Project Jupiter now has a financing lesson attached to its power constraint. Reuters reported on September 24, with an update on September 25, that Oracle issued a force-majeure notice tied to the New Mexico data-center campus, citing delays in securing power. A person familiar with the matter told Reuters the program faces a one-year delay.
The reported delay does not mean the project has been abandoned. Blue Owl said the notice does not change the parties’ commitments. But it does turn an infrastructure problem into a test of contract design, cash timing and risk allocation—the same issues already visible in the project’s loan-distribution pressure.
The judgment
- Implication: In AI infrastructure, power availability is not merely an operating dependency. It can determine when construction-period capital begins earning its expected return and whether lenders will finance the next project on similar terms.
- What would change the conclusion: A revised power and construction schedule that preserves the original completion economics—or contractual remedies that clearly compensate capital providers for the delay—would reduce the evidence that execution risk is migrating into financing risk.
- Management action: Separate demand, construction, power, financing and cash-return milestones in the investment model. Assign an owner to each dependency and show explicitly who bears the cost when a milestone moves.
The reported facts below come from Reuters and the cited company materials. The conclusions about finance, controls and management response are Numbers & Judgment analysis.
What changed at Project Jupiter
Reuters reported that Oracle invoked force majeure because of delays securing power for the site. A person familiar with the notice and the deal said the delay is expected to last a year. The same source said Blue Owl has about $3 billion of equity invested and earns a lower return during construction, with returns increasing after completion.
That last point is the financial hinge. A construction delay can defer the moment when an asset begins producing its higher contracted return, even if the customer remains committed. The result is not just a later opening date. It can be a longer period of capital tied up at a lower return, a different interest-carry profile and a changed refinancing timetable.
Reuters also reported that the event is affecting discussions around other proposed data-center financings. The six largest U.S. technology companies are projected by Moody’s to spend about $1 trillion on AI-related capital in 2027, according to the report. At that scale, one project’s contract provisions can become a market precedent.
AI infrastructure has three markets that must clear
The useful finance framework is to separate three markets that are often collapsed into one AI-growth narrative.
- The customer market: Is there contracted demand for compute?
- The physical-infrastructure market: Can land, power, permits, equipment and construction be assembled on schedule?
- The capital market: Will lenders and investors finance those assets at a price that preserves acceptable returns?
A signed compute contract can help clear the first market. It does not automatically clear the other two. That distinction also matters when evaluating large backlogs: contract value is not the same as recognized revenue until infrastructure, financing and service delivery arrive in the right sequence.
Force majeure is also a capital-allocation clause
Force-majeure provisions are usually discussed as legal protections. In project finance, they also allocate economic exposure. The important questions are concrete:
- When does the customer’s payment obligation begin?
- Who funds interest, taxes and construction overhead during the extension?
- Does a delay trigger additional equity, lender consent or a revised completion guarantee?
- Can committed financing still be drawn on the original schedule and terms?
- Which party absorbs a lower construction-period return?
Those terms can determine whether a delay is manageable or whether it changes the project’s risk-adjusted value. Reuters quoted a lawyer saying data-center contracts are evolving rapidly. That is unsurprising: contracts written when capital was cheaper and power timelines seemed more certain may not allocate today’s risks cleanly.
The earlier loan signal now matters more
This article originally examined reports that roughly $18 billion of Project Jupiter loans were being quoted by syndicate banks at about 89–91 cents on the dollar. Below-par pricing did not prove the project would fail. It did indicate that investors required a discount for credit, execution and liquidity risk.
The force-majeure notice connects that market signal to a specific operating dependency. A lender evaluating the next AI campus can now ask not only whether demand is contracted, but whether power delivery, completion triggers and delay remedies are bankable. This is why the broader shift of AI financing toward insurers and other long-duration investors deserves attention: long-duration capital does not eliminate construction and concentration risk.
A better CFO scorecard
Management should track the project as a linked set of milestones rather than one completion percentage:
- contracted customer demand and payment start dates;
- committed versus still-needed debt and equity;
- loan distribution, secondary pricing and lender concentration;
- power availability, interconnection and backup-generation certainty;
- construction, permitting and equipment milestones; and
- cash yield and weighted cost of capital under revised schedules.
The scenario model should distinguish an on-time case, a delay with preserved economics, and a delay that requires more capital or produces a lower return. The goal is not to invent a precise probability from one news report. It is to show which contractual or operating fact would move value. The free Forecast Scenario Planner provides a simple structure for separating those cases.
Delay is not failure, but it is information
Blue Owl said Oracle, STACK Infrastructure and Blue Owl remain aligned and that the notice does not change their financial commitments. That statement matters and should keep the analysis proportionate. A one-year delay can be absorbed by a well-capitalized project with durable customer demand and effective contractual remedies.
But the new information narrows the question. Project Jupiter is no longer only a case about distributing large loans. It is a live test of whether the financing structure can absorb a delay in the physical infrastructure on which the revenue and return schedule depend.
For AI infrastructure, financeability now depends on more than who wants the compute. It depends on whether power, contracts and capital arrive in the same sequence.
Sources
- Reuters, “Oracle, Blue Owl project delay sends ripples through AI financing, sources say,” September 24, 2026; updated September 25, 2026.
- Reuters, “Oracle’s $18 billion data center debt under pressure, FT reports,” September 18, 2026.
- Oracle, “Oracle Issues RFP for 2 GW of Renewable Energy in New Mexico,” September 8, 2026.
- Oracle, “Project Jupiter Statement on Construction Permitting,” September 14, 2026.
By Robert Young | Numbers & Judgment. Published September 20, 2026. Materially updated September 26, 2026 to include the reported force-majeure notice and financing implications.

Leave a Reply