Navigating dedicated large load tariffs: Siting and power costs in the AI era

The surge in AI-driven data center development is prompting many utilities to adopt dedicated large load tariffs, as data center electricity consumption has grown from 1.9 percent of total US electricity usage in 2018 to 4.7 percent in 2024, according to a June 2026 Lawrence Berkeley National Laboratory report sponsored by the US Department of Energy. Tariff design now plays a significant role in determining a data center’s long-term power costs. For developers and investors, navigating this evolving tariff landscape is critical, as differences in tariff design and cost-allocation mechanisms can materially impact project-level economics and energization timelines.

Historically, utility planning and cost recovery frameworks were designed around gradual, diversified, and slow(er) load growth. Today, the rise in AI-driven data center development creates a fundamentally different challenge – concentrated demand requiring substantial investments in generation and transmission infrastructure. The industry’s primary response has been the adoption of dedicated data center large load tariffs. To date, utilities have typically offered special rates for large industrial customers. However, the latest generation of tariffs are designed specifically to assign the cost of new infrastructure, and other appropriate system costs, associated with serving large loads to data center customers, while shielding general ratepayers from those costs.

Traditional large load tariffs were designed to attract large customers

Large load tariffs were designed to accommodate industrial and manufacturing customers with high load-factor demand, with the tariff terms focusing primarily on service terms and demand commitments. These rate structures helped attract large customers and reward efficient utilization of the electric system, often through discounted pricing, because their steady, high-load-factor demand increased sales and spread fixed costs across greater energy consumption. Utilities generally built the necessary infrastructure and recovered those costs through broader rate mechanisms, resulting in infrastructure investments being shared across the utility's customer base (i.e., commercial, industrial and residential).

Today’s large load tariffs are designed to protect residential ratepayers

Today’s data center tariffs need to contemplate significantly larger customers hungry for electrons. Rather than attracting new load, the primary purpose of this new cohort of tariffs is to allocate the costs associated with serving large new loads to the customers driving those needs. Modern large load tariffs increasingly include the following provisions that have increased power supply costs to data centers and, in some cases, elongated energization timelines:

• Minimum demand provisions – often requiring customers to pay for a minimum of 85 percent of contracted capacity regardless of actual usage.
• Phased load-ramp schedules – with enforceable milestones, ensuring customers cannot reserve hundreds of megawatts of capacity while delaying development after infrastructure to serve them has already been constructed.
• Long-term service agreements – typically 10 to 15 years and, in some cases, up to 30 years, aligning customer commitments with utility asset lives.
• Financial collateral requirements – designed to protect utilities and ratepayers from project delays, downsizing, or cancellation.
• Early exit and termination charges – requiring repayment of up to 100 percent of construction costs.
• Direct assignment of network – upgrade and infrastructure costs: assigning a greater share of the new infrastructure costs to large loads.

The pace of rollout of these new tariff structures is remarkable, as shown in Figure 1. Regulators approved 29 dedicated data center tariffs in 2025 alone. By mid-2026, more than 30 states had utilities with at least one approved data-center-specific tariff, with dozens more under consideration. States shown in light grey represent a combination of utilities continuing to address large load requests through bespoke commercial arrangements, and large load tariffs under development (e.g., Illinois, Wisconsin, and New York).

Beyond utility tariff filings, a growing number of states are becoming more directly involved through legislation. Oregon demonstrates that large load cost responsibility is increasingly being shaped not only through utility tariffs and PUC proceedings, but also through broader state policy and legislation, adding political and regulatory complexity to siting decisions.

PacifiCorp’s 2024 tariff revisions focused on cost causation and demand forecasting accuracy, introducing enhanced cost responsibility for large load customers. However, more recently, Oregon’s POWER Act has built upon and formalized these principles through legislation. The POWER Act:

• Mandates a separate rate class: for data centers and cryptocurrency facilities.
• Expands cost responsibility: where large customers must fund 100 percent of the grid’s network upgrades needed to serve them, and facilities above 100 MW pay a one-cent-per-kWh surcharge to support residential bill relief.
• Requires a minimum level of clean-energy generation: for large facilities, integrating cost allocation and decarbonization objectives within a single framework.

For data center developers, this evolving landscape is becoming an increasingly important siting consideration. Two otherwise technologically identical campuses can face materially different power supply costs depending on the serving utility, even within the same state. As tariff structures continue to evolve, developers must look beyond immediate power supply costs. They must also evaluate how the generation and infrastructure needed to serve future demand will be developed, how those costs will be allocated, and how policy and regulatory requirements may change. The resulting cost, reliability, operating flexibility, and service conditions could influence customer siting and operating decisions over time, materially affecting project-level economics.

We help developers and investors navigate this rapidly changing environment by evaluating regulatory developments, tariff structures, and cost allocation mechanisms. Our retail electric rate team translates these factors into clear forecasts of future retail electric rates and power supply costs to support critical siting, planning and investment decisions.