As someone who spent 14 years at the National Security Agency and the past 11 years in marketing technology, with the last three years dedicated almost exclusively to helping businesses use AI responsibly, I have seen firsthand how critical the right infrastructure is for North Carolina’s economic future. Many of the companies I work with are based right here in our state. They range from small family-owned operations to large enterprises that need practical ways to connect with customers through AI-powered tools. That is why I believe North Carolina must adopt public policies that prioritize and incentivize modular and edge data centers across the state.
Hyperscale data centers play an essential role in training frontier AI models. That work is already concentrated on the West Coast, in Texas, and in Northern Virginia. Those facilities require massive investments, enormous amounts of land, and billions of gallons of water for cooling. While they are important for a limited set of high-intensity training applications, they are not the model North Carolina should emphasize. Instead, our state has a clear opportunity to claim national leadership in the next phase of AI development: a distributed network optimized for inference models and real-world outputs.
Inference is where AI delivers value to everyday North Carolinians in their daily lives. It powers personalized recommendations that make shopping and entertainment more relevant, customer service automation that delivers instant and helpful support, predictive maintenance that keeps vehicles and home systems running smoothly, fraud detection that protects personal finances, health care applications such as remote patient monitoring and faster diagnostic support that improve access to quality care, cost savings through smarter energy management and optimized operations that reduce household and business expenses, transportation improvements like synchronized traffic signals that ease congestion and shorten commute times, and by democratizing marketing technologies for small businesses to level the playing field against enormous corporations.
These workloads do not require the same scale as frontier model training. They benefit far more from being placed closer to the point of use. A strategic network of smaller, modular edge data centers spread throughout North Carolina would meet that need without the drawbacks of hyperscale construction.
The advantages of this distributed approach are substantial. First, modular edge facilities require far less land than hyperscale projects, making them suitable for placement in a wide variety of communities rather than a handful of large industrial zones. Second, they consume significantly less water because many modern modular designs use efficient or even water-free cooling technologies. Third, they avoid the need for enormous capital projects to upgrade power substations and transmission lines, allowing faster deployment and lower infrastructure costs for both developers and taxpayers.
Additional benefits include reduced latency, which improves the performance of AI applications that rely on real-time responses. A distributed network also enhances resilience; if one site experiences an outage, others can continue operating without widespread disruption. From an economic standpoint, these facilities can be built incrementally and scaled as demand grows, creating construction and technology jobs across urban, suburban, and rural areas of the state rather than concentrating them in a handful of counties. This spread of opportunity supports North Carolina’s diverse regional economies and helps smaller businesses gain affordable access to AI tools without having to depend on distant hyperscale providers.
Environmentally, the model aligns with responsible AI principles by minimizing resource strain and allowing for more precise energy management tailored to local needs. It also supports data privacy and security goals that many of my clients care about, since processing can occur closer to the source rather than routing everything through a few centralized hubs.
North Carolina already possesses key advantages, including reliable power in many regions, a skilled workforce, and a business-friendly climate. By enacting targeted incentives, such as streamlined permitting for modular facilities and public-private partnerships for strategic site selection, state leaders can accelerate this vision. These policies would not replace hyperscale development where it makes sense for training, but they would deliberately shift our focus to the area where we can lead: practical, responsible, and widely accessible AI inference.
The future of AI will not be defined solely by who builds the largest models. It will be defined by who deploys those capabilities most effectively to create real value for businesses, communities, families, and everyday North Carolinians across our state. North Carolina has the chance to lead in that deployment phase. By prioritizing a distributed network of modular edge data centers, we can strengthen our economy; protect our natural resources; create broad-based opportunities; improve quality of life through better health care access, transportation efficiency, and meaningful cost savings for households and communities alike; and show the nation how responsible AI infrastructure should be built.