Texas has moved from an expansion-led data-center market to an audit- and compliance-led regime. Electricity access—not land, capital, or demand for computing capacity—is becoming the gating factor for Meta Platforms’ artificial-intelligence infrastructure strategy. Texas remains attractive because of its large energy market, available land, and technology ecosystem 26. JLL reportedly expects the state to become the world’s largest data-center market 13. But the scale of proposed load has forced direct state intervention. Governor Greg Abbott ordered the Public Utility Commission of Texas and ERCOT to audit data-center projects and temporarily halt approvals for new grid connections 3,5,7,17,18,21.
The queue explains the intervention. Texas has received approximately 474 GW of electricity interconnection requests—roughly five times peak demand—with about 90% attributed to data centers 6,8. ERCOT paused new approvals while reviewing the queue 33. Estimates suggest that 49.8 GW of queued infrastructure could be delayed 29. The headline figure is not firm demand. The queue reportedly contains speculative, duplicate, unfunded, and otherwise non-executable projects 32. Probability-adjusted and executable megawatts are the relevant investment metrics 32.
The number still matters. It has made grid interconnection a governance point and a strategic bottleneck for AI infrastructure in Texas 4,18. For Meta, the exposure is immediate. Meta and BlackRock announced a $14 billion transaction to develop a 1 GW Texas data center 10,27. The facility is expected to require 1 GW—approximately 40% of all other El Paso Electric customer consumption during a hot summer afternoon 9. Meta may request another 400 MW from El Paso Electric if necessary 9. The project demonstrates Meta’s capital strength. It also demonstrates the political and infrastructure risk created by hyperscale concentration.
Key Insights
The regulatory pause is an execution risk
Texas has halted or paused approvals for new data centers seeking grid connections. Abbott’s directive is supported by three sources 5,7, and the audit order by three sources 3,17,18. A separate audit claim has two sources 4,6,25. The stated purpose is to determine whether new facilities can be integrated without overburdening existing infrastructure and to protect public safety and grid reliability 2,14. State officials have characterized the intervention as a capacity and infrastructure measure, not political pressure 2.
The suspension remains pending an assessment of outstanding requests and a comprehensive audit of projects seeking approval 5,21,22. This is not routine administration. It is state oversight of utility interconnections 5,18. The delayed ERCOT Batch Zero transmission-planning study shows that the intervention can affect network planning as well as individual data-center permits 6.
Queue rationalization could ultimately improve the market. Removing duplicative, speculative, and non-firm requests would allow credible projects to receive clearer network-study results 29,32. The interim cost is uncertainty. Approval timelines are unclear 5. Projects face postponement 5, and regulatory review can produce delays, modifications, or outright denials 26.
For Meta, the correct response is a wider execution range. Construction schedules, energization dates, and project returns must account for permitting, interconnection, construction, and expansion risk 22,31. Tighter rules can increase lead times or constrain growth 12. Grid delays may force cloud, AI, GPU, and data-center companies to reassess site selection and development timelines 2. Capital could move toward regions or countries with more available electricity and faster interconnection 18. Texas exposure may become less attractive relative to regions with more reliable supply 2.
The new rules reward contracted, self-funded projects
Texas has introduced a statewide compliance framework requiring more information before PUCT or ERCOT authorizes connections 24,26. Applicants must disclose electricity consumption, water usage, community impacts, ownership and controlling interests, and public funding 24,26. The stated purpose is transparency and operational accountability 26, while limiting effects on the grid, water supplies, and nearby communities 24,26. Abbott’s guidelines direct regulators to scrutinize power use, water consumption, community effects, and developer incentives 11.
The framework shifts review from isolated interconnection requests to cumulative resource impacts. It also raises the burden of proof. Queue reform reportedly requires evidence of capital, site control, equipment plans, customer commitment, and adherence to responsible-growth standards 32. Projects that fail the audit or ERCOT requirements may be denied connection 6,26. Already approved projects may still face scrutiny 25. Meta’s stated agreement to comply with Abbott’s standards on grid demand, water use, and community impact 11 reduces the risk of a direct compliance dispute. It does not eliminate the risk of restudy, modification, or additional infrastructure obligations.
Cost allocation is the central economic issue. Abbott’s guidelines reportedly require data centers to pay for their own electric infrastructure 28. The objective is to prevent large loads from shifting upgrade costs to other customers 15. The standards may reduce ratepayer cost transfers 26 and support grid stability 26. They will also increase compliance, disclosure, infrastructure, water-management, cooling, and on-site-generation costs 26.
El Paso Electric is seeking approval for a special data-center rate structure 28 that includes a credit guarantee, ensuring existing customers do not bear infrastructure costs built specifically for a data center 28. The utility’s position is direct: new large-load customers should bear the costs required to serve them 28. The wider policy debate remains focused on cost recovery and grid-upgrade allocation 12.
