Texas Data Center Audit: What Abbott's Pause Reveals (and Conceals) About Cost Shifting

In early August 2026, Governor Greg Abbott directed the Public Utility Commission of Texas (PUCT) and the Electric Reliability Council of Texas (ERCOT) to pause and audit all data center projects in the interconnection queue, roughly 474 gigawatts of pending capacity, about 90 percent of it from data centers^[1][2], before any new grid connections proceed. The move prompted a rapid corporate response: Skybox Datacenters, Digital Realty, Mara, and others publicly committed to comply with Abbott's stated standards covering infrastructure costs, water use, load flexibility, and community transparency^[1][7][8].

But here is what the headlines obscure. Abbott's audit directive names a problem, that data centers must "pay their own way" and "not shift costs onto Texas families"^[8], without specifying the mechanism that would enforce it. Data center operators pledging "existing or pledged practices" covering grid infrastructure costs is corporate theater if there is no binding tariff, no minimum-demand ratchet, no collateral requirement, and no cost isolation to keep the burden off residential ratepayers^[1]. A voluntary commitment to transparency is not a protective tariff. The real audit that matters is not visible: What share of new generation and transmission built to serve these loads will be assigned to the data center class itself, versus socialized system-wide? What minimum take will the operator pay, and over what term? Are there exit fees covering stranded assets if the load underperforms? None of these questions appear in the public record yet, which is the problem.

The pause itself is important, it stops the ratepayer risk of funding capacity that never materializes, but it is also a narrow circuit breaker. Texas has no large-load tariff comparable to Virginia's GS-5 or Ohio's data-center rider. Without one, individual special contracts will be negotiated behind closed doors, filed with the PUCT, and then redacted before ratepayers can see the price or the cost-allocation mechanism. The audit can demand disclosure, but only a standing tariff with standardized protections, high minimum-demand ratchets matched to the asset life of dedicated infrastructure, collateral requirements, and cost isolation, can prevent the next generation of deals from repeating the pattern: confidential discounts for the load, socialized recovery for the grid operator.

Power companies have already signaled they view the audit as manageable, even validating their commercial models^[1]. That confidence is the warning sign. If utilities believed the audit would force them to absorb the cost of new infrastructure, they would say so. Instead, they are treating it as a box-checking exercise that clears the path to building rate-based assets on behalf of data centers. The audit must answer a specific question before any project is approved: Is the cost of new generation and dedicated transmission being assigned to the data center customer class (with appropriate ratchets and collateral), or is it being spread across all ratepayers? Until that is public, "accountability" is a word, not a rule.

Senate Bill 6, signed into law in June 2025, established disclosure and curtailment obligations for large loads of 75 MW or more^[5], laying the regulatory foundation for Abbott's pause. That is the statutory authority for a protective large-load tariff. The window to file one is now. If PUCT issues a new standardized tariff before the audit concludes, requiring data centers to pay for their own dedicated capacity, carry a high minimum-demand charge over the life of the assets, post collateral, and consent to flexible-load or bring-your-own-generation terms, then the audit's findings can be locked into a rule. If not, individual deals will be litigated case-by-case in confidential special-contract dockets, and ratepayers will learn what they funded after the bill arrives.