• Project Braid has encountered delays at several proposed US data-centre sites while retaining its target of bringing 500MW online in 2027
  • Permitting, transformer availability and grid reviews are affecting individual locations, leaving the venture dependent on several sites progressing towards construction

The fact

Google and Blackstone's new cloud venture has encountered delays at several proposed US data-centre locations as it prepares to offer access to Google's AI processors. The venture, known internally as Project Braid, was launched in May with an initial US$5 billion equity commitment from Blackstone and a target of bringing 500MW of capacity online in 2027.

One affected project is in Cheyenne, Wyoming, where Crusoe had been developing a large data-centre campus. Crusoe said in June that it had paused work at its customer's request, and Google later took over the development. Bloomberg, republished by The Star, reported that Google is reapplying for permits and plans to reduce the project's scale. Google has not publicly confirmed the reported reason for replacing Crusoe.

Other proposed locations face different constraints. Bloomberg reported that one candidate site did not have suitable transformers. In Texas, Governor Greg Abbott ordered an audit in August of data-centre projects moving through ERCOT's interconnection process before they can proceed. Project Braid chief executive Benjamin Treynor Sloss said individual sites progress at different rates and that the venture still expects to meet its 500MW target.

The assessment

The 500MW target now depends on more than the timetable of any one site. Each candidate location has to clear its own permitting, grid and equipment constraints before it can contribute usable capacity. If one site slips, the venture has to recover that capacity elsewhere in its development pipeline.

Project Braid's wider pool of locations gives it that option. Cheyenne shows how a change in developer can put permitting work back into the schedule, while Texas introduces a separate grid-review hurdle. A transformer shortage can stop another location for a different reason. What matters is whether enough sites can reach construction readiness in time to support the same overall delivery target.

For BTW readers, the 500MW commitment is now a portfolio-delivery problem as much as a capacity target. Land and capital do not make a site usable on their own. Permits, grid approval, electrical equipment and the development team all have to line up early enough for construction to deliver the promised capacity in 2027.

What to watch

Watch whether Google secures revised permits for Cheyenne and which other Project Braid sites enter construction. The 2027 target will become more credible if construction-ready alternatives can replace delayed locations without reducing the planned 500MW.