How AI Data Centers Are Driving Commercial Electricity Prices Higher (And What Businesses Can Do About It)

There's a force reshaping commercial electricity prices across the United States, and most business owners have no idea it's coming for their energy budget. It's not a pipeline failure or a geopolitical crisis. It's the servers training large language models, processing genomic data, and rendering AI-generated content — and the power they consume is restructuring wholesale electricity markets in ways that haven't been seen since the shale gas revolution of the 2010s.

US data center electricity demand is projected to grow from approximately 80 GW today to 150 GW by 2030, according to analysis from the Lawrence Berkeley National Laboratory. That 70 GW of new load — roughly equivalent to adding the entire state of California's current peak demand to the grid in five years — is not being distributed evenly. It's concentrating in specific regions: Northern Virginia, the Dallas-Fort Worth metroplex, Columbus, Ohio, and the Chicago suburbs.

For businesses operating in those regions, the electricity grid is being fundamentally repriced. For businesses operating anywhere in the PJM, ERCOT, or MISO markets — which together cover most of the eastern United States and Texas — the ripple effects of data center load growth are now showing up as higher capacity auction prices, tighter reserve margins, and elevated forward electricity curves.

This isn't a distant threat to monitor with vague concern. The 2026/2027 PJM capacity auction cleared 22% above the prior year, driven in significant part by data center load growth in Virginia and Ohio. Texas's ERCOT grid is warning of negative reserve margins for summer 2026 as data center interconnection requests pile up faster than new generation can be built. Businesses that understand the mechanism behind these price increases — and that take proactive procurement steps — will maintain competitive energy costs. Those that don't will absorb them passively.

The 80-to-150 GW Demand Curve: Why Data Center Load Is Reshaping the Grid

To understand why data center growth affects your electricity bill, you need to understand how electricity markets price capacity and energy — and why a massive new load anywhere in a regional grid affects prices everywhere in that grid.

How Regional Grid Markets Work

US wholesale electricity markets — PJM, ERCOT, NYISO, ISO-NE, MISO, CAISO — are regional systems. All generators bid into a common pool; all load pays a common (or zone-modified) market clearing price. When demand grows faster than supply — new load, retirements, or both — two things happen: the energy price rises, and the capacity market clears higher to incentivize new generation.

Data center growth is causing exactly this dynamic, and it's happening faster than the grid can respond.

The Interconnection Bottleneck

New generation (solar, wind, batteries, natural gas) connects to the grid through an interconnection queue managed by each ISO/RTO. These queues have become catastrophically backlogged. PJM's interconnection queue contained over 290 GW of proposed projects as of early 2026 — most of it waiting years for interconnection studies and grid upgrades to accommodate new capacity.

Meanwhile, new data center load can connect to the grid far faster than new generation can. The result is a widening gap between demand growth and supply response — precisely the condition that drives capacity prices higher.

Why 150 GW of Data Center Load Is Different From Normal Growth

Historically, US electricity demand growth has been modest — roughly flat since 2007 as efficiency improvements offset economic expansion. Data center growth breaks that pattern. A single hyperscale campus (Google, Microsoft, Amazon, Meta) can consume 500 MW to 1,000 MW — comparable to a mid-sized city. These campuses run at 90-95% utilization 24/7/365, creating an extremely high load factor that strains grid infrastructure differently than residential or small commercial load.

The AI-powered energy management systems that smaller businesses are adopting to reduce consumption are being overwhelmed, in aggregate pricing terms, by the systems those same AI companies need to train and run.

Regional Hot Spots: Virginia, Texas, Ohio, and Illinois Price Pressure

Data center load growth is not uniform. Four regions stand out as the most significant price pressure points — and if your business operates in any of them, the impact on your electricity costs is already measurable.

Northern Virginia: Ground Zero

Loudoun County, Virginia — dubbed "Data Center Alley" — hosts more data center capacity than any other location on earth. AWS, Microsoft, Google, Meta, and dozens of colocation operators have built, are building, or have announced more than 70 GW of planned data center capacity in the greater Northern Virginia region.

