Economic development has always evolved alongside the defining infrastructure of its era. Canals connected markets. Railroads industrialized regions. Rural electrification transformed manufacturing productivity. Interstate highways rewrote the logic of logistics. Telecom and broadband reshaped the service economy. Each generation believed its buildout was singular. In reality, the pattern has repeated itself for more than a century. 

Today the AI and data center boom is the latest turn of that wheel. For economic developers, the lesson from more than a hundred years of cycles is not that the infrastructure was a mistake. It is that the practitioners who understood the full arc of the development of that infrastructure made better decisions than those who only saw the opening surge. 

Infrastructure booms create lasting value. They rarely create it smoothly. 

The arc is consistent. Capital floods toward strategic infrastructure. Regions compete hard to capture it. Excess capacity or geographic imbalance emerges. Markets restructure, often painfully. And then the infrastructure that looked overbuilt becomes the platform for the next era of growth. 

Railroad speculation produced consolidation and bankruptcies, then accelerated industrialization far beyond passenger transport. The fiber overbuild of the dot-com years was financially brutal for the companies that financed it. But, that same dark fiber derided as waste later became the foundation for broadband, cloud computing, and video streaming. That backbone now helps enable the AI boom itself. As with many cycles, the infrastructure outlived the bubble that built it. 

The AI economy is producing a new infrastructure race. 

This is not primarily a technology story. It is an infrastructure story.  

Data centers consumed about 4.4 percent of U.S. electricity in 2023 and are projected to reach roughly 6.7 to 12 percent by 2028, according to a 2024 Department of Energy report. Meeting that demand is driving one of the largest expansions of electric infrastructure in decades.  A July 2025 Department of Energy analysis projected that total U.S. electricity demand will require roughly 100 GW of new peak generating capacity by 2030, with data center growth expected to account for about half of it. That represents one of the largest expansions of electric infrastructure in decades. 

The implications extend far beyond the technology sector. Expanding generation is only part of the challenge. New transmission lines, substations, and other grid infrastructure must also be planned, permitted, financed, and constructed, often over many years. The result is that electric infrastructure itself is becoming a competitive asset.  

That shift changes the calculus for economic developers. Utility capacity, transmission access, industrial sites, and water availability increasingly determine which projects a region can credibly pursue.  

Water deserves equal attention. Large data centers, power plants, combined-cycle gas turbines, and many industrial processes all place significant demands on water resources. In water-constrained regions, access to reliable water supplies is becoming as consequential as access to power.  

Energy and water are no longer cost lines to manage. They have become structural constraints on what can be built and where. 

In the near term, this creates real competition. The communities recruiting data centers are often pursuing advanced manufacturing at the same time, and in some markets the two now draw from the same pools of megawatts, industrial land, and skilled trades. That pressure is important to highlight, but history points to the more important and optimistic story. 

The generation, transmission, substations, and fiber going in to serve AI are the same assets that determine industrial competitiveness. History suggests this is the pattern nearly every major infrastructure cycle has followed.  

Railroads built for one purpose industrialized regions far beyond it. Rural electrification, built to light homes, transformed factory productivity. Fiber laid for the dot-com economy became the backbone of cloud computing. In many regions today, hyperscale demand is accelerating grid investment that would otherwise have been years away.  

There is a real difference between consuming infrastructure capacity and catalyzing its expansion. The distinction matters because today's largest users may also become tomorrow's justification for investments that benefit an entire regional economy. The buildout justified now by the spike in digital demand is likely to improve the viability of energy-intensive industry tomorrow: battery production, semiconductors, electrified industrial processes, robotics-heavy manufacturing. 

The risk is not the infrastructure. It is strategic imbalance. 

History suggests the danger has never been the investment itself. The danger is failing to anticipate second-order effects: overconcentration in a single infrastructure class, workforce strain, rising power and water costs, the potential for stranded assets, and slower timelines for industrial users crowded out by larger loads. 

Not every community in this boom will emerge equally positioned once the market matures. Some will secure transformational, durable assets. Others will inherit constrained systems, higher costs, and reduced industrial flexibility. The difference will not come down to who moved first. It will come down to who understood the long-term tradeoffs behind today’s decisions. 

What economic developers should be asking now. 

The defining question is shifting. "How do we win this project?" is giving way to harder, more strategic ones: 

• What is our long-term infrastructure strategy, beyond the next announcement? 

• How much power concentration in a single use is healthy for our region? 

• Which industries create complementary ecosystems rather than competing for the same scarce inputs? 

• How do we protect optionality so today's win does not foreclose tomorrow's? 

• Are our workforce systems keeping pace with both construction and operations demand? 

• How are our utilities, EDOs, manufacturers, and communities actually coordinating, rather than working in parallel? 

 Economic development has always been strongest when practitioners balanced growth attraction with long-term resilience. Success of the next century will depend on managing infrastructure transitions without losing economic balance. 

The AI buildout will reshape the economic geography of North America, just as railroads and electrification did before it. The transmission, generation, and connectivity going in today may strengthen industries well beyond information technology. But realizing that value will require coordination. As economic development enters its second century, communities that treat infrastructure expansion as part of a broader industrial strategy, rather than a single-project recruitment exercise, will be better positioned when the cycle matures. 

History suggests the communities that benefit most are rarely the ones that move fastest. They are the ones that think furthest ahead. 

Didi Caldwell is the President and CEO of Global Location Strategies Pillar: Readiness Wins.