Oracle Invokes Force Majeure as Stargate AI Hits Power Grid Bottlenecks

Oracle has issued a force majeure notice to developers of Project Jupiter, a key New Mexico data center, citing power supply and regulatory delays that threaten its 2028 launch. The action highlights growing infrastructure bottlenecks constraining the Stargate initiative, a joint artificial intelligence computing network backed by OpenAI, SoftBank, Oracle, and MGX. Announced in January 2025, Stargate targets up to 500 billion dollars in US-based AI infrastructure over four years to support next-generation model training and inference. Project Jupiter, developed with Blue Owl Capital, was designed to operate on a 2.45-gigawatt microgrid. After initial plans for gas turbines faced environmental pushback, developers pivoted to Bloom Energy solid oxide fuel cells. The revised setup still requires new natural gas pipelines and state-level air quality permits. Regulatory proceedings have faced repeated legal challenges, with New Mexico courts temporarily blocking approvals earlier this year. While facility construction proceeds, the standalone energy infrastructure remains unpermitted, directly delaying operational readiness. The delays have triggered immediate financial repercussions. Syndicated debt tied to the project, originally valued at approximately 18 billion dollars, now trades at a discount between 0.89 and 0.91 cents on the dollar, reflecting heightened investor caution. Oracle’s stock dipped as the company seeks to align payment obligations with delayed milestones and mitigate contractual risk. The situation underscores the vulnerability of massive AI capital expenditures, where prolonged construction cycles threaten to stall revenue conversion from cloud and compute contracts. Project Jupiter’s challenges reflect a systemic industry constraint. As data center power demands reach gigawatt scales, development timelines are increasingly dictated by energy grid capacity, utility partnerships, and permitting processes rather than hardware procurement. Texas regulators have already paused new data center approvals to reassess grid strain, with over 470 gigawatts of capacity currently queued for interconnection. While Stargate’s Abilene site has begun initial operations, most planned facilities remain in early development. The divergence between projected AI capacity and physical deployment realities has emerged as a critical bottleneck for the sector. Capital commitments and hardware orders scale rapidly, but energy generation, transmission infrastructure, and environmental clearances operate on multi-year timelines. For Stargate and its partners, Project Jupiter demonstrates that scaling artificial intelligence infrastructure requires parallel acceleration of grid modernization and regulatory frameworks, or face continued delays that ripple through financing, leasing, and compute delivery.