The benchmark for openness. The Texas grid operator (ERCOT) runs an energy-only market, meaning generators are paid only for the power they deliver, and it lets a large load connect quickly in exchange for accepting occasional cutbacks when the grid is tight. That is the fastest route to power in the United States, cheap gas lets a developer supply its own electricity, and the data-center build is already at scale. Carbon is the only weak spot, and it carries little weight in the default scoring.
See live grid data →Why: Energy-only connect-and-manage (connect first, ease congestion later) keeps ERCOT the fastest large-load path in the US; the regulatory watch is readiness-tested queue management (a 75 MW threshold and posted security), not closure.
What would change the read: Restrictive new large-load interconnection rules following a major reliability event.
For new supply: Energy-only connect-and-manage is the fastest generation build path in the US, which is why supply and demand are racing each other here.
Beyond the grid: Water and local consent are the main non-grid frictions: a University of Texas poll found 56 percent of Texas voters, and 62 percent in rural areas, oppose a data centre in their community, and the state still does not require most centres to report water use even as it faces scarcity. Texas Tribune UT Austin
Open and fast. The Texas grid operator (ERCOT) lets a large new load connect quickly and then manages any constraints, on the condition that it can be dialled back when the grid is tight, rather than making it wait years for firm supply. This is one of the quickest paths to power anywhere.
Tight at peak but workable: summer scarcity is real, but cheap on-site gas (behind-the-meter) floors the practical availability for a developer who can self-supply.
Cheap power and the cheapest self-supply: low wholesale plus abundant gas make Texas one of the lowest recurring-cost markets for a large load.
Proven momentum: more than 1 GW of data-center capacity and widely called 'the next Virginia'; ~233 GW of large-load interconnection requests, over 70% data centers.
Carbon is the weak pillar (~310 g), gas-heavy; immaterial in the default lens, but it pulls ERCOT toward Tier 2 under a carbon-weighted buyer.
Extreme-cold fragility was demonstrated in Winter Storm Uri (Feb 2021), when ERCOT ordered about 20,000 MW of rolling blackouts, the largest manual load shed in US history, with outages lasting days; freezing and fuel issues drove roughly 76% of unit failures. Post-Uri winterization rules have improved subsequent winter performance but the tail risk remains.
What this means: A large new load faces occasional exposure here. Monitor conditions and keep a contingency plan.
event-based · FERC/NERC Feb 2021 cold weather final report
Average customer interruption (SAIDI) of 97.27 minutes per year in 2024, excluding major event days as filed.
The market figure above is an average across these territories. Your site sits in one of them, not in the average. Each row below is that utility’s own filed record, which is the evidence a Project Case can actually rest on.
| Serving utility | SAIDI min/yr | Customers | Major-event method |
|---|---|---|---|
| Oncor Electric Delivery Company LLC (TX) | 67.28 | 4,013,062 | own method |
| CenterPoint Energy (TX) | 150.09 | 2,847,806 | IEEE 1366 |
| City of San Antonio - (TX) (TX) | 61.8 | 938,905 | IEEE 1366 |
| AEP Texas Central Company (TX) | 149.6 | 913,164 | IEEE 1366 |
| Austin Energy (TX) | 76.03 | 557,039 | IEEE 1366 |
| Pedernales Electric Coop, Inc (TX) | 57.44 | 425,109 | IEEE 1366 |
| Denton County Elec Coop, Inc (TX) | 26.9 | 319,523 | IEEE 1366 |
| Texas-New Mexico Power Co (TX) | 118.37 | 268,934 | own method |
| AEP Texas North Company (TX) | 102.6 | 199,119 | IEEE 1366 |
Filed to EIA Form 861 for 2024, excluding major event days. Territory-wide averages: see the layer note above for what they do not establish.
NERC's 2024 Long-Term Reliability Assessment rates ERCOT elevated-risk: about 20 GW of new large loads plus a more variable, less dispatchable resource mix raise the risk of long-term shortfalls.
What this means: A large new load should expect genuine exposure to forced power cuts (curtailment), price spikes, and delays in getting connected, and should design in backup power or firm supply contracts.
assessed · NERC 2024 Long-Term Reliability Assessment
ERCOT's open-and-fast access is explicitly conditioned on a large load accepting curtailment when the grid is tight, so the fast route to power is a non-firm route. Firm service is a different, slower product.
