Analysis

The key underwriting question is no longer demand but capital intensity and financing durability. Oracle is effectively converting a software-like earnings base into a utility-style infrastructure model before proving steady-state returns; the resulting mix shift can cap the earnings multiple even if revenue growth remains strong. Equity issuance alongside deeply negative free cash flow creates a recurring overhang: unless operating cash conversion improves materially over the next two quarters, incremental buildout is likely to compete with buybacks, dividends, and per-share earnings growth.

The renewal datapoint is more useful for suppliers than for Oracle shareholders in the near term. It confirms acute AI-compute scarcity and supports demand visibility for NVDA, Broadcom (AVGO), Arista (ANET), and power/cooling vendors, while Oracle retains the execution risk of commissioning sites, securing power, and filling capacity at acceptable margins. Longer term, stronger economics for older accelerators could reduce replacement cadence, a modest negative for GPU unit demand in 2028-29, but that is outweighed today by new-cluster procurement.

Consensus may be too focused on the reported earnings multiple and insufficiently focused on the balance-sheet duration mismatch: multiyear contracted revenue does not equal cash collection or credit-quality certainty. The bullish case is falsified if gross margin fails to recover as new facilities mature, if capex remains elevated beyond fiscal 2027, or if RPO converts into slower-than-expected cloud revenue. Conversely, two consecutive quarters of gross-margin expansion with capex held within plan would justify rerating ORCL materially from current depressed sentiment.

Key Decisions for Investors

• Maintain a neutral-to-underweight ORCL stance for the next 1-3 months; do not treat the lower headline P/E as a value signal until quarterly operating cash flow turns positive and cloud gross margin expands sequentially. Thesis fails on demonstrated margin recovery plus capex discipline.
• Express the near-term infrastructure demand signal via a 3-6 month pair: long NVDA or AVGO versus short ORCL in equal beta-adjusted dollars. Suppliers monetize backlog immediately, while ORCL bears deployment and financing risk; exit if ORCL reports material gross-margin expansion or NVDA demand guidance weakens.
• Use any 10%+ ORCL rebound ahead of the next earnings release to add downside hedges rather than chase the equity; a put spread is preferable to outright shorting given high short-covering risk after the stock's large drawdown. Size only after confirming option implied volatility and strike liquidity.
• Set a quarterly watch trigger: upgrade ORCL only if cloud gross margin rises sequentially, capex guidance does not increase, and share count stabilizes. Failure on any two of those metrics keeps the stock structurally constrained over the next 6-18 months.