Data Centers: Necessary Infrastructure with Legitimate Local Costs

Digital technology still has a physical footprint

Cloud computing can feel almost weightless. A request leaves a screen, a result returns, and the machinery remains out of sight. Behind that experience are servers, cooling systems, electrical equipment, networks, construction, land, and people. AI makes this relationship harder to ignore because advanced models require substantial computing capacity.

I believe we should continue exploring AI, robotics, and other emerging technologies. They may improve research, accessibility, productivity, health, education, and public services. But support for progress is stronger—not weaker—when it acknowledges the infrastructure and community decisions that make progress possible.

The scale is growing, and the range matters

Berkeley Lab's 2025 update estimates that U.S. data centers could account for 11.8 percent of national electricity use in 2030, with scenarios ranging from 9.5 to 15.3 percent. That range is important. It is not a single guaranteed outcome; it reflects uncertainty involving technology, deployment, efficiency, utilization, and broader electricity demand.1

The earlier Department of Energy-sponsored 2024 study estimated that data centers used about 4.4 percent of U.S. electricity in 2023 and could reach 6.7 to 12 percent by 2028. The estimates have moved as the market has changed. Communities and utilities therefore need current scenarios, transparent assumptions, and plans that can adapt—not a promise that one forecast will remain correct.2

Growth at this scale can affect generation, transmission, local distribution, land use, water demand, construction, tax policy, and decisions about who pays for new capacity. Those are legitimate public questions even when the underlying technology is valuable.

A false choice makes the conversation worse

One side of the debate can sound as if every data center is proof of progress and every objection is resistance to the future. The other can sound as if the infrastructure has no public value and should simply be stopped. Neither position is strong enough for the actual decision.

The benefits and burdens are not automatically distributed to the same people. A region may gain construction work, investment, tax revenue, and stronger digital infrastructure while particular residents worry about utility costs, water, land, noise, or whether promised jobs will last. Responsible analysis should identify which effects are measured, which are projected, who receives them, and who carries the uncertainty.

The Department of Energy's data-center design guidance emphasizes reducing energy use, improving power and cooling efficiency, and considering water and renewable-energy strategies. Technical choices can reduce impact, but good design does not replace public accountability. Both are needed.3

A practical test for responsible growth

I would evaluate a proposed data-center development through a set of questions that can be answered with evidence and enforceable commitments. The goal is not to create an impossible standard. It is to make approval, investment, and community benefit more credible.

• Demand: What power and water ranges are expected at opening, full buildout, and peak operation?
• Efficiency: Which measurable design and operating standards will apply, and how will performance be reported?
• Infrastructure: What generation, transmission, distribution, or water upgrades are required, and who pays for them?
• Resilience: How will the project respond to drought, grid stress, emergencies, equipment changes, or demand that differs from forecasts?
• Community value: Which jobs, tax benefits, infrastructure improvements, or other commitments are durable and verifiable?
• Accountability: What information will remain public after approval, and what happens if the project misses its commitments?

Progress earns trust through transparency

Technology development and community protection should not be competing slogans. They are design requirements for the same future. If AI and cloud services are important, then building their infrastructure responsibly is also important.

My current position is therefore conditional but optimistic: continue investing, experimenting, and building—but expect specific evidence, efficient design, realistic infrastructure plans, and clear community commitments. The future will need computing capacity. It will also need public trust.