This framework favors financially strong hyperscalers with long-term power contracts, substantial balance sheets, and the ability to fund dedicated generation and transmission. It penalizes speculative developers, merchant plants, and projects dependent on socialized infrastructure. Current ERCOT prices are reportedly insufficient to support broad merchant generation without bilateral hyperscaler contracts, capacity-market support, or regulated cost recovery 32. Meta’s scale is therefore an advantage only if it accepts greater upfront capital intensity and explicit responsibility for power infrastructure.
El Paso creates jurisdictional risk, not a clean exemption
El Paso is a special case. ERCOT covers most of Texas but excludes El Paso 28. Abbott’s directives specifically reference ERCOT and PUCT without expressly naming El Paso Electric or other utilities outside ERCOT 28. The applicability of the state requirements outside ERCOT remains unresolved 26. Enforcement mechanisms and the durability of the guidelines are also legal and political uncertainties 28.
Abbott has argued that PUCT jurisdiction extends beyond the ERCOT grid and that the standards apply both within ERCOT and to PUCT-regulated projects 26. The City of El Paso stated that a binding PUCT requirement would apply to the Wurldwide El Paso project through the state process 26. The result is tension, not exemption. Meta may avoid precisely the same ERCOT queue mechanics as projects in Central or West Texas, but the project remains exposed to PUCT approval and state-level standards.
El Paso Electric is seeking approval to construct the McCloud plant, a set of gas-fired generators adjacent to Meta’s data center 28. The proposed $500 million plant is designed to provide 225 MW of continuous power 9. The utility expects to build additional generation and distribution infrastructure to meet demand 28, while Meta’s planned electricity supply would initially add 250 MW to the utility’s system 28. The proposed plant requires PUCT approval 9. Private or dedicated supply does not remove permitting risk.
The scale mismatch is material. Meta’s 1 GW requirement is large relative to El Paso Electric’s existing customer base, and S&P analysts indicated that Meta could seek another 400 MW 9. A private-grid model has also been reported for Amazon’s West Texas AI project 23. Hyperscalers are therefore moving toward dedicated generation and power procurement. That architecture can reduce dependence on the bulk grid. It also adds environmental permitting, emissions compliance, fuel-supply, construction, and financing risks. For Meta, adjacent generation may become a prerequisite for timely energization rather than an optional optimization.
Power and water are now part of the license to operate
Reliability is the primary concern. Integrating high-load facilities into the ERCOT and Texas systems is described as a reliability and safety issue 2. Grid failure is identified as a principal risk of expansion 2,8. Rapid data-center growth has raised concerns about long-term ERCOT capacity and reliability 16, while state authorities identify grid reliability and water availability as potential bottlenecks 21. Texas’s weather-adjusted annual peak load grew approximately 3% from 2022 through June 2026 20. Other estimates attribute three percentage points of load growth to non-data-center sources and two percentage points to data centers 32. The evidence supports meaningful load growth, not the claim that all recent demand expansion is data-center driven.
Water is the parallel constraint. Texas has mandated a comprehensive audit of data-center power and water usage 5, with scrutiny expected to include cooling efficiency and overall resource consumption 5. Water supply pressure is a state concern 5,26. Residents and officials in Central Texas have raised concerns about local water supplies and other resources 30. The proposed San Antonio moratorium likewise reflects concern over electricity and water demand 39. Meta’s compliance with the new standards therefore affects permitting, community acceptance, and its broader sustainability position.
The political perimeter is expanding. Texas has seen backlash over grid reliability, water use, neighborhood disruption, taxpayer subsidies, and cost allocation 28. Data centers can generate noise, lighting, traffic, and emergency-response concerns 26. Opposition is reported across the political spectrum 13, and approximately 75 U.S. data-center projects face local opposition 33. The concern is national. A Reuters/Ipsos poll found that 77% of Americans worry data-center electricity demand could increase household utility bills 1,33. Meta’s largest projects will attract the most scrutiny where local customers believe a hyperscaler receives incentives while externalizing infrastructure costs.
Queue size is not realized demand
The 474 GW queue is a signal of prospective AI demand, not a dependable forecast of realized load. It is approximately five times Texas peak demand 6, and roughly 90% of requests originate from data centers 6. The queue is reportedly being rationalized to remove projects that distort planning 32. Credible developers are delayed by the oversized feasibility process 32. BloombergNEF estimates that the review could place roughly 20% of the U.S. data-center development pipeline at risk 31 and cost developers as much as $15 billion through delays 29. Those figures are scenario estimates, not confirmed losses.
