Oracle’s proposed investment in 2 gigawatts of renewable energy projects for New Mexico comes as local opposition could delay development of the Project Jupiter data center that Oracle is building for OpenAI.
The two tech companies are developing the $165 billion Project Jupiter data center in Santa Teresa, New Mexico, as part of the broader Stargate AI infrastructure project announced by President Donald Trump in 2025. But Project Jupiter faces local protests and court battles over concerns about its environmental impacts—and the latest Oracle announcement on September 8 seeking proposals for renewable energy projects does not change the fact that the data center will be powered by fuel cells that consume natural gas.
“Like all matching programs, this would be synthetic in the sense that 2 GW of renewables wouldn’t directly power the data center,” wrote Michael Thomas, CEO of the Cleanview data platform that tracks renewable energy and data center projects, in a LinkedIn post.
Oracle and OpenAI had originally planned to build 2.2 gigawatts of natural gas generation that could have resulted in 14 million tons of CO2 emissions per year, Thomas pointed out. That would have effectively “wiped out 20 years of carbon emissions reductions in the state of New Mexico,” he wrote in his post.
State residents and environmental groups have rallied in opposition to Project Jupiter with protests and lawsuits. New Mexico’s Land Office twice blocked requests to run a natural gas pipeline to the data center in New Mexico’s Doña Ana County.
Such opposition likely prompted the tech companies’ Project Jupiter plan to pivot from gas turbines to fuel-cell technology, which produces electricity through an electrochemical process without combustion.
Edited to add the actual point I was getting at, because I am dumb and didn't type it: I am thrilled to see local government and population cottoning on to this kind of farcical greenwashing.
The power requirements are indeed massive for both, but quickly get out of hand when talking about AI DCs. The average hyperscale DC is around 40MW, whereas the AI DCs are typically starting at twice that, around 100MW. The cooling requirements alone at that level are staggering!
The actual utility of DCs can certainly be argued, but with much of the modern world being underpinned by the cloud infrastructure built up to date, I think it can be considered a necessary evil. The ability to centralize so much compute can make for a more efficient use of resources compared to smaller DCs spread around, each with their own power and cooling requirements.
AI DCs are objectively much more impactful from an environmental perspective, but I would hesitate to say it matches that impact in its utility. I can’t say with a straight face that we really need such high power requirements just to make Will Smith eat spaghetti.
I think a lot of the power constraints for DCs could be solved with a solid renewable source, combined with battery storage for dips and peaks in the grid. Once again, though, the AI DCs consume so much for such, relatively speaking, middling utility that I struggle to see why we really need so many AI DCs.