America’s Power Grid Faces 160GW Data Center Crunch

According to DCD, US utilities have committed to connect more than 160GW of new large-load demand, primarily driven by data center expansion, according to new analysis from Wood Mackenzie. The report shows that utilities across the PJM Interconnection region alone anticipate 55GW of new demand by 2030 and 100GW by 2037, far exceeding planned generation capacity. Wood Mackenzie’s global head of Grid Edge Ben Hertz-Shargel warned that “PJM utilities have committed to serve twice as much new large-load demand as there is new power generation being developed to support it,” potentially leading to shortages and higher electricity costs in the Mid-Atlantic. The analysis revealed significant inconsistencies in how different utilities forecast future demand, with Virginia emerging as the primary focal point for data center growth despite Dominion Energy not being among the four utilities projecting more than 11GW of growth by 2040.

The Generation Gap Crisis

The fundamental issue here isn’t just about demand growth—it’s about the timing mismatch between when utilities commit to serve new loads and when they can actually build the generation to support them. Electricity generation projects, particularly large-scale ones, face regulatory hurdles, environmental reviews, and construction timelines that can stretch 5-10 years. Meanwhile, data center developers can secure interconnection agreements and begin construction within 18-24 months. This creates a dangerous lag where utilities are legally obligated to provide power before they have the means to generate it. The situation is particularly acute in PJM Interconnection, which serves 65 million customers across 13 states—any capacity shortages here would ripple through one of the world’s largest organized electricity markets.

Data Center Economics Meet Grid Reality

The data center industry’s explosive growth is colliding with the physical limitations of power infrastructure in ways that many investors and developers haven’t fully appreciated. Modern data centers supporting AI workloads can consume 50-100MW each—equivalent to a medium-sized city. When multiple facilities cluster in regions like Northern Virginia, they create localized demand spikes that existing transmission systems weren’t designed to handle. What makes this particularly challenging is that data center power demand is “firm”—they need reliable, 24/7 electricity, unlike more flexible industrial loads that can be curtailed during peak demand periods. This inflexibility means utilities can’t rely on demand response programs to manage the load.

The Tariff Solution and Its Limitations

The emergence of data center-specific tariffs represents a market-based response to this challenge, but it’s a double-edged sword. While requiring developers to pay for infrastructure upfront filters out speculative projects—as seen with AEP Ohio’s pipeline dropping from 30GW to 13GW—it also risks pushing development to regions with less robust environmental standards or weaker grid reliability. More concerning, these tariffs don’t actually solve the generation shortage; they merely shift the financial burden. The fundamental physics problem remains: you can’t create electrons through financial engineering. If utilities don’t build new generation capacity, no amount of tariff restructuring will prevent potential brownouts during peak demand periods.

The Dominion Energy Paradox

Dominion Energy’s position highlights the complexity of this situation. Despite Virginia being the epicenter of data center growth, Dominion isn’t among the most aggressive forecasters—yet CEO Robert Blue expects demand to double by 2039. This apparent contradiction suggests utilities may be deliberately conservative in public forecasts while planning more aggressively internally. Dominion’s plan to install 33GW of new generation over 20 years represents one of the most ambitious utility buildouts in modern history, but it still may not be enough given the scale of projected demand. The company faces the additional challenge of balancing this expansion against Virginia’s clean energy mandates and community opposition to new fossil fuel plants.

Broader Market Implications

This generation-demand imbalance will have consequences far beyond the utility sector. Electricity-intensive industries from manufacturing to cryptocurrency mining will face increasing competition for limited capacity, potentially driving up costs across the economy. We’re likely to see more situations like what happened in Texas, where crypto miners agreed to shut down during grid emergencies in exchange for favorable rates. The situation also creates opportunities for energy analysts and trading firms that can accurately predict where and when these shortages will emerge. More fundamentally, it raises questions about whether current market structures and regulatory frameworks are adequate for managing this scale of concentrated load growth.

The Path Forward

Solving this crisis requires a multi-pronged approach that goes beyond simply building more power plants. Utilities need to accelerate deployment of grid-enhancing technologies that can squeeze more capacity from existing infrastructure. Data center operators must embrace more aggressive energy efficiency measures and consider geographic diversification beyond traditional hubs. Perhaps most importantly, regulators need to create clearer pathways for approving and constructing new generation while ensuring costs are allocated fairly. The alternative—widespread capacity shortages and price spikes—would not only hamper digital transformation but could undermine economic growth across multiple sectors that depend on reliable, affordable electricity.