NextFin News - Oracle sent a force majeure notice to the developer of its flagship New Mexico AI data center, a legal shield that could let it delay payments if the project slips past its 2028 target - and Morgan Stanley analysts warn the move adds pressure to the wider market for data center debt. Oracle shares fell about 4% on Thursday as investors weighed the company's ability to deliver the capacity behind its record backlog, while roughly $18 billion of loans tied to the project were already trading near 89 to 91 cents on the dollar.
The notice, first reported on Thursday and confirmed by the developer, is not an exit from the project. Oracle remains the primary tenant. Its purpose is narrower and, in some ways, more telling: it is designed to protect Oracle from potential cost increases and to let it delay payments should Project Jupiter be derailed and fail to come online in 2028 as planned. The credit market had started pricing the risk before the notice: Oracle's credit rating sits one notch above junk, its credit default swaps have hit record highs, and a bond maturing in 2056 has pushed above an 8% yield. The stock is down more than 45% over the past year, trading near $139 with a market capitalization around $422 billion, far below its 52-week high above $322.
The Situation: One Legal Notice, a $95 Billion Buildout Behind It
Project Jupiter is a massive data center campus in Doña Ana County, New Mexico, being built to serve artificial intelligence workloads. A spokesperson for Blue Owl Capital, whose Stack Infrastructure unit is developing the site, confirmed the notice and said Blue Owl, Stack Infrastructure and Oracle "remain fully aligned on Project Jupiter," adding that it "does not change the financial commitments to this multiyear project." Oracle said the project remains on schedule.
But the context has changed. On September 10, Oracle reported a strong first quarter of fiscal 2027 - revenue of $19.35 billion and adjusted earnings of $1.92 a share, both ahead of estimates - and raised its full-year guidance. The same earnings picture revealed the scale of what the company has signed up for: capital expenditures of $28.5 billion in the quarter, up from $8.5 billion a year earlier, and a full-year fiscal 2027 capex forecast of $90 billion to $95 billion.
That spending is being financed with debt. In fiscal 2026, Oracle raised $43 billion in debt and $5 billion in equity. For fiscal 2027, the company expects to raise approximately $40 billion through a combination of debt and equity, including a previously announced $20 billion at-the-market equity offering, and it does not expect to issue additional debt in calendar 2026. Total borrowings now stand at roughly $105 billion, and long-term leasing commitments tied to data centers approach $250 billion, many with 15- to 20-year lives that Oracle expects to sublease on shorter time horizons.
The trigger for the notice appears to be power. A person familiar with the matter said Oracle issued the notice citing potential delays in securing power for the project, and that securing power for the site is Oracle's responsibility under the contract. The project is reportedly at least seven months behind schedule, with permitting, power infrastructure and local opposition complicating development; a request connected to supplying natural gas for planned 2.2-gigawatt gas turbines was rejected.
In a statement, Oracle said force majeure notices are commonplace in developments of this scale and are often used to preserve contractual rights among project partners. In large buildouts, such notices function as a reserving device - preserving the right to claim relief on timelines and penalties without conceding that a project is off track.
The Transmission Channel: From One Project to the Whole Debt Market
The first question is why a single legal notice on a single project rattled more than Oracle's stock. The answer is that Project Jupiter is not an isolated deal. It is a test case for the financing model that the entire AI data center boom now depends on.
The model works like this: a cloud provider signs a long-term contract with an AI customer, then borrows against the future revenue to build the facility, often through a third-party developer. The debt is meant to be distributed to a broad set of lenders and investors, who underwrite it on the assumption that capacity will come online on schedule and the customer will pay. Oracle's force majeure notice signals that the company itself now sees enough execution risk to warrant legal cover.
"The notice signals the company's own risk assessment on execution has shifted enough to warrant legal cover," Rishi Jaluria, an analyst at RBC Capital Markets, wrote in a note to investors. "Any slippage in the 2028 timeline has direct implications for capacity delivery and revenue recognition given Oracle's supply-constrained posture."
The mechanism is a mismatch in timing. Debt service does not wait for revenue recognition. If a project is delayed, interest and principal payments keep coming while the cash from the customer has not started. For a company already carrying more than $100 billion of debt and planning roughly $90 billion to $95 billion of capex in a single year, that mismatch compounds quickly.
Morgan Stanley's warning lands here. If the anchor tenant can invoke force majeure to delay payments, the debt tied to that project becomes harder to distribute, harder to price, and more likely to remain on the balance sheets of the banks that originated it. Reports have indicated that banks including Morgan Stanley, SMBC, MUFG and Bank of America have been working to distribute and de-risk large data center loans tied to Oracle projects, with more exposure left on their books than originally anticipated.
One Delay Is Cyclical. The Financing Model Is the Structural Question.
Two forces are being conflated in the market reaction, and separating them matters.
The cyclical leg is straightforward. Project Jupiter is reportedly at least seven months behind schedule because of permitting, power and local opposition. These are execution hurdles that can, in principle, be worked through. If the project merely slips by a year, Oracle's payment obligations shift with it, and the long-term contract with the customer - widely reported as a roughly $300 billion, 10-year agreement with OpenAI - remains intact. On that reading, the force majeure notice is prudent, even sensible, risk management.
Evercore called the reported notice "a pretty sensible move" to align Oracle's revenue and expenses if the project is delayed, and said it "is not an indication that this project is going away."
William Blair made the same point: the immediate financial impact should be limited because Project Jupiter is not expected to contribute revenue in Oracle's current fiscal year.
The structural leg is the harder problem, and it is where the Morgan Stanley concern sits. The structural claim is this: the AI data center buildout is being financed with a maturity and risk profile that does not match the underlying assets. Data center loans are often originated at 70% to 80% loan-to-value, according to Morgan Stanley's own research on bridging the data center financing gap. The assets are bespoke and illiquid, dependent on a single tenant and a single technology cycle. The revenue is promised over 10 to 20 years, but the debt is being distributed now, priced on the assumption that capacity arrives on time.
