In a bold fusion of cutting-edge technology and America’s fossil fuel heartland, a massive new AI data center is set to rise in West Texas, right in the Permian Basin—the undisputed epicenter of the U.S. fracking revolution.
Dubbed “Horizon,” this ambitious project is spearheaded by Nvidia-backed AI startup Poolside in partnership with cloud infrastructure giant CoreWeave. Spanning over 500 acres on the historic Longfellow Ranch owned by the oil-rich Mitchell family, the facility promises to harness the region’s abundant natural gas resources to power the next wave of artificial intelligence computing. But as AI’s insatiable hunger for energy collides with supply chain bottlenecks, questions loom about the sustainability of this boom.
The Horizon Project: AI Meets Oil Country
At the core of Horizon is Poolside’s vision to build an “AI megaplex” capable of delivering up to two gigawatts of computing power—equivalent to the output of the Hoover Dam. The initial phase, anchored by CoreWeave as the primary tenant, will bring 250 megawatts online by the end of 2026, with reservations for an additional 500 megawatts in expansions. Construction is slated for completion in early 2027, employing modular off-site building techniques to slash costs below the industry’s typical $16 billion price tag for a similar-scale facility (excluding chips).
What makes Horizon stand out is its strategic location. Nestled amid the Permian Basin’s sprawling oil and gas operations, the site benefits from proximity to natural gas production, processing plants, and pipelines. An existing on-site gas plant, originally developed by Occidental Petroleum, will generate power directly for the data center, allowing incremental scaling without overwhelming the local grid. This self-reliant approach addresses a key pain point in the AI industry: the scarcity of ready-to-deploy computing infrastructure. As Poolside co-founder and CTO Eiso Kant noted, the real bottleneck isn’t just gigawatts—it’s the speed of delivery in a multitrillion-dollar AI arms race.
Poolside, focused on developing AI with “humanlike intelligence” for coding, has secured Nvidia’s backing and is amid a $2 billion fundraising round valuing the company at $14 billion. CoreWeave, meanwhile, will provide Nvidia AI resources starting this December, positioning Horizon as a hub for high-stakes AI development.
Do you have a Tax Burden in 2025?
Natural Gas: The Fuel Powering AI’s Future
West Texas’s fracking boom has flooded the market with cheap natural gas, making it an ideal energy source for power-hungry data centers. Unlike many U.S. facilities that rely on utility grids, Horizon’s on-site generation model leverages the Permian’s infrastructure to cut costs and ensure reliability. This isn’t just a one-off; it’s part of a broader trend where AI’s explosive growth is driving unprecedented demand for natural gas.
Globally, AI computing is projected to require an additional 55 gigawatts of power capacity by 2030, much of it potentially met by natural gas-fired plants.
In the U.S., utilities like PPL are greenlighting new gas power plants—such as 1.3 gigawatts in Pennsylvania—to handle the surge from data centers.
Natural gas’s role in displacing coal while providing baseload power makes it a pragmatic choice, even as environmental concerns mount.Yet, this reliance highlights AI’s hidden energy footprint. Data centers could amplify natural gas consumption, with older plants being patched up to meet demand rather than retired.
In regions like Texas, where OpenAI and Oracle are building massive facilities, the grid is already feeling the strain, prompting new rules to curtail power to large users during emergencies.
The Turbine Shortage: A Critical Bottleneck
While natural gas abundance is a boon, a looming shortage of gas turbines—the workhorses of natural gas power generation—threatens to derail the AI energy rush. Supply chain issues are making it harder to bring new capacity online quickly, potentially crippling grid reliability.
This vulnerability is exacerbated by AI’s rapid pace, forcing data centers to fall back on older, less efficient simple-cycle turbines that guzzle more fuel and emit higher emissions.
Experts warn that without an emergency push to ramp up turbine production, the U.S. could face blackouts or delays in AI deployments.
Discussions around the AI boom often highlight grid challenges and energy innovation, but the turbine crunch is a stark reminder of manufacturing bottlenecks in the supply chain.
Even as companies repurpose sites—like imploding coal plants for new data centers—the equipment shortage persists, underscoring the need for investment in turbine manufacturing.
Implications for Energy and Beyond
Horizon exemplifies how AI is reshaping the energy landscape, blending tech innovation with fossil fuel infrastructure. By generating its own power, the project sidesteps some grid issues plaguing other developments, but it also spotlights the broader risks: skyrocketing energy demands that could strain resources and delay the transition to renewables.
In Ireland, for instance, power shortages have already forced Amazon to scrap a planned AI rack building facility.
For the Permian Basin, this could mean a new economic lifeline, extending the fracking era through AI-driven gas demand. However, it raises environmental questions—will AI’s growth lock in fossil fuel dependence, or spur cleaner innovations? As the industry races forward, addressing turbine shortages and grid vulnerabilities will be key to ensuring this boom doesn’t bust.
Stay tuned to Energy News Beat for more on how AI is transforming the energy sector.
Got Questions on investing in oil and gas? Or do you have a Tax Burden in 2025?
Crude Oil, LNG, Jet Fuel price quote
ENB Top News
ENB
Energy Dashboard
ENB Podcast
ENB Substack
Be the first to comment