A scaled-back Stratos Project signals a new era of negotiated hyperscale development in Utah
Kevin O’Leary’s O’Leary Digital has agreed to materially reduce the land footprint of its proposed Stratos Project data center complex in Box Elder County, Utah, surrendering 19,430 acres near the Locomotive Springs area—nearly half of the originally discussed 40,000-acre expanse. The move follows direct political intervention from Utah Senate President J. Stuart Adams, who pressed for a far steeper reduction to protect the adjacent Locomotive Springs Waterfowl Management Area and to address mounting local concerns around water consumption, energy demand, and ecological disruption.
The revised footprint—now approximately 20,570 acres—is more than a cartographic adjustment. It reflects a broader shift in how large-scale digital infrastructure is being bargained, permitted, and socially legitimized. In today’s market, hyperscale data centers are no longer evaluated solely on their promise of capital investment and job creation; they are increasingly measured against their hydrological impact, grid burden, habitat adjacency, and community benefit commitments.
Notably, O’Leary Digital maintains that the project’s employment projections remain unchanged, citing roughly:
- 4,500 construction jobs on average over the build-out period
- 2,500 permanent operations roles once fully operational
- Potential advanced-manufacturing spinoffs, still under evaluation
That continuity in job claims—despite a major land concession—will likely become a focal point for both supporters and skeptics, as stakeholders seek clarity on what exactly is being built, at what density, and with what resource profile.
Power at the scale of a small grid: why 7.5–9 GW changes the conversation
The Stratos Project’s most consequential detail may not be acreage at all, but its estimated 7.5 to 9 gigawatts (GW) of power demand—a figure that places it among the most ambitious U.S. hyperscale proposals on record. For context, a multi-gigawatt campus can function like a new city’s worth of load, forcing utilities and regulators to confront questions that go well beyond routine interconnection:
- Grid reliability and congestion: Can transmission and generation scale without displacing other economic activity or raising consumer costs?
- Resource adequacy: What firm capacity backs intermittent renewables, especially during peak demand or extreme weather?
- Infrastructure sequencing risk: If compute halls arrive before substations, lines, and generation, the project can become a stranded asset—or a political flashpoint.
- Decarbonization credibility: Large loads increasingly face scrutiny over whether “green” procurement is additional and local, or merely contractual.
This is where the AI boom becomes tangible. The surge in AI training, inference, and cloud workloads is translating into a physical arms race for land, power, and cooling capacity. Stratos illustrates the new baseline: hyperscale developers are being pushed to show not only that they can buy electricity, but that they can do so in a way that is grid-supportive, low-carbon, and verifiably resilient.
As a result, the most bankable projects are likely to be those that pair scale with modern grid strategies—such as long-term power purchase agreements (PPAs), on-site or grid-scale battery storage, and demand-response programs that allow operators to modulate load when the grid is stressed.
The water–cooling dilemma in an arid region, and the rise of “social license” metrics
If power is the headline, water is the permitting battlefield—especially in the American West. Data centers have multiple cooling pathways, and the public debate often hinges on whether a campus relies on evaporative cooling, chiller-based systems, or newer approaches that can reduce freshwater dependency. In Utah, where the Great Salt Lake and surrounding ecosystems have become symbols of broader water stress, the Stratos Project enters a regulatory environment increasingly shaped by water accounting and ecosystem outcomes.
Crucially, the project remains at an early stage: no permits or approvals have been filed, and comprehensive environmental review—including environmental impact statements, hydrological modeling, and wildlife habitat studies—is still pending. Also unresolved are written commitments regarding water allocation and any contribution framework tied to the Great Salt Lake.
That sequencing matters. In many jurisdictions, the era of “announce first, permit later” is giving way to a more conditional model in which developers must demonstrate, early and quantitatively, how they will manage:
- Water sourcing and consumption intensity (including seasonal variability)
- Cooling technology selection and contingency operations during drought
- Habitat buffers and construction impacts near protected areas
- Community benefit agreements (CBAs) that translate scale into local value
Utah’s intervention—demanding major land reduction to protect a wildlife management area—signals a maturing posture: economic development remains welcome, but only if paired with enforceable safeguards. For the data center industry, this is a preview of a future in which “social license to operate” is not rhetorical; it is negotiated in acreage, water terms, and monitoring requirements.
What Stratos may foreshadow for AI infrastructure siting across the U.S.
The Stratos Project’s evolution—from expansive proposal to politically moderated footprint—captures a broader recalibration in data center site selection. States and rural communities are learning that hyperscale developers need them as much as they need the investment, and that leverage is increasingly exercised through environmental conditions, resource covenants, and public-value capture mechanisms.
Several forward-looking implications stand out for investors, utilities, and technology operators:
- Cooling innovation becomes a permitting advantage: Expect stronger momentum behind closed-loop, direct-to-chip, and immersion cooling designs that reduce freshwater reliance.
- Energy procurement shifts from “available” to “defensible”: Projects will be judged on whether their power strategy is additional, reliable, and locally grid-compatible.
- Edge and distributed architectures regain appeal: Not as a replacement for hyperscale, but as a risk hedge—smaller facilities can be sited nearer renewables, transmission, or less constrained water basins.
- Environmental offsets broaden beyond carbon: Water credits, habitat enhancement, and watershed restoration may become standard components of large digital infrastructure deals.
For O’Leary Digital, the downsizing is a tactical concession that may reduce political friction while increasing execution complexity. For the industry, it is a strategic signal: the next generation of AI-era data centers will be permitted not just with capital and kilowatts, but with measurable stewardship, transparent modeling, and durable community agreements that can withstand scrutiny long after the ribbon-cutting.
By
By
By
By
By
By
By
By