JPMorgan's asset management division made the same point in its January 2026 outlook, calling Oracle and Meta "two debt outliers" that are starting to finance data centers via the debt markets. It noted that OpenAI has committed to pay Oracle $60 billion a year - an amount OpenAI does not yet earn - for cloud facilities Oracle has not yet built, requiring 4.5 gigawatts of power. Since JPMorgan first highlighted Oracle's challenges in September 2025, the stock was down roughly 35% and credit spreads had widened by about 90 basis points as of mid-December 2025.
This is a structural shift, not a cyclical dip. A cyclical problem mean-reverts when the project catches up. A structural problem does not self-correct when one project is delayed - it requires the financing model itself to be repriced. The industry has moved from cash-flow-funded capex to debt-funded capex at a scale and speed that has no clean precedent.
The Second-Order Risk: A Repricing of the Whole Asset Class
The first-order effect is obvious and already priced: Oracle's stock falls, its CDS widens, its borrowing costs rise. The second-order effect is what Morgan Stanley is flagging, and it is cross-asset.
If data center debt tied to the most prominent AI infrastructure story starts trading at 89 to 91 cents on the dollar, the discount does not stay contained. It spreads to the rest of the asset class: the loans being syndicated for other hyperscaler-linked projects, the asset-backed securities issued against data center rents, and the private credit facilities funding the buildout. Morgan Stanley expects about $20 billion of AI-related deals in leveraged finance markets in 2026; JPMorgan projects $150 billion over the next five years. A repricing at the margin affects the pricing of the whole pool.
There is a precedent investors are watching. In November 2025, a data center delivery delay at CoreWeave sent that stock down 16%. The difference now is scale: Oracle's program is an order of magnitude larger, and its debt is public, rated, and held across the investment-grade complex. When a BBB-minus rated issuer with more than $100 billion of debt starts to look execution-constrained, the question is not just whether Oracle can deliver - it is whether the investment-grade label on AI infrastructure debt was earned or assumed.
The third-order effect is the expectation gap. Oracle's valuation and its backlog - remaining performance obligations that analysts have estimated in the hundreds of billions - are priced on the assumption that capacity delivery keeps pace with contract signing. The force majeure notice introduces the possibility that revenue recognition lags the backlog, which would force a downward revision not just of one quarter but of the multi-year growth trajectory the stock is underwriting.
The Counter-Thesis: Oracle Is Not CoreWeave, and the Contract Is the Moat
The strongest argument against the bear case is simple: Oracle is not a speculative cloud provider, and its customer is not going anywhere. The OpenAI contract is a long-term take-or-pay commitment from the most sought-after AI company in the world. Even if Project Jupiter is delayed, the demand for the capacity is not in doubt - the global shortage of AI compute means any megawatt Oracle brings online will be absorbed.
The bull case also notes that Oracle has repeatedly said it remains on schedule, and that force majeure notices are routine in large construction and infrastructure contracts. The company raised $43 billion of debt in fiscal 2026 and has a clear funding plan for fiscal 2027. Its cloud infrastructure revenue grew 121% in the most recent quarter to $7.4 billion. This is a company with real, growing cash flow - not a balance sheet waiting to break.
Morningstar's Luke Yang, CFA, maintains a $220 fair value estimate for Oracle and treats a Jupiter delay as only a possibility at this stage. "Our base case assumes that Oracle can keep up with its delivery pace and fulfill its long-term growth targets," he wrote, adding that the shares currently look undervalued.
This counter-thesis has real force. But it rests on one assumption that the force majeure notice itself undermines: that execution risk is low enough that the financing model does not need to change. If Oracle - with the deepest pockets and the strongest customer in the business - needs legal cover against its own project timeline, then the same model applied to smaller, less-cash-rich players is being underwritten on even thinner assumptions. The bear case does not require Oracle to fail. It only requires the market to stop assuming that every project comes online on time.
What to Watch: The Signals That Decide the Thesis
The base case is that Project Jupiter eventually comes online, perhaps with a delay, and Oracle's payment obligations shift with it. In that scenario, the force majeure notice is remembered as a prudent legal move, and the stock recovers some of its losses as capacity starts contributing revenue.
The downside case is that the delay is longer than expected, or that more projects in Oracle's pipeline show similar stress. If the $18 billion of Jupiter-linked loans continue trading below 90 cents, or if Oracle's CDS spreads widen further past their record highs, the cost of capital for the whole program rises - and with $90 billion to $95 billion of capex still to fund, even a modest increase in borrowing costs compounds into billions of dollars.
The upside case is that Oracle brings capacity online faster than the market now expects, revenue recognition accelerates, and the backlog converts to cash at the pace the stock originally priced in. That would validate the debt-funded model and could re-rate the entire data center credit complex.
The falsifying signal for the bear thesis is specific: if Oracle reports on-time capacity additions over the next two quarters and the Jupiter-linked loans recover above par, the structural-repricing argument is wrong - the issue was a single-project delay, not a broken model. The falsifying signal for the bull thesis is equally concrete: if Oracle's next quarterly report shows capex rising while revenue recognition from new capacity stalls, or if the company is forced to issue additional debt in calendar 2026 despite saying it would not, the market will have to reprice the financing model, not just the stock.
The force majeure notice is not Oracle admitting defeat on a data center. It is the market learning that the AI infrastructure boom has been financed on the assumption that nothing ever slips - and that assumption is now the asset class's most leveraged position.
Data as of September 25, 2026. Market figures reflect trading on and around that date.
Explore more exclusive insights at nextfin.ai.
