Mizuho reiterates Oracle stock rating on strong cloud growth

Oracle reported Q4 revenue of $19.2 billion, up 21% year over year and above the $19.1 billion consensus, while non-GAAP EPS of $2.11 also beat expectations. Infrastructure-as-a-Service revenue surged 93% year over year to $5.8 billion, and remaining performance obligations jumped 363% to $638 billion, supporting management’s $90 billion fiscal 2027 revenue target and higher EPS guidance of $8.05. Analysts remained broadly constructive, with multiple firms raising price targets, though the stock was down nearly 15% over the past week.

Analysis

The cleanest read-through is that Oracle is no longer being priced as a software vendor but as a power-constrained infrastructure utility with software economics. The implication for winners is broader than ORCL: GPU server OEMs, networking vendors, and data-center power/cooling suppliers should keep benefiting as long as hyperscale capex remains supply-limited rather than demand-limited. The likely loser set is less obvious: traditional enterprise software peers may face valuation compression if investors start underwriting a new benchmark where growth must be paired with massive capex commitments to stay credible.

The biggest second-order risk is margin durability, not revenue. A business can grow into a gargantuan backlog and still disappoint if incremental returns on deployed capital fall below the market’s current enthusiasm; that becomes a 6-12 month story as capex cycles through the income statement. If investors begin to focus on free-cash-flow conversion rather than bookings, the stock can de-rate sharply even with intact top-line momentum. That makes the next catalyst sequence critical: any delay in deployment, change in cloud mix, or evidence of lower-than-expected utilization would hit multiple expansion first and fundamentals later.

Consensus appears to be underestimating how much of the upside is now already in the guidance architecture. The market is likely extrapolating the backlog as if it were near-certain monetization, but the real question is whether the company can translate it without sacrificing mid-cycle returns. In that sense, the move may be modestly overdone near term even if the strategic thesis remains intact over 2-3 years. The best contrarian setup is to own the beneficiaries of the buildout rather than chasing the highest-multiple beneficiary directly.

Key Decisions for Investors

• Maintain a tactical long ORCL only on pullbacks toward the low $190s; upside remains credible if execution stays clean, but risk/reward is no longer asymmetric after the rerating.
• Prefer a basket long in the picks-and-shovels layer: long SMCI / ANET / VRT over ORCL for the next 3-6 months, as these names monetize the capex cycle with less execution risk and faster earnings translation.
• Pair trade: short a software peer basket with slower growth and lower capex intensity against long ORCL or the AI infrastructure basket; thesis is that the market will reward visible AI monetization over generic enterprise software multiples.
• Buy ORCL downside protection into the next earnings cycle via put spreads, targeting a 3-6 month horizon; catalyst risk is any margin or capex-efficiency disappointment after the current optimism.
• For investors already long ORCL, trim 20-30% into strength and redeploy into infrastructure beneficiaries; the expected return profile is better in adjacent suppliers than in the main narrative stock.