Oracle's 2.45 GW Cloud Data Center Faces 2028 Compute Capacity Risk

For SaaS and cloud engineering teams, this is an infrastructure supply warning: Oracle's 2.45-gigawatt New Mexico data center is a key future GPU compute node for OpenAI and other AI workloads. A force majeure notice means capacity that was expected online by 2028 might slip, squeezing availability and pricing for AI cloud services. It exposes how energy pipeline governance, not just chip supply, now dictates cloud capacity timelines.

Oracle's decision to send a force majeure notice to the developer of Project Jupiter on September 24, 2026 is the latest signal that the AI infrastructure buildout is running into the messy realities of energy permitting and construction logistics. Project Jupiter is a Stargate data center campus in New Mexico designed for 2.45 gigawatts of capacity and intended to run on gas-powered fuel cells from Bloom Energy. Force majeure clauses are typically used to shield parties from unforeseeable or uncontrollable events, and Oracle's invocation, first reported by Bloomberg, is not an exit from the project but a mechanism that would allow the company to delay payments if the facility misses its 2028 target. Oracle insists the schedule remains intact. In a statement to CNBC, the company said, 'Project Jupiter remains on our planned schedule. We are fully committed to New Mexico and confident in our path forward.' Oracle has not named the developer in its public statement, but Bloomberg's reporting and the request for comment directed to Blue Owl suggest the private capital firm is the developer or owner behind Project Jupiter.

“

Project Jupiter is a Stargate data center campus in New Mexico designed for 2.45 gigawatts of capacity and intended to run on gas-powered fuel cells from Bloom Energy.

The timing is notable because the campus relies on a natural gas pipeline from Energy Transfer, which has slipped nearly six months to February 1, 2027 after regulators repeatedly denied permits for the line. In August 2026, Bloomberg reported that the pipeline route had been changed following those rejections, and a separate air-quality permit for the fuel cell system remains an obstacle. These are not theoretical risks; they are concrete scheduling problems that could push a 2.45-gigawatt campus beyond its 2028 operational date. The force majeure notice, therefore, reads less like a sudden act of God and more like a legal precaution tied to known energy infrastructure delays.

The Stargate program—announced in 2025 as a multi-company push to build out AI data center capacity for OpenAI—has become a barometer for how quickly large language model demand is translating into physical infrastructure. Oracle's role as a lead cloud provider and anchor tenant makes its scheduling risk a bellwether for the entire sector. The New Mexico campus is one of several gigawatt-scale builds that depend on natural gas, fuel cells, and new transmission or pipeline capacity. In many regions, utilities and pipeline operators are struggling to secure permits quickly enough to meet AI-driven demand. The Energy Transfer pipeline's repeated regulatory rejections highlight how local environmental and safety reviews can slow projects even when national policy encourages AI infrastructure buildout.

For Oracle, the notice is a financial and legal hedge. If Project Jupiter slips, the company could withhold or delay lease payments, protecting its balance sheet from paying for space it cannot yet use. For the developer, likely Blue Owl, a delayed revenue stream could complicate debt service and investor returns on a project that already carries steep construction costs. Bloom Energy's fuel cell order is also tied to the gas supply, so pipeline delays ripple through the supply chain. Energy Transfer faces credibility questions over its ability to deliver on time. The air-quality permit issue adds another layer of uncertainty.