Power-market economics also contradict the simplest load-growth narrative. ERCOT and PJM reportedly set new peak-load records in 2026 20. Yet batteries, renewable additions, and reserves are suppressing near-term scarcity and forward-price tightening 32. Wind projects can experience curtailment 36, potentially creating low-cost energy for behind-the-meter loads 36 and deployment opportunities for companies such as Soluna 37. High prospective demand therefore does not guarantee broad merchant-generation profitability. Investors must distinguish contracted, probability-adjusted load from speculative queue capacity—and demand growth from realized pricing power.
The same discipline applies to adjacent project announcements. Tecogen reports a demonstration pipeline exceeding 8 GW 35, consisting of 12 demonstrations representing more than 8 GW of potential or operating data-center capacity 35. Its natural-gas chillers are positioned to reduce grid-supplied data-center demand by 25% to 35% 35. Other announced projects include Amazon’s 7.65 GW gas-fired Pecos County plant 6, Tract’s more than 2 GW Reno campuses—nearly one-third of NV Energy’s current generating capacity 38—a potential 17 GW FRMI campus 34, a 1.2 GW AirTrunk campus 19, and a 1 GW Texas project 10. These projects demonstrate the industry’s ambition. They also explain why regulators now demand proof of financing, customers, equipment, and site control.
Implications for Meta Platforms
Meta should treat Texas as a high-upside but policy-sensitive deployment market. The proposed 1 GW El Paso facility and $14 billion Meta-BlackRock transaction show that the company is willing to commit substantial capital to AI infrastructure 10. Dedicated generation, distribution, and special-rate arrangements could give Meta greater control over power availability than smaller or less-capitalized competitors. If fully energized and utilized, the campus could also support lower unit costs in computing infrastructure.
The same scale creates concentration risk. A 1 GW facility representing approximately 40% of El Paso Electric’s non-Meta hot-afternoon load 9 can become a political focal point, especially when the utility must seek PUCT approval for adjacent generation 9,28. Meta’s potential 400 MW expansion would deepen that exposure. The strategic trade-off is clear: campus concentration can improve operating efficiency, while geographic diversification improves resilience. Texas regional concentration risk has already been identified 5. Limited electricity and water availability could affect operating costs, project economics, pricing power, and future site selection 5,22.
The regulatory regime may favor Meta relative to weaker developers while raising the absolute cost of expansion. Meta can likely meet disclosure, capital, customer-commitment, and infrastructure-funding requirements more readily than speculative projects. Its announced compliance commitment 11 may strengthen its position with regulators and preserve its social license. The guidelines can still require project modifications, impose dedicated-infrastructure costs, expose incentive arrangements to scrutiny 26, and create uncertainty for projects approved under prior local agreements 26. Community-impact assessments may increase the likelihood of opposition or redesign 24.
The valuation question is not whether AI demand exists. It is how quickly Meta can convert that demand into energized, revenue-generating capacity. The Texas pause introduces a temporary regulatory constraint 21. The legal status, enforceability, political durability, and geographic scope of the guidelines remain uncertain 28. Tail risks include weak ERCOT economics, customer deferrals, and a mismatch between infrastructure investment and realized load 32. The broader risk set includes failed interconnections, revised network costs, local opposition, water scarcity, financing shortfalls, equipment inflation, and hyperscaler redesign or phasing 32. Meta’s AI capital-expenditure and data-center ramp models must therefore use scenarios rather than treating announced megawatts as near-term capacity.
The competitive effect could still favor Meta over time. A stricter Texas framework may remove speculative queue participants and reduce congestion in network studies, improving visibility for well-funded projects 29,32. Requiring large loads to fund their own infrastructure may prevent ratepayer backlash from becoming an outright prohibition. Dedicated generation and behind-the-meter solutions can reduce dependence on ERCOT. Meta’s task is to prove that its project adds reliable, economically funded capacity without imposing disproportionate costs on residents or water systems.
What to Monitor
The decisive indicators are the outcome of the statewide audit; the criteria for removing or retaining projects in the 474 GW queue; PUCT treatment of El Paso Electric’s special rate and McCloud plant; the treatment of projects outside ERCOT; and evidence of Meta’s power-procurement and water-management commitments. The market should also monitor whether ERCOT’s delayed Batch Zero study resumes and whether the surviving queue produces actionable transmission results 32.
A favorable outcome would make the pause a cleansing event. Speculative projects would be removed, credible hyperscalers would gain clearer interconnection visibility, and Meta’s balance sheet would become a moat. An adverse outcome would extend energization delays, increase capital intensity, and redirect Meta’s incremental AI investment to other regions.
The bottom line: Texas is no longer selling access to land and cheap power alone. It is selling access to a controlled infrastructure system. Control is the prize. Meta has the capital to compete for it, but the company must underwrite Texas power as a regulated, politically exposed asset—not as a routine construction input.