Stargate is a lot of things at once: a $500 billion investment, an avatar of modern technology, the culmination of OpenAI’s algorithmic-driven ambitions, and — perhaps most importantly — a proxy for the arms race in artificial intelligence that’s taking place globally (It’s also a cult-classic 90s television show, but that’s besides the point).
It’s also a stark reminder of the transformation that’s sweeping North Texas, as hundreds of data centers bloom across the region, and Texas as a whole. Increasingly, facilities like those toured by The News’ Lana Ferguson in recent months — who was also on hand at Stargate’s Abilene campus last week — are highlighting the mismatch that exists between energy supplies and voracious AI demand.
Trillions of dollars and a proportionate degree of anticipation is buttressing the AI economy. But it’s running headlong into a real problem. Namely, there may not be enough energy to power all of the projects expected to come online. Ferguson’s excellent cover story this week is a primer on how this technology is ratcheting up demand on power providers throughout the Dallas-Fort Worth area, leading to what one executive says is a “significant scale issue that we’re dealing with.”
It raises the question of whether utilities and policymakers are fully prepared for the onslaught of companies joining the likes of Oracle, NVIDIA and OpenAI in tapping into the AI wave, and how that appetite for power will push up electricity prices even more for Average Joes and Janes like us.
There are concerns about the data center boom’s effect on local power capacity. Yet at an appearance in Westlake on Wednesday, Gov. Greg Abbott said that Texas has enough power capacity to meet the needs of the technology players flooding into the state.
The state is the largest generator and consumer of energy in the U.S., producing the most oil and gas in the nation in addition to ample renewable energy.
“We have wind, solar, nuclear, which we’re adding even more … and, of course, natural gas. And so we have low-cost power, abundant power, that is attracting the AI data centers,” he said.
Even so, there’s also a growing realization that the juice needed to power the AI race is also playing a role in pushing up fossil fuel use, which has implications for the environment.
In a recent column, Bloomberg’s eminently brilliant Javier Blas zoomed in on International Energy Agency research to illustrate how the “myth” of peak oil and gas is steadily falling apart, mainly because fossil fuel demand in major economies is expected to stay robust for decades to come.
Among other findings, the IEA expects oil to remain “the largest single fuel” well into the second half of this century, for a variety of reasons. But it’s not hard to guess at least one. AI and so-called “hyperscale projects” are driving big natural gas needs in Appalachia, Hart Energy recently reported, while large scale campuses sprout all over Texas and the Southwest.
AI’s ascendance is great for growth, but its implacable energy needs are exposing the limitations of current energy policies. Renewable energy is great in isolated cases, but has severe challenges when it comes to scaling and powering large projects like data centers. Meanwhile, nuclear is the only carbon-free energy source — and it’s undergoing a much-overdue revival — but those projects are expensive, and take years to develop.
That leaves oil, gas and coal — key ingredients in generating electricity — to fill the gap. And the inconvenient truth is that fossil fuels are the only energy source reliable and abundant enough to power industrialized economies — and all the technology coming online.