Texas Governor Greg Abbott ordered the Public Utility Commission of Texas and the Electric Reliability Council of Texas (ERCOT) on August 3, 2026 to conduct a mandatory audit of every data center project seeking to connect to the state’s power grid, effectively pausing new approvals until the review is complete. The number that prompted the order is the real story: ERCOT’s interconnection queue has ballooned to 474 gigawatts of requested new capacity, up from 233 gigawatts in January — a doubling in roughly seven months — with an estimated 90% of that demand coming from data centers. That queue is now more than five times ERCOT’s entire peak historical demand, a mismatch so large it’s no longer a theoretical planning concern but an immediate grid-stability and cost question.
Abbott had previously tried a lighter-touch approach, asking data center developers to voluntarily disclose their power and water needs. According to his own account, most simply didn’t respond — which is what pushed the state toward a mandatory audit instead. The audits Texas is now requiring will examine on-site and off-site electricity and water demand, noise mitigation, lighting controls, use of state tax incentives, and the actual ownership structure behind each project, a scope broad enough to catch speculative land-and-power reservations, not just fully committed builds. Google and Microsoft are among the major operators that have been drawn to Texas specifically because of its historically loose regulatory environment and seemingly abundant power supply — the same qualities that made the state attractive are now the ones straining under the weight of AI-era demand.
Why Texas, of All Places
Texas has spent the last several years marketing itself as one of the most permissive places in the United States to build a data center — deregulated power markets, generous tax incentives, and a business-friendly regulatory posture that made it a magnet for exactly the kind of massive AI infrastructure buildout that companies like Google, Microsoft, and a growing list of AI-focused neoclouds have been racing to complete. That same permissiveness is what let the interconnection queue balloon this far past the grid’s real capacity before state regulators stepped in — a large share of the 474 gigawatts in the queue almost certainly represents speculative reservations by developers hedging their bets across multiple sites, rather than genuine committed demand, which is exactly what an audit is designed to sort out. Rising electricity prices, which Texas residents and businesses have increasingly attributed to data center and cryptocurrency-mining growth, along with growing public opposition to unchecked AI infrastructure expansion, gave Abbott’s move real political cover.
A Preview of the Global AI Power Conversation
Texas’s freeze is a concrete, high-profile example of a tension that’s building in every jurisdiction courting AI infrastructure investment: the same qualities that attract data center capital — cheap land, permissive regulation, available power — are precisely what get strained fastest once that investment actually arrives at scale. Regulators everywhere are now facing the same basic choice Texas just made: continue approving projects on a first-come basis and hope the grid keeps up, or pause to verify which requests are real before committing scarce power capacity to speculative reservations. Expect more jurisdictions worldwide to follow Texas’s lead on mandatory disclosure and audit requirements as AI-driven power demand keeps outpacing grid planning cycles everywhere it lands.
What This Means for Philippine Founders
The Philippines is actively courting the exact kind of AI data center investment Texas is now struggling to manage — the Clark AI hub buildout and related national investment priorities depend on convincing global capital that this country can reliably supply power at scale. Texas’s experience is a useful, concrete cautionary example to watch closely, not because the Philippines is anywhere near Texas’s scale of demand yet, but because the underlying dynamic — permissive early approval processes attracting more interconnection requests than the grid can realistically support — is the same dynamic any government pursuing AI infrastructure investment will eventually face. For Philippine founders building AI products that depend on stable, competitively priced compute, it’s also a reminder that the AI supply chain’s real bottleneck increasingly isn’t chips or capital, but the unglamorous work of grid planning and power procurement — the jurisdictions that get that right, with real audited capacity rather than speculative queues, are the ones likely to offer the most reliable, least volatile compute pricing over the next several years.
Share this article