Oracle reports Q2 earnings on December 10. Ahead of the print, the cost to insure its debt over the next five years has surged to levels not seen since 2009, as concerns mount over elevated debt and aggressive AI-related capex.
Last quarter, Oracle raised $18 billion in debt to accelerate large-scale data center construction. Oracle is investing heavily today to secure future growth, boasting a forward three-year annualized revenue growth rate of 29.7% and EPS growth of 21.6%, while trading at a forward P/E of 29.0x. Only seven Russell 1000 companies offer higher growth projections at a lower valuation multiple.
Fueling its AI ambitions comes at a cost—draining liquidity and squeezing free cash flow. StarMine SmartEstimate projects Q2 free cash flow (FCF) at -$5.3 billion, versus consensus at -$5.1 billion, signaling a negative Predicted Surprise of -2.9%. If realized, this would be the lowest quarterly FCF in Oracle’s 34-year history as a public company. FCF is expected to remain negative until 2030. For example, the 2027 SmartEstimate sits at -$11.6 billion, nearly double the consensus of -$6.9 billion, suggesting further downward revisions are likely.
Credit risk is flashing red: Oracle’s 5-year CDS spread hit a 16-year high at 128 bps, the widest among major hyperscalers. While hyperscalers broadly plan >50% capex growth over the next 12 months, Oracle sits at 63%. The concern instead is the scale of capex relative to cash flow and revenue. Oracle’s capex-to-sales ratio is 58%, and capex-to-CFO a staggering 147%—both the highest by far in the peer group. Crucially, Oracle is the only hyperscaler unable to fund its next 12 months of capex solely from expected operating cash flow.
To frame credit risk, we leverage the StarMine’s Credit Combined Risk (CCR) model, a multi-factor model that integrates three standalone models – SmartRatios, Structural, and Text Mining to assess default and bankruptcy risk. The CCR model assigns Oracle a StarMine Implied Rating of BB+, six notches below its current S&P A+ rating—a gap that historically signals potential downgrades.
Our research finds that the StarMine Implied Rating is predictive of future revisions and that when agency ratings diverge significantly from implied ratings, agencies move toward the implied rating 4–5x more often than away.
Prior Editions of the StarMine Spotlight:
#002: StarMine Flags Strong Analyst Sentiment Ahead of Nvidia Earnings – Nov 18. 2025
#001: Meta’s Spending Spree – StarMine Signals $105 Billion CapEx by 2026 – Nov 5. 2025
