TO: Mayor and Members of City Council
THRU: Jodi Phelps - Assistant City Manager
Dr. Gerald Newton, AICP - Development Services Director
FROM: Willie L. Henry, CGCIO, Chief Information Officer
Demetrios Moutos, AICP - Planner II
Craig Harmon, CZO - Senior Planner
DATE: April 6, 2026
RE:Title
Telecommunications and Data Storage Facilities (Data Centers)Title
end
end
COUNCIL DISTRICT(S):
Council District(s)
All
b
Relationship To Strategic Plan:
The history and evolution of data centers are closely connected to the City of Fayetteville’s FY2025 Strategic Plan, presenting both opportunities and challenges. Modern data centers advance the Plan’s focus on economic diversity, strategic growth management, innovation, and data-driven governance. As these facilities have transitioned from small, specialized computer rooms to major infrastructure assets, they have become integral to the modern economy. This progression aligns most directly with Goal II: Diverse and Viable Economy and Goal III: City Investment Planning, as data centers can broaden the nonresidential tax base, attract substantial private investment, and necessitate coordinated planning for land use, utilities, broadband, and long-term infrastructure capacity.
Data centers are most consistent with the Strategic Plan when recognized as major infrastructure uses that require comprehensive land-use review, fiscal discipline, utility coordination, and transparent public engagement. The historical development of data centers thus supports the Plan’s objectives but underscores the importance of balancing economic opportunities with long-term public service, infrastructure, and community impacts.
Executive Summary:
Data centers have transitioned from small, internal computer rooms serving government, research, and corporate institutions to large, standalone facilities supporting cloud computing, digital communications, and artificial intelligence. As this transformation has occurred, data centers have become major physical infrastructure, requiring substantial power, cooling, fiber connectivity, security, and long-term site planning. North Carolina now reflects this broader national trend, with recent investments demonstrating that data centers are no longer a niche land use. For local governments, this historical context is significant because it necessitates viewing data centers not as generic office or warehouse space, but as a distinct development type whose siting, infrastructure requirements, and community impacts can influence land-use decisions for decades.
TA26-01 addresses this shift by closing a gap in the Unified Development Ordinance (UDO). The current UDO does not define data centers as a distinct land use, requiring staff to classify proposals on a case-by-case basis and resulting in inconsistent entitlements, review standards, and long-term enforcement. The proposed amendments introduce a clearer and more predictable framework by distinguishing incidental server rooms from principal-use data centers, aligning approvals with operational intensity, and focusing regulation on external effects that most significantly impact surrounding properties and City systems, such as exterior equipment, generator activity, noise, utility coordination, screening, and airport compatibility. The proposal was developed through cross-departmental review and informed by peer-community experiences, which indicate that generic industrial classifications often fail to address the impacts most associated with modern data centers.
The evolution of data centers underscores the need for a balanced and pragmatic approach by the City. These facilities can generate substantial private investment, expand the nonresidential tax base, and enhance economic competitiveness. However, as infrastructure-intensive and low-employment developments, their fiscal and community value extends beyond capital investment alone. The benefits of data centers depend on effective revenue capture, strategic planning for infrastructure and utility demand, and long-term compatibility with surrounding areas. Collectively, the industry’s history, the proposed text amendments, and broader local government experience indicate that data centers require a more precise and enforceable regulatory framework than traditional zoning categories can provide.
Background:
In late summer 2025, the Development Services Department began receiving inquiries about data centers, mainly concerning zoning districts and regulatory standards. In response, staff initiated research to evaluate how this use could be incorporated into the City’s existing Unified Development Ordinance (UDO) framework.
Through this analysis, staff found that the current Use table did not adequately address the classification, location, or regulation of data centers. As a result, staff began developing potential regulatory approaches, drawing from ordinances in other North Carolina jurisdictions and nationwide best practices.
