For decades, the digital backbone of the global economy was concentrated in a few well-known coastal hubs. From the sprawling campuses of Northern Virginia to the tech corridors of California, data center construction typically followed the proximity of venture capital and established tech talent. However, a significant geographical shift is underway, as the “center of gravity” for the industry is moving decisively inland.
This migration is not a matter of coincidence but a strategic response to the crushing energy demands of the artificial intelligence era. As the industry scales to meet the needs of large language models and massive compute clusters, the traditional coastal hubs are hitting a ceiling. The search for cheaper land and, more importantly, stable and affordable power is pushing the world’s largest tech operators toward the American heartland.
The trend is part of a broader global consolidation of infrastructure. According to research from Broadband Breakfast and the Synergy Research Group, hyperscale operators are projected to account for 67% of all global data center capacity by 2031.
The Great Migration: From Coasts to the Heartland
Data center construction is increasingly moving toward the center of the United States, shifting away from traditional coastal locations. This transition is being driven primarily by the availability and cost of power, which have become the most critical variables in the planning phase of new builds. While the end user may not notice a slight increase in latency as a server moves from a coast to the Midwest, the structural implications for the industry are profound.
Texas has emerged as the primary beneficiary of this shift. Reports indicate that Texas is currently leading the pipeline for new capacity, with approximately 22.5 GW coming online in the near term. Other states are also making aggressive plays to attract this investment; for example, New Mexico is seeing 8.3 GW of capacity coming online, and Pennsylvania is pursuing a major push to integrate into the data center ecosystem, with 7.1 GW in the pipeline.
Beyond the Southern states, the Midwest is seeing a surge in interest. States including Wisconsin, Indiana, Michigan, and Missouri are growing rapidly in importance. These regions have attracted multiple major projects from the industry’s heaviest hitters, including Amazon, Google, Meta, Microsoft, OpenAI, and CoreWeave. For these companies, the interior of the country offers the spatial and electrical headroom that coastal markets can no longer provide.
Powering the AI Boom: The Economics of Location
The shift toward inland investment is fundamentally an energy play. Synergy Research Group has highlighted that the availability and cost of power are the major driving factors behind this inland shift. The power-hungry nature of AI workloads requires an electrical infrastructure that can handle massive loads without destabilizing the local grid or incurring prohibitive costs.
John Dinsdale, chief analyst at Synergy, notes that while this shift is critical for the shape and structure of the underlying infrastructure, it is less impactful for the average end user. However, for the operators, the move is essential. The cost of power is a recurring operational expense that can develop or break the margins of a hyperscale facility, making the energy-rich environments of the Midwest and Texas highly attractive.
The Infrastructure Gap: Supply Chain Hurdles and Delays
Despite the aggressive planning and the shift in geography, the actual pace of data center construction is being tempered by significant headwinds. Market intelligence from Sightline Climate suggests a stark disconnect between what is planned and what is actually being built. For the current year, while 16 gigawatts (GW) of data centers are set to open in the U.S., only about a third of that—roughly 5 GW—is currently under construction. Sightline expects between 30% and 50% of these projects to face delays.
The outlook for the immediate future remains strained. For 2027, approximately 25 GW of data centers have been announced for deployment, yet only 6 GW of that infrastructure is actually under construction. The gap widens even further when looking toward 2028 through 2032, where the vast majority of planned launches have yet to break ground.
These delays are attributed to two primary factors:
- Component Shortages: The industry is facing a critical lack of essential hardware. This is not limited to high-profile shortages in memory and storage, but extends to foundational electrical components. Batteries, electrical transformers, and circuit breakers are in short supply. While these components may account for less than 10% of the total construction cost, they are indispensable. Andrew Likens, energy and infrastructure lead at AI data center provider Crusoe, has noted that if a single piece of the supply chain is delayed, the entire project cannot deliver.
- Localized Resistance: There is a growing “rebellion” against the proliferation of data centers from both citizens and local governments. This pushback often centers on land use, water consumption, and energy strain. A notable example of this resistance is the Seminole Nation of Native Americans, who have banned the construction of data centers on their tribal lands.
What This Means for the Global Tech Landscape
The migration of data centers to the U.S. Interior represents a maturing of the industry. The “land grab” phase of the AI boom is colliding with the physical realities of the electrical grid and the global supply chain. The fact that hyperscale operators are expected to control 67% of global capacity by 2031 indicates a concentration of power—both electrical and computational—in the hands of a few dominant players.
For the Midwest and the South, this represents a massive influx of capital and a shift in economic identity. For the tech industry, it is a race to secure the physical resources necessary to sustain the AI revolution. The transition from the coasts to the heartland is not just a change in address; it is a fundamental realignment of how the digital world is powered and housed.
Key Takeaways on Data Center Trends
- Geographic Shift: Construction is moving from coastal hubs to inland states, with Texas, New Mexico, and Pennsylvania leading the current surge.
- Primary Driver: The availability and cost of electrical power are the main factors influencing new site selections.
- Execution Gap: A significant portion of announced projects are delayed due to supply chain bottlenecks in electrical components like transformers and circuit breakers.
- Social Pushback: Local and tribal governments, such as the Seminole Nation, are increasingly restricting data center development on their lands.
- Hyperscale Dominance: A slight number of massive operators are on track to control two-thirds of global capacity by 2031.
As the industry continues to navigate these supply chain “puzzles” and regulatory hurdles, the next few quarters will be critical in determining whether the announced gigawatts of capacity can actually be realized. Industry analysts will be watching for updated construction starts in the Midwest and Texas to see if the 2027 targets remain viable.
Do you think the shift to inland data centers will revitalize the American Midwest, or are the infrastructure delays too severe to overcome? Share your thoughts in the comments below.