Texas legislates, and the capacity market sends a bill
On June 20, 2025, Texas Governor Greg Abbott signed Senate Bill 6, a law written largely with data centers in mind. It applies to loads of 75 megawatts or more in the territory of the Electric Reliability Council of Texas, or ERCOT, and allows the Public Utility Commission of Texas to lower that threshold. It requires large customers to pay a transmission study fee of at least one hundred thousand dollars and to disclose to utilities whether they have made potentially duplicative requests for service elsewhere in Texas. New large loads connecting after December 31, 2025 must install equipment that allows them to be disconnected remotely during an emergency when the grid operator is forced to shed firm load. The law also directs a review of arrangements in which large loads sit beside existing power plants.
Where Ohio and Virginia addressed cost allocation through rate cases, Texas addressed it through statute, and it reached further than cost. The disclosure rule targets the speculative requests discussed earlier in this course: if a developer is shopping one campus to five utilities, the utilities can now find out. The emergency provision addresses reliability, the lesson Texas learned expensively during Winter Storm Uri in 2021. And the study fee prices the option of asking, in the same spirit as the federal generator reforms.
Much of the detail was delegated. Senate Bill 6 tells the Public Utility Commission to write rules, and the commission has spent the period since opening rulemakings on large-load interconnection standards and related questions. Anyone relying on the statute should read those rules as they are finalized rather than assuming the bill text is the whole story.
There is a second channel through which data centers affect who pays, and it runs through wholesale markets rather than tariffs. On December 17, 2025, PJM Interconnection published the results of its capacity auction for the delivery year beginning June 2027. The capacity market pays power plants to be available when needed, and the cost flows through to every customer's bill. The auction cleared at the price cap that the Federal Energy Regulatory Commission had approved, for the second year running, and PJM procured less capacity than its reliability target. PJM attributed nearly all of the growth in its reliability requirement between the two auctions to forecast data center demand.
This is a subtler form of cost shifting than a tariff. A household in Maryland or New Jersey may never live near a data center, but if the regional forecast rises because of data centers elsewhere in PJM, the capacity price rises for everyone in the region. That is why governors in PJM states have taken such interest in the market's rules, and why the price cap itself became a political settlement rather than a technical parameter.
The forecasts themselves are contested. Some of the demand in them belongs to projects that will never be built, the same speculative-request problem that affects utility planning. Regulators and market monitors have argued about how much weight to put on forecasts that rest on customer announcements. If the forecasts prove too high, customers will have paid for capacity that was not needed. If too low, the region risks shortfalls. Neither error is free.
Faraday's read is that the question of who pays has quietly become the question of who gets served first. A jurisdiction that has settled its cost allocation, whether by tariff, statute or market rule, can connect large loads with less political friction than one where the argument is still open. Settled rules are rarely generous to developers; their value is that they can be priced.
For someone assessing a site, the useful exercise is to trace a megawatt from the meter to the bill. Which tariff does it fall under? Which capacity or reserve market does it draw on? Which statute governs its interconnection and curtailment? Each answer names an institution, and each institution has a record that can be read.