Oracle Reports Thursday. Its Capital Budget Is More Than Twice the Cash Its Business Produces.

Oracle plans roughly $70 billion of net cash capital expenditures in fiscal 2027, potentially translating to $90 billion-$95 billion of reported capex after customer prepayments and timing effects, versus $32 billion of fiscal 2026 operating cash flow. After spending $55.7 billion on capex in fiscal 2026, Oracle generated negative $23.7 billion of free cash flow and expects to raise about $40 billion through debt and equity, including a $20 billion at-the-market stock program. While cloud infrastructure revenue rose 77% to $18.1 billion in fiscal 2026, the AI data-center expansion depends on rapid cash-flow growth and creates material leverage, dilution, and execution risk.

Analysis

Oracle’s valuation is increasingly a financing-duration trade rather than a pure AI-demand trade. The relevant question is whether contracted cloud revenue converts to cash before interest expense, lease obligations, and potential share issuance reset per-share economics; if not, a premium AI infrastructure multiple can compress even while backlog and revenue remain strong. The announced equity capacity also creates an overhang: selling stock into strength caps upside and signals that internally generated cash is insufficient at the current build rate.

Near term, the earnings release can support ORCL if capex is materially back-end loaded, operating cash conversion accelerates, or customer prepayments demonstrate that customers—not Oracle’s balance sheet—are funding incremental capacity. Conversely, front-loaded capex alongside weak cash collections would raise the probability of further debt issuance and credit-spread widening over the next 1-3 months. Watch whether net debt/EBITDA and interest coverage deteriorate faster than cloud gross profit; that is the threshold at which equity dilution shifts from optional to likely.

Second-order beneficiaries are hyperscale infrastructure suppliers with paid capacity commitments and limited customer-credit exposure, notably NVDA and networking/power vendors, but Oracle’s financing constraint could ultimately reduce the durability of its order cadence. The contrarian view is that prepayments and signed AI contracts may make the spend less speculative than the cash-flow statement suggests; however, that only matters if prepayments are non-refundable, multi-year, and adequately cover power, equipment, and construction cost escalation. A successful funding plan may remove the near-term liquidity concern but still leaves lower long-run returns on invested capital as the central risk.

Key Decisions for Investors

• Maintain a tactical underweight/short ORCL into and through earnings only if the post-report rally takes valuation higher while reported operating cash flow fails to inflect; use a 1-3 month horizon and cover if management demonstrates capex is back-end loaded with materially improved cash collections or raises cloud gross-margin guidance.
• Prefer a pair trade: long NVDA / short ORCL in equal dollar beta-adjusted size for 1-3 months. NVDA retains supplier economics and customer diversification, while ORCL bears construction, financing, and utilization risk; exit if Oracle discloses binding customer-funded commitments that substantially reduce net capex exposure.
• Monitor ORCL bond spreads, incremental debt pricing, and any ATM utilization as real-time funding signals. A meaningful spread widening or early equity issuance would be a trigger to add to an ORCL short; absence of either is not itself bullish unless paired with cash-flow improvement.
• Do not chase a broad AI-infrastructure short from this development alone. Use any Oracle-driven weakness to assess selectively long suppliers only after verifying their Oracle revenue concentration and receivable exposure; the missing data is the extent to which supplier demand is backed by customer prepayments versus Oracle-funded buildout.