The opinions expressed in this piece are solely the author’s and do not necessarily reflect the views of Carolina Journal or its publisher.
Artificial intelligence (AI) and cloud computing are becoming increasingly central to the modern economy, driving business intelligence, operational efficiency, and data-driven decision-making across industries. From generative AI tools to routine digital services, these technologies depend on an expanding web of physical infrastructure — most notably data centers.
Data centers deliver the computational power, storage, and connectivity that keep digital systems running, but they also place significant demands on energy and water resources. As demand for AI functionality surges, North Carolina has emerged as a prime destination for data center investment, placing the state at the intersection of economic opportunity, infrastructure capacity, and growing community concern.
North Carolina offers a compelling mix of advantages for data centers: relatively affordable land; access to major fiber routes; proximity to leading research universities such as North Carolina State University, University of North Carolina at Chapel Hill, and Duke University; and a historically business-friendly regulatory climate. These factors have already drawn hyperscale operators supporting cloud computing and AI workloads.
As reported by the News & Observer, the scale of proposed investment is striking: a $6.4 billion hyperscale data center project in Tarboro was projected to generate roughly $11 million annually in local property-tax revenue, while a separate $30 billion proposal in Mooresville — later withdrawn amid rezoning concerns — illustrates both the magnitude of capital at stake and the volatility of local approval processes.
It is understandable why there would be reservations about embracing data center expansion in local communities. Despite the economic promises, reporting from the previously mentioned News & Observer article and from the Charlotte Observer on a proposed data center in Matthews shows mounting resistance at the local level, with residents worried that the costs of hosting massive facilities may outweigh the benefits. In towns such as Tarboro and Matthews, concerns have centered on noise from cooling systems and backup generators, the visual scale of industrial buildings, heavy water use, and — most prominently — the strain placed on local electricity grids.
In Apex, similar anxieties have emerged around a proposed 190-acre data center near the Shearon Harris area. Residents questioned whether the project aligns with the town’s character and infrastructure capacity, particularly with respect to electricity and water demand.
“The people don’t want this,” one resident told ABC11. “The people that know about it don’t want it. And the people that don’t know about it — when they find out — they won’t want it either.”
Others raised concerns about proximity to homes and long-term livability. “I don’t want something that’s going to affect my family two miles from my home,” said Dr. Michelle Hoffner O’Connor, reflecting a broader fear that rapid industrial-scale development could permanently alter nearby residential communities.
These disputes have also exposed a deeper structural challenge in North Carolina’s land-use and permitting system. When developers seek special-use permits, local governments are required to act in a quasi-judicial role — functioning more like courts than legislative bodies. That framework sharply limits officials’ ability to consider public sentiment or broader policy implications.
“You can’t have public input,” Tarboro Town Manager Troy Lewis told the N&O. “You have to base your decision on evidence presented at the meeting.”
Even when elected councils vote to deny permits, developers can — and often do — challenge those decisions in court, introducing legal uncertainty for communities already grappling with the scale of proposed projects.
Rezoning requests, by contrast, allow greater public participation but come with their own volatility. In Mooresville, the proposed $30 billion data center project was ultimately withdrawn after officials signaled opposition, underscoring how political resistance can derail even the largest investments. Together, these cases illustrate a growing mismatch between the pace of AI-driven infrastructure development and the capacity of local governments to manage its impacts—leaving residents feeling unheard and municipalities caught between economic pressure and community trust.
In addition to local sentiments on data centers coming to towns, there are energy-use considerations that have not been made sufficiently transparent to allow communities and policymakers to effectively assess whether existing power supplies can support large-scale data center expansion. As the science journal Nature reports, generative AI models require substantially more electricity than earlier forms of computing, and their rapid growth is driving an unprecedented buildout of data center infrastructure. Unlike many other energy-intensive industries, data centers tend to cluster geographically so they can share power grids, cooling systems, and high-speed data connections — intensifying their localized impact on electricity demand.
From a national perspective, AI-driven data centers still account for a relatively modest share of total US electricity consumption. However, Nature notes that the local impacts can be far more pronounced. In Virginia, for example, data centers already consume more than one-quarter of the state’s electricity. In at least five other US states, electricity consumption by data centers has surpassed 10% — illustrating how geographically concentrated development can place significant strain on regional power grids even when nationwide impacts appear manageable.
A central challenge, according to researchers cited by Nature, is the lack of reliable, transparent data on how much energy AI systems actually use.
“The real problem is that we’re operating with very little detailed data and knowledge of what’s happening,” Jonathan Koomey, an independent researcher who has studied the energy use of computing for more than three decades, told Nature.
Estimates often rely on indirect methods — such as tracking server shipments and extrapolating their power draw — because companies rarely disclose detailed information about their AI workloads or electricity consumption. This lack of transparency complicates attempts to forecast future energy demand with confidence.
Still, the direction of travel is clear. A US Department of Energy–funded report cited in Nature estimates that data centers currently consume about 4.4% of US electricity and could double or even triple that share by 2028. For states like North Carolina — where multiple data center proposals are emerging simultaneously — this uncertainty creates a planning dilemma. Without clearer data on energy requirements, grid impacts, and long-term demand trajectories, communities and regulators are being asked to make high-stakes land-use and infrastructure decisions with incomplete information.
At the state level, North Carolina has begun to grapple with AI’s broader implications. Through an executive order, Gov. Josh Stein established the North Carolina AI Leadership Council, co-chaired by the secretaries of Information Technology and Commerce, to guide the responsible adoption, governance, and advancement of artificial intelligence. The state has also launched an AI Accelerator within the Department of Information Technology and directed agencies to form internal AI oversight teams aimed at improving public services and operational efficiency.
While the AI Leadership Council’s focus is correctly on governance, workforce development, and ethical considerations, the physical backbone of AI — data centers, energy supply, water use, and siting policy — remains a critical gap. In the absence of a coordinated statewide framework, decisions continue to be made piecemeal, community by community, often within legal and procedural constraints that frustrate residents and developers alike. Addressing these realities head-on is not an argument against innovation, but a necessary step toward ensuring that North Carolina’s AI ambitions remain aligned with its communities, infrastructure capacity, and long-term public trust.
