Oracle has issued a force majeure notice to the developer of its massive artificial intelligence data center project in New Mexico, known as Project Jupiter, to prepare for suspending certain payment obligations if the project cannot begin operations as planned in 2028.
This move signals that the super-large AI infrastructure project, with a total investment of approximately $165 billion and a planned power capacity of up to 2.45 gigawatts, is facing increasingly evident delivery risks.
The problem is not with the data center itself, but with the power.
Project Jupiter originally aimed to address the data center's power supply challenges in New Mexico's arid environment through large-scale fuel cells, but the natural gas required to operate those fuel cells depends on a gas pipeline that has not yet been approved for construction on schedule. As pipeline construction continues to be delayed, the entire data center's power supply plan has been dragged into uncertainty.
Oracle Prepares for Potential Failure to Launch in 2028
The force majeure notice Oracle sent to the project developer is essentially a risk isolation measure. In large infrastructure contracts, if a project is delayed by external factors beyond reasonable control, force majeure clauses can typically temporarily suspend certain contractual obligations.
For Oracle, the key question is whether Project Jupiter can begin operations as originally planned in 2028. If the data center cannot be delivered on time due to external approvals or infrastructure issues, Oracle wants to avoid continuing to bear massive payment obligations before the project is ready for use.
The company still states that "Project Jupiter remains on track according to the established timeline." But the force majeure notice itself shows that Oracle has already begun preparing for delay risks from both a legal and financial standpoint.
Following the news, shares of several companies directly linked to the project came under pressure. Oracle fell about 4% at one point, project developer parent company Blue Owl Capital dropped about 4%, and Bloom Energy, which is responsible for core power supply equipment, fell as much as 6%.
The 2.45GW Data Center Is a Key Part of Oracle's AI Expansion
Project Jupiter is located in southern New Mexico and is an important component of the United States' large-scale AI infrastructure buildout. The entire campus has a designed power demand of 2.45 gigawatts, equivalent to the electricity consumption of approximately 1.8 million households, reflecting the extreme power needs of current large-scale artificial intelligence training and inference infrastructure.
The total project investment is estimated at approximately $165 billion. For Oracle, this data center is also one of the key infrastructure assets for expanding its AI computing supply capacity. In recent years, Oracle has significantly increased its data center investment, hoping to secure enough computing power to serve large AI customers.
But a data center project of this magnitude is no longer just a matter of servers and GPUs. Power supply, natural gas infrastructure, land permits, and local environmental approvals are gradually becoming the key conditions determining whether the project can materialize.
Fuel Cells Solve the Water Shortage Problem but Create a Natural Gas Problem
New Mexico has an arid climate, and traditional cooling and power generation systems for large data centers could consume large amounts of water. Therefore, Project Jupiter's energy plan uses Bloom Energy's solid oxide fuel cells. Oracle plans to use fuel cells to provide power for the entire 2.45-gigawatt campus.
These devices generate electricity through electrochemical reactions and have low water requirements, making them well suited for large data centers in water-scarce regions. But this plan has another prerequisite: fuel cells need a continuous supply of natural gas. And this is precisely where Project Jupiter has encountered its biggest bottleneck.
To supply natural gas to the campus, the project needs to build an approximately 17-mile gas pipeline called the Green Chile Project. This pipeline is being built by Energy Transfer and was originally scheduled for completion in August 2026, but the expected timeline has now been pushed back to February 2027.
The 17-Mile Natural Gas Pipeline Becomes the Project's Critical Bottleneck
The core reason for the pipeline delay is that the project has repeatedly failed to obtain the necessary right-of-way permits for crossing New Mexico state trust land. Local land management authorities have rejected the relevant applications multiple times.
Opponents are mainly concerned about the climate impact of the large data center and its natural gas infrastructure, as well as the project's long-term pressure on local energy and natural resources. This has created an obvious chain reaction: the gas pipeline cannot be built on schedule, meaning natural gas cannot be reliably delivered to the data center; without a natural gas supply, Bloom Energy's fuel cells cannot operate continuously; and without functioning fuel cells, the entire 2.45-gigawatt campus lacks a stable power source sufficient to support its scale.
Therefore, a natural gas pipeline of only 17 miles has in effect become the critical link determining whether a $165 billion AI infrastructure project can begin operations on time.
Oracle Now Faces More Than Just Project Delay Risk
The force majeure notice can first help Oracle reduce its own contractual risk, but it cannot solve the problems the project itself faces. Even if Oracle can temporarily ease the pressure of payments to the developer, project delays could still affect the banks and other creditors financing the entire data center construction.
Large data center infrastructure typically relies on long-term financing, and whether a project can begin operations on schedule and generate rental or service revenue is an important basis for financial institutions assessing debt repayment capacity. If the natural gas pipeline continues to be delayed or ultimately fails to obtain the required permits, lenders will need to reassess construction progress, cash flow, and ultimate delivery risk.
The impact on Bloom Energy is particularly notable. The company was originally set to take on an extremely large power supply role in Project Jupiter, making the project an important part of its future business growth. Once data center construction is delayed, fuel cell deployment and revenue recognition timelines could also be pushed back.
Project Jupiter Exposes a New Bottleneck in AI Infrastructure
The problems currently facing Project Jupiter also reflect that U.S. AI infrastructure construction is entering a new phase. When large technology companies expanded computing power in the past, the most closely watched issues were usually chip supply and capital expenditure.
But as the power demand of a single data center campus reaches gigawatt scale, the importance of grid capacity, natural gas supply, transmission lines, land use, and local approvals is rising rapidly. Oracle can already secure the capital needed to build data centers and can also obtain servers and computing equipment, but these conditions are not enough to guarantee that the project can actually operate.
For Project Jupiter, the key factor determining progress has now become a 17-mile natural gas pipeline. Oracle's early invocation of the force majeure clause also shows that one of the biggest construction risks for large AI data centers is gradually shifting from "whether there is money and whether there are chips" to "whether there is enough energy, and whether permits can be obtained to build the energy facilities."
Comments