Utility feeds, UPS, storage, on-site generation, islanding, black start and the residual energy you still expect to lose are properties of a specific site design, not of a market. This layer only exists once there is a site, and we never assert it for you.
The average customer loses about 97.27 minutes of power per year (2024). This is the standard regulator outage measure (SAIDI); lower means a more reliable grid.
408 GW
30 months
The full queue analysis for this market, which covers the share of projects that give up and leave the queue (the withdrawal rate), how much of the queue is data centers, and the sources, is available through /api/v1/queue/ercot with a paid key. These are generation and storage queue figures: they show interconnection-system congestion, not a typical large-load connection wait.
Elevated
no capacity market (energy-only)
ongoing
ERCOT is an energy-only market with no capacity auction, so the forward cost of firm capacity is signaled through scarcity pricing (the operating-reserve demand curve) and forward energy rather than a clearing price. NERC rates ERCOT high-risk for forward adequacy from about 2029.
Source: NERC 2025 Long-Term Reliability Assessment (Jan 2026). A derived indicator: we publish the public clearing price and our read, never paywalled exchange data.
$50 to $95 per MWh
Midpoint about $68 per MWh
The all-in band sits well below Virginia's because there is no capacity charge to carry, and a load that can curtail at the annual system peaks pays materially less than one that runs flat through them. Scarcity pricing during grid stress is the main upside risk to the number.
This market's band is free here; the ranked cross-market table with every market's precise midpoint is the Analyst Desk. See the ranked view.
Basis: triangulated from public sources, in US dollars per MWh delivered to a large high-load-factor load, excluding refundable taxes. Sources: ERCOT wholesale price outlook (EIA) · ERCOT market prices.
High
3 of 4 data layers measured
Full
3 of 3 dimensions: reliability, connection, cost
June 2026
per-layer vintages below
| Layer | Method | Vintage | Freshness | Source |
|---|---|---|---|---|
| PGIQ Rating opinion | opinion | June 2026 | Current | PGIQ methodology |
| Reliability | measured | 2024 | Fresh | EIA Form 861, Schedules 3B and 3C (annual electric power industry report) (two major-event methods, IEEE; respondent-defined, customer-weighted blend of 9 utilities, regional context only) |
| Forward cost | measured | ongoing | Fresh | NERC 2025 Long-Term Reliability Assessment (Jan 2026) |
| Interconnection queue | measured | end-2025 (LBNL 2026) | Fresh | LBNL Queued Up 2026 |
| Realized cost | triangulated | 2026 | Fresh | ERCOT wholesale price outlook (EIA) |
Freshness reflects each layer's refresh cadence: reliability annual to three-yearly, capacity-auction and queue yearly, tariffs every year or two. The full cross-market freshness dashboard is at /data-quality.
The real-world events that test our Tier 2 call for ERCOT (Texas): 3 support it. Each is dated and cited, so you can check the rating against what is actually happening on the ground.
The PUCT approved (18 Jun 2026) ERCOT's Batch Zero framework, which studies readiness-screened large loads in batches with a Jul 2026 cutoff for the 2028-2032 study. Formal regulatory sign-off on the readiness-tested queue.
Source: ERCOT / PUCT ↗The large-load queue kept climbing through the first half of 2026, and ERCOT introduced evidence-based readiness screens (site control, permits, financing, equipment orders) to separate committed builds from speculative ones. A market disciplining a queue this large is one where large loads genuinely want to build, supporting the rating.
Source: Utility Dive ↗Over 70% from data centers; ERCOT models ~138 GW of large loads by 2030, the demand side now racing as fast as supply. Demand at this scale choosing Texas confirms the cheap, fast-to-connect profile behind the strong rating.
Source: ERCOT ↗How ERCOT (Texas)'s reliability-adjusted score and tier have moved, month by month. Part of your Analyst Desk.
This page is our full read on ERCOT (Texas) alone, and it is free. To decide where a large load should actually go, you need every market side by side, ranked on recurring cost and time-to-connect, with the history and alerts when your shortlist moves. That is the Analyst Desk.
With your Analyst Desk, compare ERCOT (Texas) against every market side by side, ranked on recurring cost and time-to-connect, with the full history and shortlist alerts.