• SemiAnalysis mapped more than 300 US data-centre moratoriums, with about 20GW inside restricted local boundaries
  • It estimates local measures delay 1,525MW, rising to roughly 2.3GW when New York is included

The fact

SemiAnalysis said on 15 September that it had mapped more than 300 US data-centre moratoriums. Around 20GW of announced capacity sits inside local boundaries covered by restrictions, but the research firm estimates only 1,525MW is genuinely delayed by those local measures. Including New York's statewide action takes the total to about 2.3GW.

The difference comes from where the restrictions fall. Many local moratoriums cover jurisdictions without an active data-centre project, so they increase the number of restricted areas without delaying an existing development pipeline.

Some state actions reach projects through different points in development. New York has temporarily paused certain state environmental permits for new hyperscale projects. Texas requires affected data-centre loads to pass verification before advancing through ERCOT's interconnection process. Pennsylvania has removed AI data centres from its Fast Track permitting programme, while Oregon has paused unapproved requests involving state-owned property.

The assessment

Counting moratoriums does not show how much data-centre supply has actually moved. A local pause matters directly when an active project sits inside that jurisdiction. A state permit or grid-interconnection restriction can reach several projects because they depend on the same approval process.

That explains the gap between roughly 20GW located inside restricted local boundaries and the much smaller amount SemiAnalysis identifies as genuinely delayed. For developers and investors, the more useful measures are the megawatts attached to active sites, the stage each project has reached and which approval is preventing the next step.

For BTW readers, the moratorium count is not a measure of lost US data-centre capacity. The stronger indicator is how many megawatts are delayed at a permit, land or grid milestone that the project needs before construction or energisation can continue.

What to watch

Watch whether SemiAnalysis's 2.3GW delayed estimate rises as projects move through permitting and interconnection. New York's environmental review and ERCOT's verification process are especially useful tests because both can affect projects beyond a single municipality. Project cancellations, revised energisation dates or relocations would provide clearer evidence of supply impact than additional moratorium votes alone.