Staff advised interested third parties of two paths forward: (1) allow time to develop and adopt new ordinance standards through the legislative process, or (2) submit a request for a formal interpretation to determine how the use could be classified under the existing ordinance. A formal interpretation would follow the City’s established process, including review by the City Manager and the opportunity for appeal to the Zoning Commission.
The applicant group chose to defer pursuing a formal interpretation and instead await the development of specific ordinance standards.
In October 2025, a preliminary draft proposal was presented to the Planning Commission for initial review. In January 2026, a working group of technical experts reviewed the proposed amendments. Based on this collaboration, a refined set of more comprehensive and enforceable standards was developed and presented to the Planning Commission in February 2026. These standards introduced a tiered system for defining and regulating data centers, as summarized below.
The framework aligns regulatory measures with real-world impacts by identifying four primary use types: (1) internal-only accessory facilities, defined as minor server rooms located within another principal use; (2) accessory facilities with dedicated exterior plant, where support equipment generates external land-use effects; (3) principal data centers, in which data storage and processing constitute the primary site use; and (4) principal data centers with an Energy Center, where on-site power generation results in a more intensive operational profile. This categorization enables the UDO to distinguish between minor server rooms and hyperscale campuses, while establishing clearer standards as facility scale, equipment, and operations begin to impact adjacent properties and municipal systems.
Issues/Analysis:
The Evolution of Data Centers
Data centers did not emerge solely because of artificial intelligence. They represent the latest phase in a longstanding transformation of how the United States stores, processes, and transmits information. A data center is a dedicated facility housing computing equipment along with the power, cooling, networking, and security systems needed for continuous operation. Initially, these centers were isolated rooms for large government and research computers. Over time, they developed into a nationwide network of commercial facilities supporting banking, logistics, communications, cloud services, streaming, and artificial intelligence. Today, local governments and AI policy organizations see this infrastructure as fundamental to the modern economy. As a result, data centers have become increasingly important in land-use and infrastructure planning.
Early Computing: Dedicated Machine Rooms
The earliest precursor to the modern data center was not a large warehouse filled with servers but a dedicated computer room. In the 1940s and 1950s, computers like ENIAC were large, power-intensive machines designed for military and scientific calculations. Introduced in 1946, ENIAC occupied more than 1,500 square feet, weighed about 30 tons, and used 18,000 vacuum tubes. Machines from era required specialized rooms, controlled environments, continuous maintenance, and restricted access. Thus, even before the term “data center” became common, the foundation was set: critical computing equipment needed a purpose-build physical environment.
Mainframes, Time-Sharing, and the Development of the Computer Center
Between the 1950s and the 1970s, computing expanded beyond military and scientific uses into government agencies, universities, banks, insurers, and large corporations. Mainframes formed the foundation of administrative and business computing. IBM’s System/360, introduced in 1964, was significant because it unified a family of machines under one architecture, allowing organizations to expand without discarding existing software. At the same time, time-sharing allowed multiple users to access a central machine via remote terminals. This model spread throughout the 1960s and 1970s in sectors like banking, insurance, and retail. During this era, the “computer room” evolved into the “computer center,” a centralized, climate-controlled facility where institutions concentrated their most critical computing operations.
This period also transformed Americans’ perceptions of data. In the mid-1960s, scholars and federal officials proposed establishing a U.S. “National Data Center” to centralize access to machine-readable records for research and planning. Although this proposal was never implemented, it revealed two enduring concepts that continue to influence data-center politics. First, institutions increasingly sough centralized data infrastructure to facilitate analysis, forecasting, and administration. Second, the public rapidly recognized that centralized data systems could concentrate power and pose risks to privacy. Historians Dan Bouk and Christopher Loughnane, along with William Aspray, demonstrate that the origins of modern data-center politics extend beyond hardware considerations. From the outset, the centralization of computing and records prompted questions regarding not only efficiency, but also governance, control, and public trust.
