Oracle Has $638 Billion of Contracted Backlog and a $430 Billion Stock

Oracle ended fiscal 2026 with $638B of remaining performance obligations, up 363% YoY, but only expects 12% of that backlog (~$77B) to convert to revenue in the next 12 months. Despite record $32B operating cash flow, free cash flow was -$23.7B as AI data center buildout consumed capital, with the company raising $43B of debt and $5B of equity and expecting ~$40B more in fiscal 2027. While cloud infrastructure revenue accelerated (55% to 93% YoY across the year), the piece flags weaker AI contract economics and balance-sheet strain as the key reason the market discount persists.

Analysis

The market is not arguing that demand is fake; it is arguing that the conversion of that demand into equity value is being subsidized by the balance sheet. For ORCL, the near-term earnings story can look fine while the economic story deteriorates: heavy depreciation, power costs, interest expense, and equity issuance can keep reported growth high even as incremental free cash flow remains weak. That usually caps the multiple until investors see proof that each new dollar of AI infrastructure revenue carries acceptable contribution margin after financing.

Second-order winners are the vendors upstream of the buildout, especially NVDA and the broader AI infrastructure basket, because the spend is being pulled forward now even if Oracle is the party absorbing the financing pain. The losers are capital-light software peers that trade on recurring cash flow and can look expensive relative to a newly capex-intensive ORCL. On the competitive side, hyperscalers with stronger balance sheets can use this phase to defend share on enterprise AI workloads while Oracle is still funding capacity.

The key catalyst path is 1-3 months: guidance, debt issuance, and any change in the cadence of customer prepayments. The structural test is 6-18 months: whether ORCL can show positive FCF inflection without continued dilution. Consensus is likely over-fixated on backlog size and underpricing the financing drag; the contrarian risk is that customer funding meaningfully offsets capex and the stock re-rates higher once the market believes the earnings power is real. Falsifiers: a sustained FCF turnaround, slower net debt growth, or a materially smaller-than-expected equity raise.

Key Decisions for Investors

• Underweight ORCL on rallies over the next 1-2 weeks; use any strength into financing headlines to build a 3-6 month short or put-spread expression. Risk/reward improves if management reiterates large external funding needs and FCF stays negative.
• Pair trade: long NVDA / short ORCL for a 1-3 month relative-value view. NVDA monetizes the capex cycle immediately; ORCL still carries dilution and interest-rate sensitivity. Exit if ORCL proves it can fund growth with materially less external capital.
• Watch-list rather than buy: ORCL only becomes interesting long if the next two quarters show positive free cash flow and slower-than-expected debt/equity issuance. Until then, the thesis is balance-sheet repair, not compounding.
• For sector exposure, prefer capital-light AI beneficiaries over ORCL, including software/platform names that do not need to pre-fund data centers. The market may eventually reward proof of margin durability more than backlog scale.