Oracle Declares Force Majeure on $165 Billion Project Jupiter Data Center

Oracle issued a force majeure notice on Project Jupiter, its $165 billion, 2.45GW New Mexico AI data-center project, after a key energy permit denial and with associated financing debt trading below 90 cents on the dollar; shares fell about 4% premarket. The project is targeted for 2028 and Oracle said the disruption will not affect previously stated FY27 revenue or earnings guidance, but it intensifies concerns over the company’s capacity buildout and financing needs. Oracle reported Q1 FY2027 revenue growth of 29.6% to $19.35B, cloud revenue growth of 62% to $11.61B, and $664B of RPOs, but capex of $28.50B drove free cash flow to negative $5.40B while interest expense rose 55% to $1.4B.

Analysis

The market should treat this less as a single-project delay than as a test of Oracle's financing model: AI cloud bookings require front-loaded power, construction and equipment spend, while customer cash receipts lag capacity delivery. A permit-driven delay extends negative free-cash-flow duration and raises the probability of further equity issuance or higher-cost debt, creating downside to both ORCL's earnings multiple and its cost of capital over the next 1-3 quarters. The critical verification point is whether management can preserve FY27 capacity and revenue commitments by reallocating workloads to other campuses rather than merely asserting no impact.

ORCL's cloud competitors are conditional beneficiaries. MSFT, AMZN and GOOGL have existing powered capacity and could capture incremental AI workloads if enterprise customers require contracted compute on schedule; however, the relevant risk is likely concentrated among a small number of hyperscale/AI customers, so broad cloud-share assumptions are premature. The more immediate second-order pressure is on project-specific suppliers and financiers: BE's valuation leaves little room for an order deferral, while OWL's equity risk depends on its actual funded exposure, collateral terms and whether the debt discount reflects construction risk rather than a recoverable timing mismatch.

Consensus appears too focused on the nominal backlog and too willing to capitalize it at face value. Backlog only supports valuation if power availability, financing and customer credit all convert into installed capacity; a second delayed campus, weaker capex guidance, or a slower RPO-to-revenue conversion would reset the narrative from AI growth to balance-sheet stress. Conversely, a permit resolution plus evidence that quarterly capex has peaked could drive a sharp relief rally because positioning and sentiment appear materially impaired.

Key Decisions for Investors

• Maintain an underweight ORCL for the next 1-3 months; add only on a documented permit resolution and unchanged FY27 capacity/revenue guidance. Thesis is falsified by a clear capacity-reallocation plan, stable funding costs and a credible path to positive free cash flow without additional equity.
• Use a 3-6 month pair trade: long MSFT or AMZN versus short ORCL in equal beta-adjusted dollars. The trade monetizes potential AI workload substitution and ORCL-specific funding risk; exit if Oracle demonstrates that affected capacity is fully replaced or if hyperscaler capex guidance weakens broadly.
• Avoid chasing BE despite its potential equipment exposure; treat any disclosed order deferral, revised deployment schedule or customer concentration disclosure as a downside catalyst. A tradeable long setup requires confirmation that the project commitment is binding, funded and economically material to BE revenue guidance.
• Put OWL on credit-watch rather than initiate a directional equity position until vehicle-level exposure, seniority and collateral are disclosed. If the relevant financing remains below 90 while ORCL's project timeline slips, widening private-credit marks could become a 6-12 month earnings and fundraising headwind.