Texas Puts New Data Centers on Hold Pending Grid Audits

The engineers who run the Texas grid started the year looking at 233 gigawatts of projects waiting to connect. That queue now stands at 474 gigawatts, roughly 90 percent of it data center requests, more than five times the highest electricity demand the state has ever recorded in a single moment. Governor Greg Abbott's answer arrived this week: no new data center project moves forward until it passes an audit by the Public Utility Commission of Texas and grid operator ERCOT.

What the audits will examine

The mandatory review goes well beyond electricity. Each project's on-site and off-site power demand, water consumption, noise mitigation, light controls, use of tax incentives, and ownership structure will be examined before approval. Texas previously gathered this information through voluntary surveys; compliance is now compulsory. For a state that built its brand on light-touch regulation, and attracted Google, Microsoft, and a wave of hyperscale investment precisely because of it, this is a genuine change of course.

The scope is worth stating precisely: existing facilities keep running, and the order targets new project approvals. Much of the queue consists of projects still in preliminary stages, so the effect lands on the supply curve of the coming years rather than on today's capacity. What remains unannounced is how long the audits will take, and that is where the real uncertainty sits: a fast-moving review is a formality, while months-long examinations would amount to a de facto moratorium.

The boom meets its physical limits

Data centers and crypto facilities have already pushed Texas electricity prices upward, and public resistance to new projects has been intensifying across the country. The move echoes a story we covered days ago, in which PJM, America's largest grid operator, weighed the authority to cut data center power during shortages. The pattern is consistent: the AI infrastructure boom is colliding with physical constraints, namely electricity, water, land, and the patience of neighbors.

Why this matters if you buy cloud, not land

Most businesses will never build a data center, but two consequences reach them anyway. First, cost: if supply growth slows while demand keeps compounding, the downward pressure on cloud and AI pricing weakens, which matters if you are about to sign a multi-year contract. Second, geography: as power and cooling become the binding constraints in the US, capital starts evaluating other regions more seriously, and countries with reliable grids and predictable permitting stand to gain. Texas just demonstrated that even the friendliest jurisdiction will eventually ask hard infrastructure questions. The regions that answer them before the investment arrives, rather than after, will be the ones that capture it.

Sources: TechCrunch