The electricity infrastructure consequences are severe. Dominion Energy Virginia is executing a multi-billion-dollar transmission upgrade program. PJM has imposed interconnection moratoria in Northern Virginia load zones during peak construction periods. And the locational marginal price (LMP) differential between Northern Virginia and the PJM hub has widened to 20-35% on peak days as congestion increases.

What this means for Virginia businesses: Higher transmission costs being socialized across all customers, elevated capacity zone clearing prices in SWMAAC (the PJM load zone covering Northern Virginia/DC/Maryland), and tighter grid conditions that push up energy prices during peak hours.

Texas: ERCOT Under Pressure

Texas has attracted a second wave of data center investment, drawn by abundant renewable energy, favorable land costs, and tax incentives. The ERCOT interconnection queue for new data center load exceeded 40 GW in early 2026 — load that the grid was never designed to serve with current infrastructure.

ERCOT does not have a traditional capacity market; it relies on scarcity pricing (the Operating Reserve Demand Curve) to incentivize adequate capacity. As load grows, scarcity events become more frequent and more expensive — a dynamic that will define summer 2026 and beyond for Texas commercial customers on variable or indexed contracts. The Texas commercial electricity market faces its most challenging summer conditions in years.

Ohio and Illinois: Secondary Hot Spots

The Columbus, Ohio metro area and suburban Chicago have emerged as secondary data center hubs within PJM. Ohio's proximity to the mid-Atlantic fiber corridors and Illinois's access to Chicago financial infrastructure make both attractive to major operators. Both regions are seeing above-average load growth, contributing to the elevated PJM capacity auction results discussed earlier.

For commercial customers in Columbus, Cleveland, Chicago, and their suburbs, the capacity cost increases from the 2026/2027 PJM auction are being felt with above-average intensity.

How Non-Data-Center Businesses Get Stuck With the Bill

Here's the mechanism that most commercial energy discussions don't explain clearly enough: data center load growth increases costs for all grid users, not just the data centers themselves.

Capacity Cost Socialization

When PJM's capacity auction clears higher because of increased demand from data centers, every business connected to the PJM grid pays those higher capacity charges — not just the facilities responsible for the load growth. A regional grocery chain in Pennsylvania, a manufacturer in Maryland, a hotel group in New Jersey: all face higher capacity charges because the aggregate PJM load increased, pushing the capacity auction clearing price higher.

Transmission Infrastructure Costs

New transmission infrastructure built to serve data center load is typically recovered through transmission charges socialized across all grid users. Dominion Energy Virginia's multi-billion-dollar transmission expansion to serve Northern Virginia data centers will be paid by all Dominion customers through transmission access charges — including every small business in the Dominion service territory.

Locational Price Premiums

Businesses located in or near high-data-center-density areas face an additional burden: locational marginal price (LMP) premiums from grid congestion. When the transmission system between the data center cluster and the rest of the grid becomes congested, the locational price at that congested bus rises above the regional average. Commercial customers in those zones pay the congestion premium on every kilowatt-hour.

Understanding advanced contract structures that account for locational basis risk is increasingly important for businesses in data center hot spot regions.

Procurement Playbook for Operating Through the AI Power Crunch

The good news: while individual businesses can't reverse the data center demand wave, they have meaningful procurement and operational tools to manage its cost impact.

Action 1: Audit Your Contract for Capacity Pass-Through Exposure

Pull your current supply contract and identify how capacity charges are treated. If you're on a pass-through capacity structure, you're directly exposed to every future capacity auction result. Quantify that exposure now, before the June 2026 billing cycle.

Action 2: Lock Longer-Term Fixed Contracts While Forward Markets Remain Below Worst-Case

Forward electricity curves for 2027-2028 reflect market expectations of continued capacity tightening, but they don't fully price in tail-risk scenarios (another 20%+ capacity auction increase, extreme summer scarcity events). Locking fixed-rate contracts that include capacity now — before another adverse auction result — can provide meaningful budget protection.

The best time to lock in electricity rates for businesses in PJM is typically late spring or early fall, when shoulder-season energy prices are lower and before summer capacity conditions become clearer. The AI demand surge creates a strong argument for locking sooner rather than later in the current cycle.