By the late 1960s and early 1970s, large computing centers had become prominent enough to attract political attention. Max Larson’s research on campus computer-center sabotage indicates that university computing facilities were occasionally targeted during antiwar and racial justice movements, as students saw them as symbols of institutional power, military research, and administrative authority. While these facilities were not equivalent to contemporary commercial data centers, their significance shows that large computing sites had already acquired social and political meaning well before cloud computing.
Evolution from In-House Computer Rooms to Commercial Data Centers
During the 1980s and 1990s, personal computers and client-server systems decentralized computing within organizations. Offices no longer relied solely on a single mainframe. However, this change did not eliminate the need for centralized infrastructure; it transformed its nature. As organizations connected local machines to broader networks, secure backend facilities remained essential for hosting applications, storing records, and managing telecommunications. This shift led to the rise of the commercial data center industry, providing facilities for shared hosting, interconnection, and outsourced services beyond a single company’s internal use. Studies show many third-party facilities clustered in urban areas to stay close to customers, networks, and exchange points.
Two key developments accelerated this transition. First, virtualization allowed multiple computing environments to run on the same physical hardware, greatly improving server utilization. Second, there was growing emphasis on reliability standards. By the late twentieth century, data centers were judged not only on functionality but also on their ability to stay operational during maintenance, equipment failures, or disruptions. The Uptime Institute’s tier framework formalized these expectations. As a result, data centers evolved into engineered systems designed for redundancy, high uptime, and continuous service rather than just rooms housing computers.
Cloud Computing and the Emergence of Hyperscale Infrastructure
A major transformation occurred in the 2000s, when cloud computing allowed data-center capacity to be rented on demand instead of owned by individual customers. According to Amazon, AWS launched in 2006 to reimagine IT infrastructure, first offering S3 for storage and later EC2 for computing. This change fundamentally altered the computing business model. Organizations no longer needed to build and maintain their own server rooms. They could acquire storage and processing power remotely as needed. As a result, the industry shifted toward much larger campuses run by major cloud providers.
This transition also affected the geographic distribution and ownership of facilities. Shane Greenstein and Tommy Pan Fang note that third-party data firms and cloud providers use different locations strategies: third-party firms favor urban areas, while cloud providers focus on fewer, lower-density locations to achieve greater scale. Daniel Greene adds that many core physical internet assets are now controlled not only by technology firms but also by commercial real estate owners. Their colocation buildings and interconnection facilities have become part of a broader property market. In short, data centers evolved from back-office equipment to a specialized form of infrastructure, utility demand, and real estate investment.
The Local Turn: Land, Tax Policy, and Physical Infrastructure
As data centers have grown, their local impact has become more significant. Jenna Burrell’s study of Facebook’s Prineville, Oregon, facility shows that establishing a large data center rarely results from a single corporate decision. Instead, it depends on years or decades of groundwork involving land acquisition, electricity, fiber infrastructure, tax policy, and local economic development strategies. Burrell argues rural leaders often prepared for such investments long before a specific company arrived, even if these efforts seemed incomplete at the time. This is important for local governments: data centers do not simply appear without context. They are built where physical infrastructure, policy decisions, and market conditions align.
This local aspect helps explain mixed responses to data centers. Supporters often highlight increased tax revenue, construction spending, and enhanced regional prestige. Critics point to limited permanent jobs, high utility demands, land use, and public costs tied to incentives. Materials from the National League of Cities now describe data centers as a local governance issue because their benefits and burdens are most visible at municipal and county levels, even when serving national or global markets. Thus, the history of data centers shows how local governments have realized “the cloud” is not intangible. It is physical, territorial, and heavily dependent on public infrastructure.
The AI Era and the Reemergence of Physical Constraints
The current expansion of artificial intelligence has increased awareness of physical infrastructure needs. While modern data centers still store and process conventional digital traffic, AI training and inference require much larger facilities with substantial power and cooling demands. According to the U.S. Department of Energy, data centers used about 4.4% of total U.S. electricity in 2023 and are projected to consume between 6.7% and 12% by 2028. This explains why discussions now focus more on substations, transmission, backup power, water use, and utility cost allocation. Recent developments in data center operations confirm a longstanding fact: although computing seems virtual to users, it depends on physical buildings, machinery, and public infrastructure.
