What Is a Utility Load Study for a Data Center?

When a developer asks a utility to serve a data center, the utility cannot answer from a map. It has to model the new load on its system, under stressed conditions, and see what breaks. That modeling exercise is the load study. Depending on the utility it may be called a load interconnection study, a facilities study, a system impact study or, in ERCOT, a Large Load Interconnection Study (LLIS).⁶

A load study sits inside a longer sequence. Before a utility energizes a large load it typically has to screen the request, study the system, choose a constructible solution, complete detailed engineering, obtain permits and regulatory approvals, and then build; Dominion Energy Virginia’s proposed queue process lays out exactly those stages.³ The study is the step that turns a request into an engineering answer, but it is not the commitment itself.

It also differs from a capacity screen. A utility account manager may say informally that a substation “has room.” A study tests that claim against planning criteria, contingencies and every other load and generator already in the queue. Our guide to the large-load interconnection process covers where the study falls in the full path to service, and the glossary entry on interconnection studies defines the common terms.

Study quality depends on input quality. Utilities and their planners ask for more data than many first-time applicants expect, and the requests have grown as data center behavior has come under scrutiny.

• Location: coordinates of the proposed load and, if known, the substation or line the developer hopes to tap.⁴
• Load size and ramp: peak demand, and the MW to be energized each month or year through full build.¹
• Power factor: across loading levels, measured at the high-voltage delivery point.¹⁴
• Electrical design: a preliminary one-line diagram and any reactive devices on site.¹
• Schedule: an estimated project timeline with target in-service dates for each block of load.
• Dynamic models: in ERCOT, a stability study is required for any load of 75 MW or more, so a dynamic load model or PSCAD model representing the actual load behavior must be submitted with the request.⁴

The ramp schedule deserves the most care. It drives which years the planner tests, and under newer tariffs it can become the basis for minimum bills. Our guide to load ramp schedules and phased power explains how to build one that holds up.

NERC’s 2025 Level 2 alert on large loads recommended that transmission planners and planning coordinators use steady-state, dynamic and short-circuit models to assess the reliability impacts of large loads.¹ In practice a full study tends to include the following.

Dynamic studies are where data centers differ most from older industrial loads. Since 2022 NERC has analyzed load-loss events in the Eastern and Texas interconnections caused by computational loads disconnecting unexpectedly, events that existing planning studies did not predict.¹ The alert recommends that studies explicitly account for the collective tripping of electrically close facilities, which matters at sites near existing data center clusters.¹ Expect questions about ride-through, UPS behavior and on-site generation transfer schemes.

The output is a list of facilities: the interconnection itself (a tap, a new substation or switching station), any network upgrades to lines or transformers, and the earliest dates those can be in service. Who pays for those upgrades is covered in transmission upgrades and cost allocation.

Most utilities run large-load requests through stages that move from screening to firm engineering. Dominion Energy Virginia’s proposed queue process, filed with the State Corporation Commission, describes four: project initiation (submission, viability screening and queue placement), project feasibility (system studies and selection of a constructible solution), project development (detailed engineering, permitting and regulatory approvals) and project execution (final design, construction and energization).³

Timing is set less by engineering hours than by the queue. Dominion reported roughly 70 GW of large-load delivery points in its queue, nearly three times its all-time system peak of 24.7 GW set in January 2025.³ About 25 GW had projected connection dates through 2031, and about 45 GW was still under study.³

Texas shows the shift toward group studies. ERCOT’s Planning Guide Section 9 sets out the LLIS procedures, report and follow-up steps for new or modified large loads.⁶ On June 18, 2026, the PUCT approved PGRR 145 and its companion NPRR 1325, with key provisions taking effect in July 2026.⁵ Together they create a one-time “Batch Zero” process that sorts pending requests by study maturity and commitment and studies them together in two phases.⁷ See power timelines and interconnection queues for how queue position translates into a date.

Load studies used to be free or nearly so. That has changed as speculative requests have piled up. AEP Ohio, under its data center tariff, requires new data centers or expansions of 25 MW or more to complete an intake form and pay a load study fee it estimates at $10,000 to $100,000.² Some cooperatives work through a sequence of agreements instead: Tri-County Electric Cooperative in Texas starts requests of 2 MW or more with a large load service request, then requires a reimbursable services agreement, a construction development agreement and an electric service agreement at later milestones.⁸

Fees are the small part. The bigger commitment comes after the study, when the customer signs a service agreement that may include long terms, minimum bills and collateral. Georgia’s Public Service Commission in January 2025 approved rules for new Georgia Power customers above 100 MW that allow contract terms of up to 15 years, minimum billing and PSC review of the contracts.⁹ Our guide to large-load tariffs and electric service agreements covers those terms.

The cost of a study filters the queue. After AEP Ohio’s tariff took effect, its reported data center pipeline fell from more than 30 GW of requests to about 13 GW.¹⁰ Expect any request you file to be treated as a real project, with the obligations that follow.

A study report usually gives a point of interconnection, a scope of work, a cost estimate (often with a wide range at early stages) and an in-service date that depends on equipment procurement and permitting. Read it for four things:

1. 01The date for each phase of your ramp, not only the first block of load.
2. 02Upgrades that depend on someone else’s project, such as a new line another customer is funding.
3. 03Conditions on service, such as reduced firm capacity or curtailment until upgrades are complete. See firm vs. interruptible power.
4. 04How long the results stay valid, and what changes (load size, location, ramp) would trigger a restudy.

A load study is not the first step in site screening. It is expensive in time, and it commits you to a load profile. Before filing, it pays to know which utility serves the parcel, which substations and lines are realistic taps, and whether public planning data already shows constraints. Our guides to power capacity data sources and utility will-serve letters cover that groundwork.

• Confirm the serving utility and whether an RTO or ISO process also applies.
• Fix a credible peak load and ramp, and a target power factor.
• Prepare a preliminary one-line and, for large loads, a plan for dynamic models.¹⁴
• Ask the utility for its current fee, deposit, study scope and expected duration in writing.²
• Budget for what follows the study: agreements, collateral and upgrade costs.

If you are deciding which parcels deserve a study request, BlackForge can get a site reviewed against nearby transmission, substations and published constraints before you commit.