Action 3: Enroll in Demand Response Programs

Demand response enrollment reduces your ICAP tag (lowering capacity costs for subsequent years), earns you revenue, and helps the grid manage the load growth challenge that data centers are creating. For PJM commercial customers, demand response revenue of $50,000 to $500,000 per year is achievable for large facilities — partially offsetting the higher capacity charges you're absorbing.

Action 4: Evaluate Load Flexibility and Operational Scheduling

Businesses that can shift discretionary load — manufacturing batches, HVAC pre-cooling, EV charging, industrial processes — away from the peak demand hours that drive both capacity tags and scarcity energy pricing will benefit disproportionately as the AI power crunch intensifies. Invest in building automation and smart controls that enable this flexibility.

Action 5: Monitor Regional Interconnection Developments

Data center construction in your region is publicly documented through utility interconnection filings and zoning applications. Businesses that track these developments can anticipate localized transmission congestion premiums and factor them into procurement strategy earlier than competitors.

Working with an energy advisor who monitors commercial energy procurement case studies and regional market dynamics provides a significant intelligence advantage in this environment.

Conclusion

The AI data center buildout is the most significant structural change to US electricity demand since the internet boom of the 1990s — and its effects on commercial electricity prices are already visible in capacity auction results, forward market curves, and regional congestion pricing. This is not a cycle that will reverse when interest rates change or when one tech company cuts capex.

For commercial businesses, the imperative is to treat energy procurement strategically — not as a commodity transaction to be handled when the contract comes up for renewal. The businesses that will manage energy costs most effectively through the AI power crunch are those that understand their grid region's supply/demand dynamics, protect themselves with well-structured supply contracts, participate in demand response programs, and invest in load flexibility.

Commercial Energy Advisors monitors PJM, ERCOT, MISO, and other regional markets continuously, giving our clients the market intelligence and procurement expertise to navigate exactly this kind of structural price shift.

Call 833-264-7776 or contact us today for a complimentary analysis of your facility's capacity cost exposure and a procurement strategy review calibrated to your region's data center load growth trajectory.

Frequently Asked Questions

Why do data centers increase electricity costs for businesses that aren't data centers?

Electricity markets price capacity and energy on a grid-wide basis. When data center load increases overall demand in a regional grid, capacity auction clearing prices rise for all customers, transmission infrastructure costs are socialized across all ratepayers, and locational congestion premiums emerge near data center clusters — affecting every commercial user of that grid.

Which US regions face the highest electricity cost pressure from data center growth?

Northern Virginia (SWMAAC load zone in PJM), Texas (ERCOT), Columbus Ohio, and suburban Chicago are the most affected regions. Secondary impacts are felt across the broader PJM and ERCOT footprints through higher capacity clearing prices.

How much have data centers contributed to the 2026/2027 PJM capacity auction increase?

Data center load growth in PJM-served territories was a significant contributing factor to the $329.17/MW-day clearing price — a 22% increase from the prior year. Analysts attribute 30-40% of the capacity price increase to new load growth from data center development in Virginia and Ohio.

Can a fixed-rate electricity contract protect my business from data-center-driven price increases?

A fully fixed contract that buries all capacity charges in the locked rate protects you from capacity auction pass-throughs for the contract term. However, at renewal time, new contracts will price in forward capacity expectations — meaning data center-driven price pressure eventually passes through even to fixed-rate customers. The solution is a combination of contract protection and demand management.

What is ICAP and why does it matter for my electricity costs?

Your Installed Capacity (ICAP) tag is determined by your facility's coincident peak demand during PJM's five highest-load hours. It determines your share of the regional capacity obligation, which is then priced at the auction clearing price. Higher data center load increases the probability that your facility's peak hours coincide with grid peak events, potentially increasing your ICAP tag.

What is the long-term outlook for data center electricity demand growth?

Most industry forecasts project continued rapid growth through 2030, with US data center electricity demand reaching 150 GW or more. AI training workloads, cloud computing expansion, and electrification of enterprise IT infrastructure all support this trajectory. New nuclear, storage, and renewable projects are expected to gradually ease supply tightness — but the 2026-2029 window is likely to remain challenging.

Word count: 2,784