Oracle Stock Has Lost More Than Half Its Value in 6 Months. It May Finally Be Time to Buy.

Oracle reported fiscal Q3 revenue of $17.2 billion, up 22% YoY, and EPS of $1.79, up 21% YoY; cloud infrastructure revenue was $4.9 billion, up 84% YoY. Remaining performance obligations jumped to $553 billion, up 325% YoY, driven by large AI contracts many of which involve customer prepayments or customer-supplied GPUs. Management guided roughly $50 billion of capex for fiscal 2026 to support AI data-center build-out, creating near-term financing and free-cash-flow risk. Shares have fallen >50% over six months and now trade at ~19x forward P/E (vs ~21x market), which the author views as an attractive entry for long-term investors despite short-term risks.

Analysis

Oracle’s pivot to contract structures where customers prepay or supply GPUs materially changes who carries capex and inventory risk: Oracle buys optionality on long, sticky services revenue while shifting immediate hardware spend and allocation risk to customers. That reduces near-term balance-sheet strain but creates a new execution dependency on third-party GPU allocation, procurement cadence, and the timing of revenue recognition — conversion of booked obligations into free cash flow will largely drive re-rating over the next 12–36 months.

The most valuable second-order beneficiary is the GPU supplier base; sustained, large-scale commitments from a non-hyperscaler increase bargaining power and could lift wholesale GPU pricing and lead-times, reinforcing Nvidia’s pricing power and benefiting upstream memory/packaging suppliers. Conversely, customers supplying their own hardware reduces Oracle’s margin capture on equipment, compressing gross margins on the infrastructure business even as service revenue expands — a subtle margin mix-shift investors often underweight when looking solely at RPO growth.

Key tail risks are procurement shocks (supply shortages or sudden GPU price spikes), client credit or demand reversals before conversion, and tightening financing markets if management issues incremental debt to fill any residual capex gap. Near-term catalysts to watch are: monthly/quarterly conversion rates of booked obligations into recognized revenue, reported gross margins on cloud infrastructure, and any disclosed GPU allocation agreements with major suppliers; these will move the stock on a 3–12 month cadence.

From a framing perspective, this is a classic asymmetry: downside is operational/financing noise over 6–18 months; upside is multi-year margin and valuation re-rating if conversion executes. That suggests option structures and pair trades that cap near-term pain while leaving substantial upside if backlog monetization proves durable.

Key Decisions for Investors

• Buy ORCL 18–30 month LEAPS call spread (debit): size 3–6% of tech allocation. Rationale: limits premium paid while capturing a 2x+ payoff if RPO converts and cloud margins stabilize within 24 months; max loss = premium, target return >100% if shares re-rate on demonstrated conversion.
• Buy ORCL shares with a 6–9 month protective put (collar-lite): allocate 4–8% position, buy 20% OTM puts to cap downside through the next heavy capex reporting window. Rationale: protects from execution/financing shocks over the next 6–12 months while preserving upside to a longer-term re-rating.
• Directional play on GPU tightness: buy NVDA 3–6 month out-of-the-money call skew (small-sized, 1–2% portfolio): high gamma bet that supplier tightness and cloud contract pull-through force higher ASPs or upside beats. Reward asymmetric vs premium paid; keep exposure small due to crowdedness.
• Pair trade for capture of execution delta: long ORCL equity (larger size) funded by a smaller short INTC position or purchase of short-dated INTC puts (net neutral cash). Rationale: express conviction that Oracle captures services revenue from contracted AI workloads while Intel will lag in capturing accelerated GPU-driven spend; this reduces pure market beta and isolates the infrastructure win/loss outcome.