President Trump's AI Order: The Missing Minerals, Megawatts, and Democratic Oversight Question

Highlights

• Trump's June 2026 AI executive order focuses on cybersecurity and innovation but omits the physical industrial base required to support AI deployment.
• AI data centers demand massive electricity, water, rare earth magnets, copper, and semiconductors—none of which the order addresses in supply chain terms.
• Communities from Utah to New York are pushing back against data center expansion, raising democratic questions about who pays for grid upgrades and water risk.
• A proposed Utah AI campus at 9 gigawatts would exceed double the state's average electricity use, sparking nearly 4,000 public objections and a gubernatorial response.
• Without addressing material sovereignty—minerals, power, cooling, and public consent—declarations of AI dominance remain incomplete industrial strategy.

President Donald Trump’s June 2, 2026 executive order on artificial intelligence (opens in a new tab) presents itself as a national-security blueprint for American AI dominance. Its language is sweeping: frontier models, cybersecurity, critical infrastructure, innovation, and rapid deployment. But from a Rare Earth Exchanges™ perspective, the order has a striking omission. It treats AI mostly as software.

In reality, AI is also a hard-asset industrial system. It requires electricity, water, land, cooling systems, semiconductors, copper, aluminum, gallium, germanium, rare earth magnets, specialty chemicals, grid equipment, and backup power. The executive order promotes AI deployment and cybersecurity coordination, but it says little about the material and energy burden required to support that ambition.

More Promotion Than Policy

The order does not create a new AI regulator. It does not impose a licensing system. It explicitly says it should not be interpreted as creating mandatory governmental preclearance for AI models. Much of the document is therefore political positioning: pro-innovation, anti-bureaucracy, cyber-focused, and closely aligned with private AI developers.

The operational changes are narrower. Agencies are told to prioritize cyber defense, expand cybersecurity services, identify existing grant funds, create an AI cybersecurity clearinghouse, and build classified benchmarks for advanced models. These steps matter, but they are administrative accelerators rather than a full national AI industrial strategy.

The Data Center Collision

The larger national tension is already visible in Utah, Rare Earth Exchanges’™ home state. A massive proposed AI data center campus in Box Elder County, known as Stratos, has reportedly been planned at up to 9 gigawatts of power capacity—more than double Utah’s current average electricity use, according to Tom’s Hardware and other reports. Intense local backlash emerged: nearly 4,000 objections and a referendum effort after approval of a project spanning roughly 40,000 acres.

Utah Gov. Spencer Cox responded with an executive order tightening expectations for data-center development, including protections for water, the Great Salt Lake, wildlife, air quality, public utility rates, and public engagement.

This is not just a Utah story. New York lawmakers are weighing a one-year pause on new AI data centers. Ohio is seeing ballot-style resistance and reconsideration of tax breaks. North Carolina is considering new guardrails tied to water, noise, and electricity costs. Coachella, California has faced major public protests over a proposed data center campus.

The Populist Pushback

The emerging backlash is not simply environmentalism. It is populist infrastructure politics. Communities are asking basic questions: Who gets the jobs? Who pays for the grid upgrades? Who owns the water risk? Do residents get higher utility bills while AI firms receive tax breaks? Will rural landscapes be converted into industrial compute zones with limited local consent?

Those questions are democratic questions. They are also supply-chain questions.

A 2026 academic analysis warned that U.S. data centers could require hundreds of millions to more than one billion gallons per day of additional water capacity by 2030 under high-growth assumptions. Another recent study found AI data-center growth is becoming a structural stressor for regional power systems, especially where compute capacity is concentrated.

What the Order Leaves Out

The executive order says little about the physical economy beneath AI.

It does not discuss grid expansion. It does not address water-intensive cooling. It does not connect AI deployment to critical mineral supply chains. It does not explain how the United States will source the metals, magnets, power electronics, transformers, chips, cables, cooling equipment, and backup systems needed for a national AI buildout. That omission matters. AI dominance cannot be declared into existence. It must be mined, refined, manufactured, wired, cooled, powered, and permitted.

The Democratic Oversight Problem

The order also raises a governance concern. Much of its important work happens inside the executive branch through classified benchmarks, cyber directives, voluntary industry frameworks, and public-private coordination. Congress is largely absent. So are state legislatures, public utility commissions, local governments, and affected communities.

The order may be defensible as a cybersecurity measure. But if AI policy increasingly moves through classified processes and private-sector partnerships, while the physical burden lands on local grids, aquifers, ratepayers, and landscapes, democratic oversight becomes dangerously thin.

Food for Thought

This executive order is partly policy and partly performance. It accelerates federal AI-cybersecurity coordination. It signals that Washington wants American AI dominance without heavy model licensing. But it does not offer a serious account of the industrial base required to support that ambition.

From the REEx perspective, the missing sentence is simple: There is no AI sovereignty without material sovereignty. The next AI fight will not only be about algorithms. It will be about power, water, cooling, copper, rare earths, semiconductors, grid equipment, and public consent. The communities resisting data centers are not standing in the way of the future. They are asking who will pay for it.