The Box Elder data center and the Great Salt Lake: a water reality check
Utah's largest proposed data center would draw enormous amounts of water and energy at the north end of a lake that's already shrinking. Here's what's actually on the table — and the lever most of the coverage misses.
A 40,000-acre data center and on-site power campus — the "Stratos Project," backed by investor Kevin O'Leary — has been proposed in Box Elder County, at the northern edge of the Great Salt Lake. Box Elder County commissioners approved it on May 4, 2026, though developers are still raising money and full operation is likely many years away. The project has become a national flashpoint, and the loudest argument is about water. We build water-efficiency projects for Utah farms, so people keep asking us what to make of it. Here is a straight, sourced read.
The short version
- Scale: up to 9 gigawatts of power at full build — roughly double Utah's entire 2025 peak electricity demand — generated on-site by burning natural gas. Built in phases, with an initial cap near 1.5 GW.
- Water (power generation alone): Utah Clean Energy estimates roughly 2 billion gallons/year for gas engines, up to 16.6 billion gallons/year for combustion turbines — before counting cooling. Total demand has been estimated in the tens of thousands of acre-feet per year.
- Heat: about 16 gigawatts of waste heat dumped into Hansel Valley — a physicist likened it to "about 23 atom bombs' worth of energy" a day.
- The law changed: Utah's HB60 (effective May 6, 2026) narrowed the state's ability to reject a water right over broad public-welfare harm.
- Guardrails: Gov. Cox has said it will not run on 100% gas and has directed regulators to protect the Great Salt Lake.
What's actually being proposed
The Stratos Project is pitched as a combined data center and energy campus spanning roughly 40,000 acres. At full build it would need about 9 gigawatts of electricity — more than double what the entire state draws at peak — and the developers plan to generate all of it on-site by burning natural gas rather than drawing from the grid. It would be built in stages, with the first phase capped near 1.5 GW.
Timeline matters here. An initial O'Leary–Cox meeting happened in January 2026; the county commission approved the project in May; but the developers still need air-quality and natural-resource permits, still need to secure water rights, and by some accounts are still raising the money. Experts at the county hearing suggested the facility likely wouldn't be operating for roughly a decade. In other words, this is a proposal with a long runway, not a done deal — which is exactly why the rules being set now matter.
The water question
Water estimates vary widely because the technology isn't finalized and public information is limited. Utah Clean Energy modeled the two most likely gas-generation methods: reciprocating engines would use on the order of 2 billion gallons of water a year just to run, while combined-cycle combustion turbines would use roughly 16.6 billion gallons a year — about 25,000 Olympic swimming pools — and that's before the water needed to cool the data center itself. Other estimates put total demand for a turbine build near 50,000 acre-feet per year. Developers have discussed buying on-site water rights and have water under contract from the nearby town of Snowville.
Crucially, the public has already moved the project. After thousands of Utahns wrote to the Division of Water Rights, the backers withdrew an application to convert agricultural water to industrial use. That's a meaningful concession — and a sign that transparency and public pressure are working as intended.
Why the Great Salt Lake changes the math
The Great Salt Lake is a terminal basin: water flows in and only leaves by evaporation. When less reaches it, the lake shrinks and exposes a lakebed laced with arsenic and heavy metals, which can blow as toxic dust across the Salt Lake metro. The Stratos site sits at the lake's north end, near sensitive wetlands and the migratory-bird habitat the basin is famous for. Add the heat: the campus would release roughly 16 GW of waste heat into Hansel Valley — the "23 atom bombs" figure a Utah State University physicist used to describe the daily thermal load — raising local temperatures and, potentially, evaporation in the worst possible place.
The accountability gap
Two things happened almost in parallel. Utah passed HB60, which took effect May 6, 2026 and trimmed the state engineer's authority to deny a water right based on broad public-welfare, recreation, or environmental harm — narrowing the grounds on which ordinary Utahns can protest. At the same time, Governor Cox publicly drew lines: the project will not run on 100% natural gas, and he directed the Department of Natural Resources to require the most water-sensitive cooling technology and to protect the lake. So the public's formal leverage shrank, while the executive branch promised guardrails. Whether those guardrails are written into enforceable, verifiable conditions is the open question.
The part most coverage misses
Here's the context that rarely makes the headline. Even at the high end, a data center drawing tens of thousands of acre-feet a year is a real impact — but agriculture accounts for roughly 71% of the water depleted from the Great Salt Lake, and alfalfa, hay, and pasture cover more than 80% of Utah's irrigated land. No single industrial user comes close to that footprint. That's not a reason to wave the data center through; it's a reason to keep the scale honest and to point at where the water actually is.
And the agricultural lever is the one with money behind it. Utah's Agricultural Water Optimization Program cost-shares 50–75% of efficient-equipment costs, and the state has committed historic sums to getting water back to the lake. Every acre-foot a farm conserves and shepherds downstream is cheaper and faster to secure than fighting over a project that may be a decade out.
Where BasinWard stands
We're optimists about AI, and we use it every day to help farms save water. We believe the same force driving this story — better technology — should be making large language models, and the data centers that run them, dramatically more efficient: more intelligence per watt, and per gallon. Efficiency is the whole game. Just as we help farms get more crop per drop, the computing industry should be chasing more compute per drop and per watt, and competing on efficiency rather than raw scale.
That cuts both ways. Data centers should pursue those efficiency gains and take their environmental and ecological impact seriously — especially water, energy, and air at the edge of a shrinking Great Salt Lake. Efficiency and stewardship aren't opposites; they're the same discipline. A project of this size should be held to the standard we'd hold any water user to: measure your water and energy, publish a verified plan, use the most water-sensitive cooling, prove no net loss to the lake — and keep getting more efficient. Bring the data center; make it a model of efficiency and responsibility.
And while that gets sorted out, the fastest, cheapest water for the lake is already upstream — on farms that want to modernize but need a partner to design the project and win the funding. That's the work we do, and it's available today, not in ten years.
Farm in the Great Salt Lake basin?
See what you could save — in water and in cost-share — in about 30 seconds.
Run the free savings calculator →Sources
- Utah Clean Energy — Estimated emissions and water consumption from the proposed Stratos data center
- Utah News Dispatch — Renderings show vision for massive data center; permits could take years
- Utah News Dispatch — New water-rights law (HB60) gives less weight to broad-impact pushback
- Axios Salt Lake City — Utah law limits opposition to Box Elder data center
- The Salt Lake Tribune — Data center heat island and the Great Salt Lake
- Grist — Gov. Cox on the natural-gas plan
- Grist — Utah's fragile desert and the hyperscale data center
This analysis summarizes publicly reported information as of June 2026; figures are estimates drawn from limited public disclosures and will change as the project is reviewed. It reflects BasinWard's perspective and is not legal or policy advice. BasinWard is a Salt Lake City water-efficiency consultancy; it is not affiliated with the Stratos Project or its developers.