Why Oracle Stock Slumped on Thursday

Microsoft's fiscal Q2 2026 results showed revenue of $81.3 billion (up 17% YoY) and adjusted EPS of $4.14 (up 24%), but $37.5 billion of capex and commentary that AI demand exceeds capacity sent MSFT shares down over 12% and pressured cloud/AI peers. Oracle, which reported December-quarter revenue of $16.1 billion (up 14%) and adjusted EPS of $2.26 (up 54%), disclosed a remaining performance obligation of $523 billion (up 438%) and spent $35.5 billion on capex; its shares fell as much as 6.5% and have dropped over 50% from peak. Investors are concerned that elevated capex for AI infrastructure may not translate into commensurate revenue, even as Oracle's backlog and a ~31x earnings valuation make the stock look more interesting to some investors.

Analysis

Market structure: The immediate winners are AI-infrastructure suppliers (NVDA, switchgear/power suppliers and hyperscaler-capex contractors) as demand still exceeds compute capacity; near-term losers are large cloud spenders whose margins will be watched (MSFT) and legacy software with uncertain cloud monetization (some peers to ORCL). Oracle’s 438% RPO jump to $523B and a lower forward multiple (~31x) imply substantial locked-in demand, which shifts pricing power toward suppliers of specialized AI compute and data-center operators (DLR, EQIX) while pressuring free-cash-flow sensitive incumbents.

Risk assessment: Tail risks include regulatory curbs on high-end GPUs or export controls (6–18 month horizon), large capex write-downs or RPO impairments if client projects stall, and a demand pull-forward that cools after capacity ramps. In days–weeks expect elevated IV and technical-driven volatility; over 3–12 months watch conversion of RPO to recognized revenue and free-cash-flow recovery as the key binary; hidden dependency—RPO is backlog, not cash, so conversion timing is critical.

Trade implications: Favor asymmetric, time-boxed exposure to ORCL’s backlog via 9–15 month call spreads or small LEAPs (2–3% portfolio) with 15% hard stop and 35–45% upside target if cloud revenue/FCF inflects in two quarters; run a relative-value pair long ORCL vs short MSFT (equal-dollar) to express backlog monetization vs capex risk. Use put spreads on MSFT or short-dated strangles to monetize elevated IV; rotate 2–4% from broad AI-hyperscaler longs into data-center REITs and enterprise software names with clear RPO-to-cash paths.

Contrarian angles: The consensus fixates on headline capex; it underestimates structural pricing power if capacity stays tight—this historically mirrors early AWS/Azure scale-ups where heavy capex preceded outsized margins (3–36 months). The sell-off may be overdone for ORCL if management converts even 10–20% of RPO annually into revenue with stable margins; unintended consequence—accelerated capex could raise energy/commodity demand and benefit power/commodity-sensitive names.

Key Decisions for Investors

• Establish a 2–3% portfolio long in ORCL via a 9–15 month call debit spread (buy 30–40 delta LEAP call / sell a higher strike) sized to risk a 15% max loss; add another 1–2% only if RPO-to-revenue conversion improves for two consecutive quarters (target 35–45% upside in 12–18 months).
• Implement a pair trade: go equal-dollar long ORCL (1.5–2% portfolio) and short MSFT (1–1.5%) to express valuation/revenue-conversion divergence; hedge MSFT leg with a 3-month 10–15 delta put spread if implied vol spikes above historical 60-day mean by >25%.
• Reduce gross exposure to hyperscaler/AI infra longs (MSFT, AMZN-sized positions) by 2–4% and redeploy into data-center REITs (DLR, EQIX; 1–2% combined) and cash-flowing enterprise software names with clear backlog conversion; reassess after two earnings cycles (next 60–120 days).
• Monitor regulatory and supply catalysts explicitly: track (1) NVDA/ASML export-control headlines and chip allocation notices, and (2) ORCL quarterly RPO conversion/margin metrics at earnings over the next 60 days—if either shows adverse change >10% vs consensus, tighten stops or unwind positions.