One of America's biggest investors Michael Burry is betting against Oracle, the company whose 'biggest customer' is Sam Altman's OpenAI; says: I do not like how it is ...

Michael Burry has taken bearish positions against Oracle, including put options and direct shorting over the past six months, citing concerns about the company’s strategic positioning and costly data‑centre buildout. Oracle carries roughly $95 billion of debt and faces large capital demands after signing a $300 billion, five‑year cloud deal with OpenAI in September 2025 (compute purchases to begin in 2027); the stock has been volatile—rising 36% in one session in September but finishing 2025 about 40% below that peak. Burry’s actions, alongside broader worries over rising capex and execution risk, may pressure sentiment and derivative pricing around Oracle shares.

Analysis

Market structure: Oracle’s aggressive data‑centre buildout and $95bn debt load redistribute winners toward hyperscale compute buyers (OpenAI, large enterprise cloud consumers) and data‑centre suppliers, while harming ORCL equity and credit if utilization lags. Competitors (MSFT, GOOGL) gain relative pricing power because their diversified software/ads franchises insulate them from pure‑play infra margin compression; expect ORCL cloud ASPs to face downward pressure if supply outstrips demand after 2027. Cross‑asset: anticipate ORCL credit spreads and 5y CDS to widen by 50–150bps in stress scenarios; equity implied vol to trade 30–60% vs. peers.

Risk assessment: Tail risks include (1) OpenAI pullback or contract renegotiation pre‑2027, (2) a 200–300bps jump in funding costs raising annual interest expense ~ $1.9–2.85bn (for 200–300bps on $95bn), and (3) regulatory limits on AI services. Immediate (days): knee‑jerk volatility; short‑term (weeks/months): guidance and capex updates; long‑term (years): debt servicing and asset write‑downs. Hidden dependency: Oracle’s free cash flow and covenant sensitivity are non‑linear to utilization; a 10% lower-than-forecast utilization can swing cloud margins negative.

Trade implications: Direct: tactical 9–12 month put positions on ORCL (15–20% OTM, delta ~0.25) sized 2–3% NAV; buy 5y ORCL CDS protection sized to hedge potential $1bn–$3bn downside to enterprise value. Pair: short ORCL / long MSFT or GOOGL equal‑dollar for 6–12 months to capture secular software resiliency. Options: consider ORCL calendar or diagonal spreads to monetize elevated IV while limiting carry; for NVDA a small (1–2% NAV) 6–9 month 10–25% OTM put spread as a volatility hedge.

Contrarian angles: Consensus underestimates Oracle’s sticky database and licensing cash flow; if utilization ramps faster than expected or OpenAI accelerates purchases pre‑2027, ORCL equity could rebound materially (>30%). Historical parallel: early hyperscaler capex cycles (AWS) showed multi‑year pain then margin recovery; therefore short conviction should be conditional—avoid permanent capital exposure. Unintended risks for shorts include retail squeeze/cover dynamics and potential strategic asset sales or buybacks that could re‑rate ORCL.

Key Decisions for Investors

• Establish a 2–3% NAV directional bearish position on ORCL using 9–12 month puts 15–20% OTM (target payoff if ORCL falls ≥25%); set stop‑loss at 15% adverse move and trim on realized gains at 25% decline.
• Buy 5‑year ORCL CDS protection sized to hedge ~$50–100m notional exposure (cost acceptable if spreads <150bps); reduce or liquidate protection if CDS narrows by >75bps or Oracle issues repurchase guidance improving FCF by >10% YoY.
• Implement a pair trade: short ORCL equal‑dollar long MSFT (or GOOGL) ~3% NAV each for 6–12 months to capture relative resilience; unwind if ORCL/MSFT spread compresses to historical mean or ORCL issues >20% capex reduction guidance.
• Allocate 1–2% NAV to an NVDA downside hedge via a 6–9 month 10–25% OTM put spread to limit premium spent while protecting portfolio from a sentiment-driven AI sell‑off; close if NVDA implied vol falls >40% from current levels.
• Wait for ORCL quarterly release (next 30–45 days) and only increase bearish exposure if management raises capex guidance >20% YoY or projects cloud utilization <60% in near term; otherwise cap conviction and prefer credit protection over naked equity shorts.