News at a Glance
California Governor Gavin Newsom signed seven bills on September 20, requiring AI data centers to bear the costs of grid upgrades and water fees themselves, rather than passing them on to residents. The new laws aim to ease utility cost pressures from the rapid expansion of data centers and protect residents’ electricity bills. The bills direct the California Public Utilities Commission to create new rate categories for data centers and require them to pay for infrastructure upgrades, a move expected to raise cloud computing and AI operating costs.
Background
In recent years, surging demand for AI training and inference has caused data centers’ electricity and water consumption to multiply. As a major hub of the global AI industry, California has seen record-high grid loads and rising residential electricity bills. Previously, data centers often applied for preferential rates on the grounds of large-scale electricity use, with costs shared by all taxpayers, sparking social controversy. This legislation is California’s response to that imbalance and also reflects the divergence among U.S. states over AI infrastructure regulation—Texas and others are actively attracting data centers, while California has chosen to prioritize residents’ interests and grid stability. The enactment of these bills could affect data center tax and energy policies in other states.
In-Depth Analysis
In Liu Gong’s view, California’s seven bills are not a simple tax increase but a regulatory brake on the AI boom. In the past, data centers used “job creation” as leverage to shift grid upgrade and water costs onto ordinary households—essentially a subsidy. Now California has chosen to return the costs to companies, a correction of the “tech-first” logic. Compared with Texas, which is still using tax incentives to attract data centers, California’s approach may push some companies to cheaper locations, but in the short term, companies that depend on California’s computing ecosystem can only absorb the costs. The next thing to watch is how the new rate categories are defined—if they are based on peak load, small businesses will be driven out of California; if they are allocated by actual electricity usage, the big players may still accept it. In short, someone has finally taken a serious look at AI’s cost ledger.
Perspectives
Further Thoughts
- AI companies will be forced to reassess data center site-selection logic, and internalizing energy costs could become the new industry norm.
- California’s regulatory correction of infrastructure externalities provides a replicable policy template for other states.
- The new rules may accelerate the relocation of data centers overseas or to the U.S. interior, thereby affecting the global distribution of AI computing power.
Source and Original Text
This item comes from The Verge AI (published on September 21, 2026, at 20:29:45). This site provides Chinese rewrites and commentary on overseas AI developments; the original copyright belongs to the original author.
Daily aggregation of the latest overseas AI developments and in-depth commentary. Bookmark this site and don’t miss any important signal; Return to homepage for more.