Local backlash and reliability concerns are challenging the vision of an AI boom built quickly using large off-grid data centers.
Why it matters: If giant projects stumble, trillions of dollars in AI investment and the industry's plans to rapidly expand data centers could be at risk.
The big picture: In an effort to speed AI development, some companies are building data centers powered mainly by onsite generation rather than waiting years to connect to the electric grid.
There are 59 data centers with a combined capacity of about 90 gigawatts that plan to build their own power "behind-the-meter" using sources like gas turbines, generators and fuel cells, according to a report from research firm Cleanview.
A smaller subset is trying to run large campuses mostly on behind-the-meter power as a way to move more quickly than the local utility and the grid can.
Infrastructure risk analysis firm Occam Edge tracks 12 projects where onsite power is the primary serving supply, representing about 10.6 GW of announced capacity.
State of play: Recent hiccups are highlighting the challenges that come with bypassing the grid.
Earlier this month, New Mexico's top land official rejected a gas pipeline meant to supply onsite fuel cells for Oracle's 2.5 GW "Project Jupiter" data center campus, part of Oracle and OpenAI's Stargate initiative.
Local residents complained of burning lungs and 60-decibel level noise, and a supervisor called for new laws to regulate generators.
Catch up quick: Chief developers of this off-grid approach include OpenAI, and its partners Oracle and Crusoe, which have worked on OpenAI's Stargate campuses.
A Stargate project in Abilene, developed by Crusoe, reportedly went offline for days at a time due to issues with power and cooling equipment.
Another off-grid player is Elon Musk, who rattled the energy world when he built xAI's Colossus 1 using mobile gas turbines in just a few months. While it was initially off-grid, it's now grid-connected and selling compute to Anthropic.
Despite a lawsuit from the NAACP over Colossus 1, Musk is now using gas turbines to power Colossus 2 (with compute sold to Google), and earlier this month bought a mobile gas turbine company.
What they're saying: Critics argue off-grid data centers could prove slower and more costly to deploy, less reliable to run, and more expensive than expanding the electric grid.
"This is a flimsy way to deploy AI," says energy investor Jigar Shah, who predicts much of the bullish off-grid deployment figures won't materialize.
Power engineers are worried that off-grid data centers won't meet reliability targets, says Occam Edge founder and CEO Christian Okoye, who compares "Dark Gigawatts" to the "Dark Fiber" of the dotcom bust.
"It's a time of desperation," says Josh Wong, founder and CEO of ThinkLabs AI, which helps utilities find more capacity on their current grids. "The [off-grid] business model is bolt-on."
Follow the money: Investors are lacking insight into the reliability of off-grid data centers, says Okoye.
Investors faced with the uncertainties of off-grid data centers could decide not to underwrite them.
S&P Global Ratings recently downgraded Oracle's long-term issuer credit rating to BBB- from BBB, just one step above junk status, due to its massive data center spending, which includes investment in onsite power infrastructure.
Trillions of dollars are riding on whether large off-grid data centers can work as advertised, justify their added cost and win local support.
What's next: More high-profile problems building and running off-grid data centers could lead to a reckoning that the AI boom will be a lot more grid-powered than many AI companies hoped.
Facts Only
* 59 data centers plan to build onsite power using sources like gas turbines, generators, and fuel cells.
* Infrastructure risk firm Occam Edge tracks 12 projects where onsite power is the primary supply, representing about 10.6 GW of announced capacity.
* New Mexico's top land official rejected a gas pipeline for Oracle's Project Jupiter data center campus.
* Local residents complained about burning lungs and noise regarding the rejected pipeline.
* A Stargate project in Abilene reportedly went offline for days due to power and cooling equipment issues.
* Elon Musk used mobile gas turbines to build xAI's Colossus 1.
* Subsequently, Musk is using gas turbines to power Colossus 2 and acquired a mobile gas turbine company.
* Critics argue off-grid data centers could be slower, more costly, less reliable, and more expensive than grid expansion.
* Energy investor Jigar Shah predicts many bullish off-grid deployment figures will not materialize.
* Occam Edge founder Christian Okoye stated that off-grid data centers may not meet reliability targets.
* S&P Global Ratings downgraded Oracle's long-term issuer credit rating to BBB- due to massive data center spending including onsite power infrastructure.
Executive Summary
Efforts to accelerate AI development have led some companies to build data centers utilizing onsite power generation, bypassing reliance on the main electric grid infrastructure. A subset of these projects aim for this self-sufficiency to move faster than utility expansion allows. Fifty-nine data centers collectively plan to use onsite generation, such as gas turbines and generators, for power. Infrastructure risk analysts track twelve projects where onsite power serves as the primary supply, accounting for approximately 10.6 GW of announced capacity.
Recent incidents have highlighted the challenges associated with this off-grid deployment. For example, a gas pipeline rejected by New Mexico officials was intended to supply fuel cells for Oracle's Project Jupiter data center campus, resulting in local complaints about noise and health impacts. Operational issues also arose at other sites, such as a Stargate project in Abilene experiencing power and cooling equipment failures.
Critics argue that this off-grid approach may be slower, more costly, and less reliable compared to grid expansion. Experts express concern that the reliability targets for these systems may not be met, leading some to question the viability of the deployment figures. Furthermore, investors are exhibiting caution, evidenced by a downgrade in Oracle's credit rating due to its data center spending, reflecting uncertainty surrounding the reliability and financial footing of off-grid infrastructure.
Full Take
The narrative surrounding the rapid deployment of off-grid AI data centers presents a tension between the pursuit of speed and the established realities of physical infrastructure and systemic reliability. The core conflict lies in prioritizing an expedited developmental timeline over proven energy delivery systems, creating a scenario where emergent technological ambition intersects with tangible physical constraints and social acceptability.
The pattern observed is a reliance on novel, high-risk solutions—mobile generation and localized power loops—as a shortcut to meet aggressive growth targets for AI infrastructure. This mirrors historical patterns where speed of deployment leads to deferred systemic risk; the focus shifts from managing known grid uncertainties to creating unknown operational vulnerabilities in decentralized systems. The shift in perspective demanded by this situation is understanding that "speed" achieved through bypassing established systems introduces latency into reliability and financial stability, as evidenced by investor hesitancy and technical warnings regarding "Dark Gigawatts."
The implications suggest a potential decoupling of AI progress from robust energy planning. If the market rewards these localized solutions without fully accounting for long-term operational costs or infrastructure fragility, it risks creating a precarious foundation for the entire AI boom. The risk is not just that projects will fail, but that the necessary societal and regulatory mechanisms to ensure reliable energy access are sidelined in the name of innovation velocity.
Bridge questions: What metrics should be prioritized—deployment speed versus sustained grid reliability? How can investment frameworks be adapted to price in the systemic risk introduced by decentralized energy solutions? If off-grid deployment becomes the norm, what new governance structures are required to ensure public safety and equitable resource distribution?
Sentinel — Human
This analysis appears to be human-authored investigative reporting that synthesizes specific, albeit anecdotal, events and expert opinions regarding the infrastructure risks associated with off-grid AI data centers.