Oracle Stock Is Down 60% From Its High. Here's Why I'm Still Not Buying.

Oracle Cloud Infrastructure revenue surged 121% year over year to $7.4 billion in fiscal 2027 Q1, while remaining performance obligations reached $664 billion, underscoring strong AI-compute demand. However, the company is planning $95 billion of capex, generated negative $5.4 billion of quarterly free cash flow, and had $125 billion of long-term debt plus $288 billion in unstarted data-center lease commitments. With shares nearly 60% below their September 2025 peak, the article argues Oracle's leveraged AI infrastructure buildout poses materially greater downside risk than that of larger hyperscalers, particularly if Project Jupiter is delayed or AI demand weakens.

Analysis

ORCL’s equity problem is not simply leverage; it is the mismatch between long-dated, concentrated contracted demand and near-term, largely irreversible infrastructure funding. If capacity commissioning slips, the company can experience negative operating leverage: interest and lease obligations begin before utilization revenue, while depreciation and power/network costs are already embedded. This makes quarterly free-cash-flow guidance, rather than headline RPO, the key valuation catalyst over the next 1-3 months.

The second-order readthrough is selectively positive for better-capitalized AI infrastructure buyers. MSFT, AMZN, GOOG, and META can fund incremental capacity from internally generated cash and spread fixed data-center costs across broader cloud, advertising, and consumer ecosystems; any Oracle execution setback could redirect enterprise AI workloads toward these platforms. NVDA is less exposed to an Oracle-specific funding disruption than investors may assume, but an industrywide shift from speculative build commitments to financed-and-powered deployments would pressure near-term GPU shipment timing and supplier multiples.

Consensus may be over-indexing to an all-or-nothing AI demand outcome. A delay or renegotiation at a major campus would not necessarily impair end demand, but it would expose that Oracle’s backlog has lower cash-conversion quality than revenue-equivalent backlog at peers. The critical downside scenario is a widening of Oracle credit spreads alongside reduced capex or OCI growth guidance; that combination would force an equity multiple reset even if reported revenue remains strong. Conversely, sustained positive FCF conversion and disclosed customer prepayments would invalidate the bearish balance-sheet thesis over 6-18 months.

Key Decisions for Investors

• Maintain ORCL underweight/short versus MSFT for the next 1-3 months: use a market-neutral long MSFT/short ORCL pair, sized modestly ahead of earnings. The trade monetizes divergence in financing capacity and execution risk; cover if ORCL provides credible evidence of positive FCF inflection, material customer-funded capex, or materially reduced lease exposure.
• Buy ORCL downside optionality only if implied volatility is below the post-results range: target 3-6 month put spreads rather than outright puts, with strikes around 10-20% below spot. The catalyst is a capex, FCF, financing-cost, or deployment-timing revision; risk is that backlog conversion and customer prepayments remove liquidity concerns faster than expected.
• Favor MSFT and AMZN over ORCL within AI infrastructure exposure through the next two reporting cycles. Their relative upside comes from potential workload substitution if Oracle delivery timelines slip, while their balance sheets can absorb continued capex without a comparable equity-financing or lease-coverage narrative.
• Set a credit-equity alert: materially wider ORCL bond spreads, rising interest expense, or any disclosure of delayed contracted capacity without corresponding rent deferral should trigger an increase in the ORCL hedge. Do not add solely on share-price weakness; require confirmation that cash conversion is deteriorating.