California Puts AI Data Centers on Separate Power Rates

California Gov. Gavin Newsom signed seven bills on September 21, 2026 intended to keep AI data centers from passing utility costs to residents. The package creates a separate way to assign costs to very large power users and links approval and grid connection to water and power duties. The Verge

The bills were AB 1577, AB 2383, AB 2469, AB 2619, SB 886, SB 887 and SB 1168. SB 886 is titled the California Technology Innovation and Ratepayer Protection Act. The Governor's office described the package as the most comprehensive data-center laws in the nation. Governor's Office

Rate design is at the center. The laws require the California Public Utilities Commission, which sets electricity prices, to create a new rate group for data centers. A Commission proceeding, a formal review, will consider price rules for data centers and other large load customers. The stated goal is to prevent data centers from shifting grid costs onto low-income customers.

That new pricing is paired with direct payment rules. Data centers must pay for upgrades to local power grids and water systems caused by their demand. In utility terms, this moves connection-related building costs off the general rate base, the shared costs paid by all customers, and onto the connecting customer.

Disclosure comes before approval. Proposed data centers in California must report estimated water use to local governments. They must also share details on water supply, efficiency and drought planning with water suppliers and local governments, plus information on energy efficiency. The Governor's office presents this as giving communities more control on water, electricity, and land use.

Rules on power buying and environmental review are tightened in parallel. Data centers must add new clean energy supply to the grid and follow state energy procurement requirements. They are no longer eligible for broad environmental exemptions. Access to a faster approval process is conditional. To qualify, facilities must meet requirements on energy, water, and fuel use.

The legislative path was short. California lawmakers reached a deal on data-center regulation in late August 2026. State Sen. Steve Padilla (D-Chula Vista) authored a data-center bill backed by TURN. When signing the package, Newsom criticized President Trump for inaction. Los Angeles Times

The broader context here is cost allocation during steady load growth. Separate price plans for hyperscale and colocation facilities, meaning very large and shared data centers, customer-paid substation and line upgrades, and written water budgets are familiar tools in utility regulation. California is now using them together, specifically for AI computing.

Looking at what this means for operators, the compliance picture widens on three fronts at once. Price planning becomes a site-selection factor alongside network delay and fiber access. Securing water becomes a permit factor alongside securing power. And adding new clean energy becomes a condition for grid connection rather than a voluntary sustainability pledge. None of these rules ban construction. They change who pays for the infrastructure around it.

In my view, worth flagging is the conditional fast track. California has combined stricter standards with a quicker process for projects that meet them. For infrastructure teams used to choosing between speed and efficiency, that link will shape early choices about cooling design, power usage effectiveness, which measures extra power beyond computing, water usage effectiveness, which measures cooling water use, and on-site fuel use. The technology to meet those standards largely exists. The rules make those design choices visible to local authorities.

The long arc here favors this kind of transparency. I have watched my children grow up as home internet, smartphones, and cloud services each moved from novelty to normal, and each shift worked best when the physical costs were clear. AI systems will need very large amounts of reliable power and cooling. Making load forecasts, water plans, and grid-upgrade payments visible to utilities, water suppliers, and county planners places those physical costs on the balance sheets driving demand.