Why Kevin O’Leary’s AI Project Could Overload Utah’s Grid

By VirentaNews Staff — May 10, 2026

In a proposal that has stunned energy analysts and environmental observers alike, billionaire investor Kevin O’Leary has unveiled plans for a 9-gigawatt hyperscale data center in Utah—one that would consume more electricity than the entire state currently uses. If built, the facility would require enough power to run approximately 7 million U.S. homes, and its waste heat output would equal the thermal energy released by 23 Hiroshima-sized atomic bombs every day. This staggering energy footprint underscores the rapidly escalating demands of artificial intelligence infrastructure, where training large language models and running massive computing clusters now rival the energy needs of small nations. As AI adoption surges, the environmental and infrastructural consequences of such projects are forcing a reckoning over how sustainable this growth truly is.

A New Benchmark in Computing Power

O’Leary’s proposed data center, dubbed a “hyperscale” facility, represents a quantum leap in digital infrastructure. At 9 gigawatts (GW), it would dwarf even the largest existing data centers—most of which operate at less than 1 GW. For context, the entire state of Utah consumes approximately 4.5 GW of electricity on average, according to the U.S. Energy Information Administration. The project, likely designed to support AI model training, cloud computing, and high-frequency financial modeling, signals a new era where private enterprise can outpace regional energy capacity. While O’Leary has not disclosed all technical or financial details, the project reflects broader industry trends: companies like Google, Microsoft, and Amazon are rapidly expanding their data center footprints to meet the computational demands of generative AI, which requires exponentially more processing power than traditional software.

The Anatomy of a Power-Hungry Data Center

The proposed facility would be located in Utah’s Salt Lake City metro area, leveraging the region’s fiber-optic connectivity, relatively low land costs, and political support for tech investment. O’Leary, known for his role on “Shark Tank” and his shrewd investments in tech startups, has described the project as a “sovereign AI campus” designed to provide secure, high-performance computing for enterprise clients, including financial institutions and government contractors. While he has not named partners, the scale suggests involvement from major chipmakers like NVIDIA or cloud providers such as Oracle. The 9 GW demand would likely be met through a mix of existing grid infrastructure, direct power purchase agreements with renewable providers, and possibly on-site natural gas generation. However, experts question whether Utah’s grid, already strained during peak winter demand, can support such a load without major upgrades and transmission bottlenecks.

Energy, Heat, and Environmental Calculus

The environmental implications of the project are profound. Beyond its electricity consumption, the data center would generate an estimated 6.9 terajoules of waste heat per day—equivalent to the thermal energy of 23 Hiroshima bombs, each releasing about 300 trillion joules. While waste heat from data centers is typically dissipated via cooling towers or liquid systems, managing this volume would require unprecedented engineering. Some European data centers now repurpose waste heat for district heating, but Utah’s climate and urban layout may limit such options. According to a 2023 report by Nature, AI-related energy demand could account for up to 3% of global electricity use by 2026, up from less than 1% in 2020. O’Leary’s project alone would represent a significant fraction of that growth, raising concerns about carbon emissions, water usage for cooling, and long-term grid stability.

Stakeholders and Systemic Risks

The proposed data center would impact a wide range of stakeholders. Utah residents could face higher electricity rates if utilities pass on infrastructure upgrade costs. Environmental groups warn of increased strain on the Colorado River watershed, already over-allocated and suffering from prolonged drought. Meanwhile, tech companies and investors see such facilities as critical to maintaining U.S. competitiveness in AI. The Department of Energy has acknowledged that data centers could account for 12% of national electricity demand by 2030, up from 4% in 2022. Municipalities, eager to attract high-tech jobs, may be tempted to offer tax incentives and regulatory leniency, potentially at the expense of long-term sustainability. The project also raises questions about energy equity: should a single private facility be allowed to monopolize a region’s power supply, potentially at the cost of residential and industrial users?

Expert Perspectives

Reactions from energy and tech experts are sharply divided. Dr. Arjun Krishnan, a power systems researcher at MIT, warns that “unconstrained data center growth risks destabilizing regional grids and undermining climate goals.” In contrast, tech economist James Manyika argues that “AI infrastructure is the new oil—strategic, essential, and worth investing in.” Some propose regulatory frameworks to cap energy use per data center or mandate heat reuse. Others suggest shifting more AI workloads to off-peak hours or to regions with surplus renewable capacity, such as offshore wind zones. The debate mirrors broader tensions between innovation and sustainability in the digital age.

As scrutiny intensifies, the fate of O’Leary’s project will likely hinge on regulatory approvals, utility cooperation, and public sentiment. If approved, it could set a precedent for similarly massive facilities across the U.S. and beyond. The key question remains: can society harness the transformative power of AI without overwhelming the very systems that sustain it? The answer may define the next decade of technological development.

Source: Sltrib