Utah Approves 9GW Super AI Campus Requiring More Than Twice the State's Total Power Consumption

Utah Bets on AI: Unprecedented 9GW Super Data Center Campus Approved

The U.S. state of Utah has officially approved a highly controversial hyperscale AI infrastructure project — an AI data center campus with a total designed capacity of 9 gigawatts (GW). Remarkably, the facility's full-load power demand would exceed more than twice Utah's current statewide electricity consumption. The decision not only marks an aggressive expansion of AI infrastructure in the United States but also pushes the global AI industry's energy anxiety to new heights.

What Does 9GW Mean?

To grasp the staggering scale of this figure, some context is needed. Utah's current total installed power capacity is approximately 4GW, serving the daily electricity needs of roughly 3.3 million residents. The planned AI campus alone would require 9GW of power supply — equivalent to the generating capacity of about nine large nuclear power plants, or enough electricity to power millions of homes.

From a global perspective, a 9GW data center cluster would be one of the largest single-purpose power-consuming facilities in human history. The world's largest data centers today typically operate in the hundreds-of-megawatts range, and even next-generation hyperscale data centers planned by tech giants like Microsoft and Google mostly fall within the 1–2GW range per facility. The 9GW master plan means Utah is attempting to build an unprecedented AI computing "super hub."

Why Utah?

Several factors make Utah the landing spot for this mega-project:

Abundant Land Resources: Utah has vast tracts of federal land and undeveloped parcels capable of meeting the physical space requirements of a hyperscale campus. Data centers need not only server halls but also supporting cooling systems, substations, and security buffer zones.

Business-Friendly Policy Environment: Utah's state government has long been known for its pro-business policies. Tax incentives and streamlined approval processes are highly attractive for large-scale tech infrastructure projects. The rapid approval also reflects the state government's high expectations for AI's economic multiplier effect.

Favorable Climate Conditions: Located in the inland western United States, Utah's dry and cool climate helps reduce data center cooling costs, which account for a significant portion of operating expenses.

Strategic Geographic Advantage: Its western U.S. location enables effective connectivity with tech hubs like Silicon Valley and Seattle while complementing other data center clusters.

Power Supply: The Biggest Open Question

Although the project has been approved, where the 9GW of electricity will come from remains the most critical challenge. Utah's current power mix is dominated by coal and natural gas, and while the share of renewables is growing, it remains limited. Providing stable power to this campus will likely require a combination of the following approaches:

• Large-Scale New Power Generation Facilities: Including natural gas plants, solar and wind farms, and potentially small modular reactor (SMR) nuclear technology
• Interstate Power Procurement: Importing electricity from neighboring states through the western grid
• Phased Construction: The campus will most likely not operate at full capacity all at once but will be built in phases with gradual expansion

Notably, such a massive surge in power demand could drive up electricity costs for local residents and businesses — one of the key points of public scrutiny the project faces.

Energy Anxiety in the AI Arms Race

Utah's 9GW project is not an isolated case but rather a microcosm of the global AI computing arms race. As the parameter counts of large language models continue to climb, the computational resources required for training and inference are growing exponentially, and tech giants' demand for data centers has entered the "GW era."

The International Energy Agency (IEA) previously projected that global data center electricity consumption could double by 2026, reaching approximately 1,000 terawatt-hours (TWh) — equivalent to Japan's total national electricity consumption. AI workloads are the primary driver behind this growth.

Within the United States, a "land grab" for AI data centers is unfolding simultaneously across multiple states. Virginia, Texas, Ohio, and others are all competing for large AI infrastructure projects, rolling out tax breaks and expedited approvals. But Utah's 9GW-class project has undeniably pushed the competition to an entirely new order of magnitude.

Controversies and Concerns

The project's approval has not come without significant controversy:

Environmental Impact: Such massive power consumption will generate an enormous carbon footprint, especially during the transition period when fossil fuels remain in use. Additionally, data center cooling systems typically require large volumes of water, and Utah, situated in an arid region, already faces tight water resources.

Grid Stability: Concentrating a load exceeding twice the state's total capacity in a single campus places extremely high demands on grid planning, construction, and dispatch capabilities. Any failure could trigger cascading effects.

Economic Risk: Infrastructure investments of this scale involve long cycles and uncertain returns. Should AI industry growth fall short of expectations or a major shift in technology roadmaps occur, the project could face enormous sunk costs.

Social Equity: Whether the massive power draw by data centers will affect local residents' electricity prices and supply reliability is the public's most immediate concern.

Outlook: A 'New Normal' for AI Infrastructure?

The approval of Utah's 9GW AI campus reflects an unmistakable trend: AI competition is moving beyond the software layer and deep into the realms of energy and infrastructure. Behind computing power is electrical power, and behind electrical power is energy strategy.

In the future, the ability to secure sufficient, stable, and clean electricity supply may become one of the key variables determining the competitive landscape of AI. Innovations in the energy sector — the nuclear energy revival, next-generation energy storage technologies, and ultra-high-voltage transmission networks — will be just as important as advances in AI technology itself.

This project also sends a signal to the world: in the AI-era infrastructure race, boldness and speed may be just as important as technology. But how to strike a balance between development ambitions and environmental responsibility and public interest is a question that Utah — and the entire industry — must answer.