North Carolina’s Role in the National Data Center Landscape
North Carolina is now firmly part of the latest phase of data center development. The state emerged early as a destination for major cloud infrastructure, with Google announcing its Lenoir data center in 2007 and Facebook opening its Forest City facility in 2012. North Carolina also offers targeted sales and use tax exemptions for qualifying data centers, showing a deliberate strategy to attract this industry. More recently, the state announced Amazon’s planned $10 billion artificial intelligence and cloud campus in Richmond County, illustrating the market’s shift from initial cloud facilities to much larger AI-focused projects. The accompanying materials document this path, highlighting the move from early western North Carolina campuses to broader statewide expansion and growing infrastructure demand.
Collectively, this historical progression reveals a distinct pattern. Data centers have evolved from specialized machine rooms serving a limited number of large institutions to centralized computer centers, to outsourced commercial facilities, and ultimately to hyperscale campuses that support cloud and artificial intelligence services. Each phase has increased the power and accessibility of computing, while also amplifying the significance of the supporting infrastructure. For local governments, this evolution underscores the need to view data centers not as a narrow technological application, but as a critical form of physical infrastructure. Their siting, utility requirements, environmental impacts, and fiscal arrangements can influence communities for decades.
Regulatory Response
As data centers evolved from internal computer rooms into large standalone campuses, land-use questions changed. Older zoning codes often absorbed small server rooms under broader office, institutional, or industrial categories because the equipment was mostly internal and off-site effects were limited. Modern data centers are different. Often, the most consequential parts are not the server halls but the exterior plant and support systems that enable continuous operation. These include cooling yards, generator yards, substations, switchgear, fuel systems, and other utility infrastructure. Routine generator testing, noise, screening, utility demand, airport compatibility, and campus-scale phasing are now common planning issues. This is why local regulations must evolve with the industry.
That shift also explains why treating data centers as generic industrial or warehouse uses is often insufficient. A conventional industrial classification may indicate where a use is allowed, but it often does not address the questions residents and staff face once a project is proposed. How close can major equipment come to homes? When can generators be tested? What information should be provided about utility demand and site operations? How should large exterior equipment areas be screened? What gets reviewed up front, and what gets verified after construction? Peer communities, including high-growth data-center markets, have learned that if these questions remain unresolved, review becomes inconsistent, applicants face uncertainty, and enforcement becomes harder over time.
TA26-01 is intended to respond to that broader change in a measured way. It starts from a basic point: not every room with servers is a data center in the land-use sense. Hospitals, offices, studios, and similar uses may contain individual server rooms, but they do not function like principal-use data centers with dedicated exterior plant and campus-scale infrastructure. The amendments, therefore, tries to draw a clearer line between minor, internal IT functions and full-scale data-center development, so the code regulates the uses that actually create distinct land-use impacts without unintentionally sweeping in ordinary buildings with back-of-house technology.
The amendments also reflect how these facilities operate. Data centers are typically infrastructure-heavy and people-light. They often require far more power, mechanical equipment, and service coordination than their employment numbers suggest. Some include on-site power-generation components that change their operating profile by adding fuel logistics, more intensive equipment areas, and greater nuisance potential. That is why the draft distinguishes between a baseline principal facility and a more intensive variant with an Energy Center. It focuses on measurable external effects rather than treating all data centers as if they operate the same way.
Modern data centers are zoning issues, but also utility coordination, operational, airport compatibility, and long-term enforcement issues. The proposed text amendments aim to give the City a more predictable framework before proposals become more frequent or complex.
In that sense, TA26-01 is best understood not as a departure from the history of data centers but as the next step. As these facilities have become larger, more specialized, and more consequential to surrounding areas, local codes can no longer rely on loose analogies to older industrial forms. The ordinance aims to bring the UDO into closer alignment with what data centers have become: a distinct land use with recognizable public benefits and identifiable off-site impacts that require clearer rules, earlier coordination, and more reliable long-term administration.
