"Stargate" AI Project Stalled by Power Constraints: $500 Billion Investment Plan, Oracle Acts First

The new plan still calls for a capacity of 2.45 GW—equivalent to building a massive, standalone microgrid exclusively for Project Jupiter. At the time, Oracle stated that fuel cells could provide continuous baseload power to the AI data center while reducing water consumption and certain pollutant emissions. But swapping out equipment did not resolve the permitting challenges. Solid oxide fuel cells still rely primarily on natural gas, meaning high-pressure natural gas transmission pipelines must still be built upstream; moreover, a 2.45 GW-scale power generation facility itself requires air quality and environmental permits. This summer, administrative and legal disputes surrounding the microgrid escalated further: on August 23, the New Mexico Supreme Court temporarily halted the air quality permit review process. Although the court rejected the emergency injunction request on September 17—allowing the permitting process to resume—the final permit remains outstanding to this day. This has created Project Jupiter’s most immediate and tangible conflict. While construction of the data center building itself can continue under county-level permits, the microgrid responsible for supplying power must clear a separate, state-level regulatory process. The server rooms can keep rising floor-by-floor—but the energy system that will actually power the servers remains stuck in permitting limbo. Any delay at the power supply node inevitably pushes back the data center’s delivery timeline. Oracle’s issuance of a “force majeure” notice is thus a proactive risk-mitigation measure: should energy-related or permitting issues cause further delays, certain payment obligations may be extended accordingly. $1.8 Billion Loan Already Affected Beyond construction progress, financing pressures behind Project Jupiter are also surfacing. According to Reuters citing the Financial Times, approximately $1.8 billion in syndicated loans tied to the project have recently traded at a discount in the secondary market: debt instruments with a face value of $1 now sell for roughly $0.89–$0.91. This price decline signals investors demanding higher risk premiums—and growing concerns over project delays and future repayment capacity. Meanwhile, Oracle itself is in the midst of an AI infrastructure investment surge. Over the past year, the company has signed a series of massive data center and cloud computing contracts with OpenAI, requiring sustained capital expenditures over the coming years—including servers, data centers, power infrastructure, and network buildouts. Accordingly, whether these projects go live as scheduled directly determines when upfront investments begin converting into cloud service revenue. The stakeholders behind Project Jupiter extend far beyond developers and tenants. The entire project links infrastructure capital firms like Blue Owl, syndicated lenders, Oracle, and ultimately OpenAI—the end user of the compute capacity. Any extension of the construction timeline transmits impact down the chain—through financing, leasing, and cloud service contracts. And similar challenges are not confined to New Mexico. Just three days before Project Jupiter issued its force majeure notice, Texas announced it would suspend approvals for new state-level data center projects and reassess grid capacity.

To Deliver Compute Power, Oracle Is Transporting Natural Gas by Truck to Power Its Data Center

This is not Oracle’s first attempt. According to reports citing informed sources, at its data center project in the Salt Lake City suburbs, Oracle relied on truck-delivered natural gas for over a year to maintain steady progress—the service provided by Certarus, a subsidiary of Superior Plus. Certarus stated that prolonged delays in connecting data centers to pipelines and the grid “have created growing demand for energy solutions that are rapidly deployable and flexibly scalable.” Similarly, at the data center campus Oracle is building for OpenAI in Shackelford County, Texas, the company adopted the same approach—with VoltaGrid LLC serving as the supplier. Oracle posted on social media: “Thank you to the VoltaGrid team for identifying a cost-effective power solution that ensures customers can access OCI (Oracle Cloud Infrastructure) compute capacity on schedule.” The pressure is most acute in New Mexico. Wall Street Insights noted that the data center project codenamed “Project Jupiter” faces severe pipeline delays. Pipeline operator Energy Transfer LP was forced to redesign its route after New Mexico regulators rejected the originally planned pipeline alignment—pushing the projected commissioning date from this summer to next year. To safeguard the project timeline, Oracle is now considering delivering compressed natural gas to the site via trucks—enabling early-stage operations before the pipeline is officially connected to the grid. Signs of urgency are already evident: last month, Oracle issued a force majeure notice to the project’s developer, potentially shielding the company from certain payment liabilities if delays persist. High Costs and Scalability Limits: Not a Long-Term Fix Truck-based delivery is no cheap solution. According to Jack Weixel, Senior Energy Analyst at energy analytics firm East Daley Analytics, the all-in cost of truck-delivered compressed natural gas—including labor, specialized equipment, and vehicle fuel—is roughly four times the price of natural gas at major pipeline hubs.

Oracle Moves Gas by Trucks to Avoid Data Center Power Delays

3 days ago

Oracle utilizes trucked compressed natural gas to bypass pipeline delays for data centers, leveraging virtual pipeline services despite quadruple the cost of traditional supply.

bloomberg

2026-10-08

Oracle Moves Gas by Trucks to Avoid Data Center Power Delays

Oracle Moves Gas by Trucks to Avoid Data Center Power Delays

FAQ