Data centers are often portrayed as invisible engines of the digital economy, powering artificial intelligence, cloud computing, streaming platforms and billions of daily online interactions. Yet behind every search request, video call and machine-learning model lies a physical infrastructure that consumes electricity, occupies land and requires enormous quantities of cooling. A new study by J. Heering and E. Voeten, published in Nature Communications, examines a question that is becoming increasingly urgent worldwide: what makes people support, or oppose, the construction and expansion of data centers?
The research focuses on three forces that may determine public opinion: sovereign control, decarbonization and energy costs. Together, these factors capture a central contradiction of the digital age. Data centers can strengthen a country’s technological independence and attract investment, but they can also intensify pressure on electricity grids and raise questions about who controls critical digital infrastructure. At the same time, their social acceptance may depend on whether the energy they consume is produced with low carbon emissions—and who ultimately pays for it.
The issue has moved rapidly from a niche planning concern to a global political flashpoint. The explosive growth of artificial intelligence has created demand for increasingly powerful computing facilities, many of which contain tens of thousands of specialized processors. These machines perform the calculations required to train and operate large AI models, but they generate substantial heat. Keeping them within safe operating temperatures requires sophisticated cooling systems, which can consume additional electricity and, in some locations, significant volumes of water.
A data center’s total energy demand is shaped by more than the number of computers inside it. Its efficiency is often assessed through power usage effectiveness, or PUE, a technical ratio comparing the facility’s total electricity consumption with the power used directly by computing equipment. A PUE close to 1 indicates that most electricity is reaching servers rather than cooling, lighting or other support systems. Even highly efficient facilities, however, can have a large environmental footprint when their computing capacity is expanded at extraordinary speed.
This creates a difficult policy challenge. Electricity demand from data centers can compete with households, factories and transportation systems that are also being electrified to reduce fossil-fuel use. In regions where power grids are already strained, new facilities may require additional generation, transmission lines or long-term contracts that affect electricity prices. If the additional power comes from coal or natural gas, the climate benefits of digital growth may be weakened. If it comes from renewable sources, the result may be cleaner—but renewable electricity is limited, variable and increasingly contested by other users.
Heering and Voeten’s study places these practical concerns alongside the idea of sovereign control. Data centers are not merely warehouses filled with computers; they are strategic nodes in national infrastructure. Governments may view domestic computing capacity as essential for cybersecurity, economic competitiveness, scientific research and the operation of public services. Hosting data and AI systems within national borders can reduce dependence on foreign providers and give governments greater authority over sensitive information. That promise of digital sovereignty may increase public support even when local communities remain concerned about costs or environmental impacts.
The study’s framing is significant because public reactions to data centers are rarely determined by technology alone. People may evaluate the same facility differently depending on whether they believe it serves a national interest, creates local jobs, contributes tax revenue or supports public research. They may also ask whether the benefits are distributed fairly. A project that supplies low-cost computing for hospitals or universities could be viewed differently from one designed primarily to increase the profits of global technology companies while placing new demands on local infrastructure.
Decarbonization adds another layer of complexity. A facility powered by wind, solar, nuclear energy or other low-carbon sources may appear more acceptable than one dependent on fossil fuels. But “renewable-powered” does not automatically mean impact-free. Large electricity users can influence when and where power is consumed, potentially increasing the need for storage or backup generation when renewable output falls. The climate performance of a data center therefore depends not only on annual energy certificates, but also on the timing of its electricity use, the carbon intensity of the local grid and the broader consequences for energy-system planning.
Energy costs may be just as influential as environmental concerns. Data centers often negotiate electricity contracts and seek locations where power is reliable and affordable. Communities, however, may worry that infrastructure built to serve large commercial users will be financed through higher bills, public subsidies or reduced grid reliability. These questions are especially sensitive during periods of inflation or energy insecurity, when households are already struggling with rising costs. Public support can become fragile if digital expansion is perceived as shifting economic burdens from technology companies to ordinary consumers.
By connecting sovereignty, decarbonization and energy prices, the researchers highlight why the future of data centers cannot be decided solely through corporate investment or technical efficiency. The public debate is ultimately about control, fairness and trade-offs: who benefits from expanded computing capacity, who bears its environmental and financial costs, and whether digital progress can be aligned with climate goals. As AI adoption accelerates, those questions are likely to determine where the next generation of data centers is built—and whether the public sees them as engines of national strength or symbols of an increasingly expensive digital economy.
Subject of Research: Public support for data centers, with a focus on sovereign control, decarbonization and energy costs.
Article Title: How sovereign control, decarbonization and energy costs shape public support for data centers.
Article References: Heering, J., Voeten, E. How sovereign control, decarbonization and energy costs shape public support for data centers. Nat Commun (2026). https://doi.org/10.1038/s41467-026-76500-9
Image Credits: AI Generated
DOI: 10.1038/s41467-026-76500-9
Keywords: Data centers, public support, sovereign control, decarbonization, energy costs, artificial intelligence, electricity demand, digital infrastructure, energy policy, climate change
Tags: AI and data center energy demandsclimate implications of data center expansioncritical digital infrastructure governanceData center public supportdecarbonization and renewable energy in data centersdigital economy infrastructuredigital infrastructure and land useenergy consumption and environmental impact of data centersenergy costs and public acceptanceglobal policy on data center developmentsocietal perceptions of digital infrastructuresovereignty and digital independence
