Electricity Rate Designs for Large Loads: Evolving Practices and Opportunities - 2026 Update

Overview

This document is a technical brief, prepared jointly by Lawrence Berkeley National Laboratory and The Brattle Group, that updates the January 2025 edition. The authors are Natalie Mims Frick, Peter Cappers, Goksin Kavlak, Long Lam, Ryan Hledik, Jadon Grove and Sebastian Rotella. It surveys US large-load tariffs and electric service agreements for data centres and classifies design elements by prevalence and trend.

The elements examined include minimum demand thresholds, contract duration, minimum billing demand, collateral, exit fees and load ramp schedules. According to the findings, large-load tariff minimum demand thresholds range from 0.3 to 150 MW, with 75% between 5 and 100 MW. Standardized minimum contract terms range from 1 to 20 years; the median rose from 5 years before 2025 to 12 years since 2025.

Minimum billing demand is typically a share of contract demand, with a median of 80%. This means the user pays for a large part of the contract power even if it does not actually draw it. The introduction cites an LBNL estimate that US data center demand grew 2.3 times from 2018 to 2024 and could grow 3.3 times from 2024 to 2028. The brief addresses US tariffs and does not apply directly to Turkish regulation; it does, however, give an international view of how large loads share network costs.

Key points

1. Large-load tariff minimum demand thresholds range 0.3-150 MW, with 75% between 5 and 100 MW.
2. Standardized minimum contract terms range 1-20 years; median rose from 5 years (pre-2025) to 12 years (since 2025).
3. Minimum billing demand is typically a share of contract demand, median 80%.
4. US data center demand grew 2.3x from 2018 to 2024 and could grow 3.3x from 2024 to 2028 (LBNL estimate).
5. Collateral, exit fees and load ramp are also examined design elements.
6. 2026 update of the January 2025 edition.

Using it in design

• The cost model should support large-load tariff features: minimum billing demand (% of contract MW), contract term, ramp schedule and collateral.
• Reflect take-or-pay exposure in lifecycle cost when the IT load ramps slowly.
• Scenario the load ramp together with contract power and minimum billing demand.
• Compare against local conditions before using US data as defaults for another country's tariff.

Who should read it

Finance, planning and design teams modelling data center energy contracts and lifecycle cost.

Publication details

Authors

Natalie Mims Frick; Peter Cappers; Goksin Kavlak; Long Lam; Ryan Hledik; Jadon Grove; Sebastian Rotella

Publisher

Lawrence Berkeley National Laboratory / The Brattle Group

Year

2026

Document language

English

Source type

Government / public

Access

Free to read

Licence

Copyright retained; citation only — the publisher’s terms apply to the original document

Citation