Budget Impact:
The budget impact of data centers has changed as the industry evolved. Earlier computing facilities were often small, internal rooms within government, university, or corporate buildings, so their local fiscal footprint was modest. Modern data centers are different. The industry shifted toward large standalone campuses built around cloud computing and AI. Local governments now see them not as ordinary office support space but as major physical developments tied to land, power, water, fiber, and long-term infrastructure planning. The National League of Cities describes data centers as a local-governance issue because they bring tax revenue, jobs, and improved digital connectivity while also creating pressure on energy demand, water use, and land use. DOE reports that U.S. data-center electricity use rose from 58 TWh in 2014 to 176 TWh in 2023, showing how much larger and more infrastructure-intensive these facilities have become.
Historically, the most important positive budget effect for local governments has been revenue, especially where data centers are taxable as land, buildings, and equipment. In Virginia’s mature data-center localities, JLARC found data-center revenue ranged from less than 1 percent to 31 percent of total local revenue, with Loudoun County at 31 percent and Prince William County at 7 percent in FY23. JLARC also found a single data center with $150 million in taxable computer equipment could generate very different local revenue over five years depending on local tax policy, from about $10.8 million to $0.4 million. Data-center equipment is expensive and often replaced on roughly five-year cycles, so the local tax stream can be substantial where the tax structure captures it. Local officials in those Virginia jurisdictions reported data-center revenue helped lower real-estate tax rates, build reserve funds, support an affordable-housing trust fund, and construct new schools.
That said, the historical record shows the budget upside is not simple. Data centers are often promoted as job creators, but economic benefits mainly come from construction rather than ongoing operations. JLARC found most of the industry’s economic benefits in Virginia came from construction, not day-to-day operations. Industry representatives reported construction peaks can bring about 1,500 workers on site, which differs greatly from the lower staffing levels once operations stabilize. This pattern matters for local budgets because communities may see large upfront investment and construction activity without equally large long-term operating employment. In other words, data centers can be strong tax-base projects even if they are not major permanent-job generators.
The historical cost side has become more visible as facilities have grown. Modern data centers require local governments to spend more time on planning, zoning, permitting, and coordinating with utilities and state regulators. They can also create indirect fiscal pressure when growth in electricity, water, and land demand affects public systems or community expectations. The National League of Cities advises local leaders to address sustainability, zoning, and infrastructure concerns proactively. It also notes that city leaders increasingly need to coordinate with utilities, regulators, and technology companies when these facilities are proposed. This pattern reflects the broader evolution of the industry. Once data centers became large, utility-dependent campuses rather than back-office rooms, their budget impact expanded from simple tax questions to broader issues about infrastructure readiness, public-service capacity, and long-range growth management.
Overall, the historical local-government budget impact of data centers is mixed but potentially significant. They can strengthen the tax base, generate substantial property- and equipment-related revenue, and provide bursts of construction-related economic activity. At the same time, their benefits vary widely based on local tax structure. Their permanent employment footprint is usually smaller than their capital investment suggests. Their growing infrastructure demands can create indirect public costs that local governments must manage. As a result, the budget story of data centers is not simply that they are fiscally positive or negative. Over time, they have become large and infrastructure-intensive enough that their local fiscal effects depend heavily on how well a community captures revenue, plans for utility and land-use impacts, and aligns approvals with long-term public-service capacity.
Options:
Next steps determined at Council’s direction.
Recommended Action::Recommended Action
end
Attachments:
• Data Center Evolution - PowerPoint
• Telecommunications and Data Storage Facilities - Data Centers - PowerPoint
• Planning Commission February 17, 2026, Agenda Packet
• Planning Commission February 17, 2026, Draft Minutes
• Data Centers Presentation from FCEDC to the County Commissioners on February 19