Data centers are becoming the new flashpoint in America's energy wars, with rural communities mounting a bipartisan backlash that echoes the decade-long battles over pipelines and fracking. From Nebraska to West Virginia, residents are organizing against massive AI facilities, citing concerns about water depletion, power grid strain, and land use. Unlike previous infrastructure fights, this opposition is happening as hyperscalers like Microsoft, Amazon, Google, and Meta are aggressively securing gigawatts of power capacity to fuel the AI boom.

The backlash is real and spreading fast. Willie Nelson, who once protested the Keystone XL pipeline, now opposes data centers. Jane Kleeb, a Nebraska anti-pipeline organizer, is now active in decentralized data center opposition efforts. Analysts expect data centers to become a top midterm election issue, with states weighing new power and water regulations and considering repealing tax breaks that lure these facilities.

Why Are Communities So Angry About Data Centers?

The opposition mirrors the political playbook from fossil fuel fights. Rural residents worry about three core issues: water consumption, electricity costs, and the concentration of wealth and power in the hands of tech giants. However, much of the public debate is built on incomplete information and worst-case scenarios stripped of context.

Take water consumption. Stories about data centers often lead with shocking numbers: one facility might use millions of gallons daily, or a state might be projected to consume billions annually. But scale matters enormously. A 2026 analysis by the Florida Water and Pollution Control Operators Association found that all U.S. data centers combined account for roughly 0.3% to 0.4% of total daily U.S. water withdrawals, based on 2021 figures. For perspective, U.S. golf courses use roughly 30 times as much water annually, while California almond production alone uses well over 100 times as much.

The electricity numbers are similarly large but require context. Lawrence Berkeley National Laboratory estimated that U.S. data centers consumed about 200 terawatt-hours of electricity in 2025, roughly 4.7% of total national electricity consumption. Yet critics often assume this demand automatically translates to higher household bills. The evidence is more complicated. Some research suggests recent data-center growth has had only limited effects on retail rates, while broader inflation and energy-policy choices may explain more of the increases consumers have seen.

How Are Tech Giants Securing Power for AI Data Centers?

Hyperscalers are taking dramatic steps to lock in reliable energy. The scale of their commitments is staggering:

• Microsoft's Nuclear Bet: Signed a 20-year agreement to restart the Three Mile Island nuclear plant in Pennsylvania, offline since 2019, specifically to supply roughly 835 megawatts of power for AI workloads.
• Amazon's Campus Acquisition: Purchased an 8,000-acre data center campus in Texas for a project that could include up to 7.65 gigawatts of on-site natural-gas generation. Also acquired a data center campus co-located with the Susquehanna nuclear station in Pennsylvania for $650 million.
• Google's Modular Reactors: Announced agreements with Kairos Power for small modular reactors to power its facilities.
• Meta's Capacity Request: Issued a request for proposals seeking up to 4 gigawatts of new nuclear capacity.
• International Expansion: Bitzero Holdings secured a 15-year, $2.6 billion lease for AI workloads at its Norway site, leveraging Nordic hydroelectric power.

These deals reflect the sheer scale of AI's energy demands. AI data center capital expenditure is projected at $5.2 trillion by 2030, with power demand surging 165% by 2030. That's not a typo: power demand is expected to nearly triple in four years.

What Are the Real Concerns Versus the Exaggerated Ones?

The strongest concerns about data centers may not be the ones dominating headlines. While water and electricity debates grab attention, the deeper issues deserve more scrutiny. As AI becomes embedded throughout society, these systems could shape what information people see, how institutions make decisions, and how individuals interact with government and business. Massive data collection could strengthen surveillance capabilities. Ideological biases could be embedded in algorithms. Dependence on AI could weaken personal autonomy and concentrate decision-making power in governments and corporations.

Job displacement is another legitimate concern, though the historical record suggests caution about apocalyptic predictions. In 2016, AI pioneer Geoffrey Hinton argued that people should stop training to become radiologists because machines would soon outperform humans at reading medical images. A decade later, radiology is hardly disappearing. Research shows the number of active U.S. radiologists rose over the past decade, salaries increased, and workloads expanded along with the use of medical imaging.

This pattern reflects what economists call "Jevons paradox": when technology makes a resource or service cheaper and more efficient, total demand can rise. If AI makes a programmer five times more productive, it's tempting to assume the economy will need one-fifth as many programmers. But economies are not static. If software becomes dramatically cheaper to produce, businesses that could never afford custom software may suddenly buy it. New applications become economical. Entirely new product categories can emerge.

Why Are Tech Companies Losing the Public Trust?

Analysts point to a fundamental communication failure. Tech firms have lost the public's trust by touting economic benefits rather than listening to residents' actual concerns. The Data Center Coalition says it's working to better demonstrate its commitment to community dialogue and fair water and power use. But the damage is already done in many communities.

Some hyperscalers are also creating barriers to rural connectivity even as they lay fiber across the country. As hyperscalers build fiber to link their data centers, some are contractually barring off-ramps to towns those lines pass through, keeping rural America disconnected even as high-capacity cable runs right by. Some contracts prohibit fiber builders from tapping the conduit for spurs to communities or even laying spare "shadow conduit" that could later serve them. Firms prize speed, cost, and simplicity, wanting clean routes with no breakouts.

Not all companies follow this approach. Zayo said it has no such restrictions and seeks to connect underserved areas along its routes, including an 800-mile open-access build in Nevada.

What Are States Doing to Address Data Center Concerns?

Some states are trying to balance economic opportunity with resident protection. West Virginia Governor Patrick Morrisey unveiled a seven-point framework to court data centers while addressing public opposition. Under a 2025 law that strips local control in favor of statewide standards, half the sector's tax revenue would go toward cutting and possibly eliminating state income taxes; the rest splits among host counties, other counties, and local infrastructure. The plan promises ratepayer and water protections and a local "voice" through a new advisory council.

Utah became the first state to require data centers to disclose their water use, a transparency push that could matter for rural communities and a drought-stressed West where the facilities compete with farms and households for scarce water. The law, signed by Governor Spencer Cox in March, makes new data centers over 10,000 square feet using at least 75 acre-feet a year report water sources, uses, and conservation to state regulators; aggregated data goes public. Nebraska has since copied the approach.

The bottom line: America faces a genuine tension between the energy demands of AI innovation and the legitimate concerns of rural communities. The only way to navigate this intelligently is to start with facts rather than numbers selected for maximum panic, while taking seriously the real questions about privacy, surveillance, and concentrated power that data centers raise.