Anti-nuclear advocacy groups are shifting tactics to oppose data center projects that propose nuclear power, challenging a key energy solution for AI buildout.

Half a century ago, raw emotion fueled an anti-nuclear energy campaign that continues to stifle clean energy development in the United States. Today, that same fear-driven playbook is being deployed to target data centers.

Between 1970 and 1990, the United States constructed 95 nuclear power plants, which continue to supply nearly one-fifth of the nation’s electricity. However, the minor incident at Three Mile Island led to the cancellation of 67 additional planned reactors. The country did not build another nuclear power plant between 1996 and 2016. Even today, efforts to revitalize the domestic nuclear industry face persistent skepticism and legacy regulatory hurdles. Small modular reactors were first tested decades ago. Had anti-nuclear fearmongering not halted the Integral Fast Reactor program in 1994, these technologies could already be powering data centers.

Many Americans are rightfully alarmed by data centers depleting local drinking water reserves and inflating electricity costs. Others question whether these facilities are part of a vast conspiracy to strip away jobs, liberties, and even lives. Goldman Sachs estimates that U.S. data center power demand will surge from 31 gigawatts in 2025 to 66 gigawatts in 2027. The transition toward an all-electric economy has accelerated demand beyond the grid’s ability to expand new capacity. As demand outpaces supply, electricity prices rise. Neither the media nor the political establishment adequately prepared the public for these rapid shifts, making it unsurprising that data centers have become a polarizing political issue.

The current divide pits a repurposed anti-nuclear playbook—which risks crippling the American economy—against the acknowledgment that certain data center projects have severely disrupted local communities. Regardless of their necessity, such disruptions represent an unacceptable trade-off. Public opposition is fueled by documented cases, such as an Oregon community discovering that nearby Google data centers were consuming a quarter of its water supply. According to Consumer Reports, grid-connected data centers also tend to elevate electricity costs. These tangible impacts have galvanized even staunch advocates of economic growth.

Many developers have struggled to communicate the everyday utility of data centers: every movie showtime lookup, digital payment, cloud backup, or search query relies on this infrastructure. They have also fallen short in addressing legitimate community concerns. In March, President Trump convened the data center industry at the White House to sign the Ratepayer Protection Pledge. However, the commitment remains voluntary, and public skepticism persists given past instances where similar pledges went unfulfilled.

The Chinese government and its affiliated networks, which capitalize on public anxiety, are financing a coordinated campaign against data centers aimed at slowing U.S. economic growth. A May report from the Bitcoin Policy Institute identified three interlocking vectors demonstrating China’s targeted efforts against American AI data centers. These vectors include propaganda from Chinese and Russian state media; a U.S.-based NGO network financed by expatriate Neville Roy Singham; and billions in undisclosed foreign funding from sources including Swiss philanthropist Hansjörg Wyss and the Oak Foundation, which is backed by British billionaire Alan Parker. Lead author Sam Lyman argues that America faces a binary choice: not between adopting AI or rejecting it, but between American AI and Chinese AI. This mirrors the strategic decision between accepting continued Chinese dominance over critical minerals or pursuing an America-first mining and processing revival.

The Trump administration’s promotion of nuclear energy—for both grid integration and standalone units powering cloud and hyperscale data centers, military installations, and remote industrial sites—forms part of a broader national strategy to expand electricity generation without burdening consumers. This initiative quietly gained momentum during his first term. Following years of policy stagnation, President Biden issued Executive Order 14141, “Advancing U.S. Leadership in AI Infrastructure,” shortly before leaving office. The order authorized leasing federal land for gigawatt-scale data centers powered “with clean energy resources” to shield public rates from sharp increases. During the Ratepayer Protection Pledge signing ceremony, numerous leading AI and technology firms committed to absorbing the full cost of new power generation for their facilities, ensuring no negative impact on consumer rates.

White House AI and Crypto Advisor David Sacks contrasted this “common sense” approach with Senator Bernie Sanders’ call for a complete moratorium on new data centers. Sacks warned that such a ban would stall a nationwide construction boom, suppress wage and employment growth for skilled tradespeople, and ultimately cede technological leadership to China. Sacks further noted that permitting AI firms to fund grid modernization and sell surplus power back to the network could actually reduce consumer electricity bills. The primary obstacle remains that small modular reactors are still in early testing phases and must navigate a complex, legacy-permitting framework—a direct consequence of decades of anti-nuclear opposition.

As emerging data centers transition to independent, off-grid power generation, anticipated spikes in electricity costs should stabilize. The pressing question then becomes water consumption. Large data centers typically depend on water-based cooling towers or evaporative chillers to prevent server overheating and hardware failure. Although many operators have adopted closed-loop cooling systems—where water recirculates through sealed pipes or coils to prevent evaporation, reducing usage by up to 70%—public perception remains focused on fears of depleted municipal water supplies. Advanced alternatives such as direct-to-chip cooling, which routes chilled liquid directly to processors, and immersion cooling, which submerges entire servers in non-conductive fluids, further minimize water dependency.

Geographic advantages are also being leveraged, with Norway locating facilities near cold-water rivers and fjords, while temperate climates allow ambient air to regulate server temperatures. Meanwhile, Elon Musk is advancing plans to deploy orbital data centers connected via satellite and Starlink, eliminating terrestrial cooling requirements entirely. SpaceX and Nvidia intend to launch the first space-based data center by 2028. However, in rapidly evolving technological sectors, even minor delays can trigger irreversible competitive disadvantages.

Regulatory moratoria, such as New York Governor Kathy Hochul’s recent ban, and Pennsylvania Governor Josh Shapiro’s decision to remove data center projects from his state’s expedited approval pathway—established in November 2024—risk transforming temporary constraints into permanent barriers. Manhattan Institute senior fellow Shawn Regan contrasts these restrictive measures with Texas Governor Greg Abbott’s approach amid a competitive reelection campaign. Texas, he says, mostly refined its existing process “to better sort through a chaotic data-center development pipeline without abandoning growth.” Under Texas regulations, “Batch Zero” projects seeking grid interconnection must undergo comprehensive audits. Developers are required to disclose projected electricity and water consumption, on-site generation plans, cooling methodologies, tax incentive structures, and mitigation strategies for noise, traffic, and other localized impacts.

The divergence in policy approaches is stark. The same fear-driven politics that historically delayed nuclear development—now amplified by “No Kings” style protests (funded by Singham NGOs)—has pushed certain politicians toward alignment with Chinese interests. Policymakers aligned with President Trump’s vision of American technological leadership recognize the need to avoid discarding critical infrastructure investments. Realizing this balance requires leveraging the capabilities of major technology firms, despite lingering public skepticism regarding their alignment with broader societal interests.

Duggan Flanakin is a senior policy analyst at the Committee for a Constructive Tomorrow who writes on a wide variety of public policy issues.