Data Centers, Large Loads, and the Utility Grid: Statistical Evidence of a Structural Shift in U.S. Power Systemsa

Executive Overview

Across the United States, electric utilities are confronting a structural shift driven by hyperscale data centers, artificial intelligence workloads, and electrification. What was once incremental load growth is now measured in gigawatts, not megawatts, forcing utilities, regulators, and planners to rethink rate design, transmission planning, substation sizing, and system resilience.

Recent utility earnings calls, regulatory filings, industry conferences, and OEM announcements show that large-load integration is no longer speculative. It is being actively planned, contracted, regulated, and financed with engineering execution emerging as the primary constraint.



This article synthesizes recent, publicly reported data to quantify how fast large-load demand is growing, how utilities are responding, and why statistically grounded planning and power system studies are now essential to protect reliability and ratepayer affordability.

1. Utility Load Growth Is Accelerating at the Gigawatt Scale

Xcel Energy: Multi-State Large-Load Expansion

Xcel Energy now expects to have 6 GW of contracted data center load in its queue by 2027, double what it anticipated only months earlier. The company reports:

• 2+ GW already under contract
• 1 GW expected to be signed in 2026
• A 20 GW potential pipeline, with 4 GW considered “high probability”
• Data center load growth contributing to a 2.2% increase in weather-adjusted electric sales in 2025

To accommodate this growth without shifting costs to existing customers, Xcel is pursuing large-load tariffs in:

• Minnesota
• Wisconsin
• Colorado
• Texas

The company has also identified $10 billion in additional infrastructure investment opportunities, on top of its existing $60 billion five-year capital plan, including a $1.5 billion, 765-kV transmission line awarded by the Southwest Power Pool.

Statistical signal:

Large loads are now influencing regional transmission planning and capital allocation decisions.

2. Florida Enters the Large-Load Era

Florida Power & Light (FPL), the nation’s largest electric utility, expects to announce its first large-load data center deals in 2026.

Key indicators include:

• 50+ large-load inquiries already received
• A long-term goal to place 15 GW of new generation into service for data centers by 2035
• Internal planning scenarios targeting 30 GW or more, including nuclear capacity
• A “bring-your-own-generation” strategy combining:
• Renewables
• Battery storage
• Gas generation
• Potential small modular reactors (SMRs)

FPL’s leadership has emphasized alignment with consumer bill protection policies, signaling that large-load integration will be conditioned on cost causation and regulatory approval.

Statistical signal:

Large-load demand is expanding into traditionally residential-heavy utility territories, not just established data center hubs.

3. Large-Load Tariffs Are Becoming a National Norm

Large-load tariffs — once niche instruments are now a core regulatory tool.

Across the U.S:

• 66 large-load tariffs were approved or pending as of late 2025
• Many require:
• Upfront payment for transmission and distribution upgrades
• Long-term capacity commitments
• Demand flexibility or curtailment provisions
• Approximately 20% explicitly require or incentivize load flexibility

California’s proposed SB 978 would apply to loads ≥75 MW, mandating:

• Cost-causation-based rate structures
• Prohibition of diesel backup generation
• Clean backup requirements
• Formal state studies on impacts to decarbonization, water, and pollution

Similar frameworks have emerged in:

• Ohio
• Oregon
• Virginia
• Wisconsin

Statistical signal:

Regulators are converging on the principle that large loads must self-fund the infrastructure they require.

4. Grid Modernization Is Shifting from Hardware to Data

Insights from the DTECH utility leadership keynote underscore that utilities are not only expanding infrastructure they are changing how the grid is operated.

SDG&E

• Operates 225+ hyper-local weather stations
• Uses AI-enabled models to forecast wildfire risk at the circuit level
• Employs AI-driven drone inspections across millions of assets
• Runs nightly wildfire spread simulations on supercomputers

PG&E

• Nearly 50% of customers have ≤150-A service panels
• Traditional electrification upgrades can cost:
• $5,000–$10,000 per customer panel
• Up to $30,000 for transformer upgrades
• Grid-edge intelligence and AMI 2.0 enable:
• Next-day electrification service
• System-wide savings measured in billions of dollars

Duke Energy

• Uses cloud computing to reduce system-wide modeling time from six weeks to six hours
• Integrates data from millions of connected grid devices
• Maintains service through record-breaking load events across multiple states

Statistical signal:

Digital modeling, forecasting, and data integration now directly affect capital efficiency and reliability outcomes.

5. Generation Supply Is Being Optimized, Not Just Expanded

While new generation is being added, utilities and OEMs are also extracting more value from existing assets.

GE Vernova reported 1.1 GW of U.S. onshore wind repowering orders booked in 2025:

• U.S.-manufactured nacelles and drivetrains
• COD expected 2026–2027
• Repowering increases:
• Output
• Asset life
• Availability

This aligns with utility strategies to meet near-term load growth without waiting for entirely new greenfield projects.

Statistical signal:

Repowering and asset optimization are becoming critical tools for near-term capacity adequacy.

6. Engineering Is the Gating Factor

Across all data points, one conclusion is consistent:

The limiting factor is no longer demand, capital, or policy it is engineering execution.

Large-load integration requires:

• Accurate load forecasting
• Interconnection studies
• Short-circuit and stability analysis
• Transmission and substation design
• Protection and control coordination
• Compliance with evolving tariff and regulatory requirements

Utilities that can execute these studies quickly and defensibly are able to move projects forward. Those that cannot face delays, cost overruns, or regulatory pushback.

Key Statistics at a Glance

Technical FAQs

Closing Insight

The data is unambiguous: large-load demand is reshaping the U.S. electric grid at a structural level. Utilities that pair statistical rigor, digital intelligence, and disciplined engineering will be able to scale reliably and affordably.

Those that do not will struggle regardless of how much demand exists.