ERCOT's update to the Texas Senate on July 29 contained a number that will be quoted for months: approximately 474 GW of large loads seeking interconnection to the grid, about 90% of them data centers, as of June 2026. For context, ERCOT's all-time peak demand record is 85,508 MW, set on August 10, 2023. The queue of proposed new load is more than five times the largest amount of electricity Texas has ever used at one moment.
That comparison is arresting, and it is also the wrong way to read the queue. The 474 GW is a list of requests. Developers file them to secure a place in line, often for the same capital looking at several sites, and they cost relatively little to submit compared with what it costs to build a campus. The useful questions are how fast the queue is growing, what share of it is likely to be built, and what ERCOT is doing to separate real projects from speculative ones.
How fast it is growing
The growth rate is the most striking feature. In its April update to the legislature, ERCOT reported about 410 GW of large load requests as of March 26, 2026, with about 87% from data centers. By the June 18 announcement that the Public Utility Commission had approved its new batch study process, the figure was more than 438,000 MW, nearly 89% from data centers. The July 29 presentation, with data as of June, put it at about 474 GW and roughly 90%. That is growth of more than 60 GW in about three months, nearly the size of the entire summer peak of some neighboring regions.
The data center share has risen at the same time. Texas has long attracted cryptocurrency mining, industrial electrification and hydrogen proposals, which were prominent in earlier versions of the large load queue. The share attributable to data centers has climbed steadily, and the rush is specifically for AI facilities seeking firm service, as ERCOT's April materials describe it.
The generation queue tells a different story
The same Senate presentation shows the supply side. As of June 30, 2026, ERCOT had 1,949 active generation interconnection requests totaling 462,785 MW. Solar accounted for 161,554 MW, energy storage for 170,945 MW, wind for 47,991 MW and gas for 77,607 MW. Solar and storage together make up about 72% of the generation queue. ERCOT's April materials noted a 271% increase in gas capacity in the queue since the Texas Energy Fund was created in 2023.
At first glance the two queues look balanced, about 463 GW of supply against 474 GW of demand. They are not comparable in that way. Solar and storage capacity does not deliver its nameplate at every hour, and a data center seeking firm service needs energy around the clock. Gas capacity in the queue, the resource most directly matched to firm round-the-clock demand, is 78 GW. Not all of it will be built either, given the multi-year backlog for gas turbines.
Why the queue cannot be studied the old way
Under the previous approach, ERCOT and the transmission service providers studied large load requests one at a time. As requests multiplied, each study was quickly made obsolete by the next, because the assumptions about what else would connect kept changing. The batch process approved on June 18 replaces that with a single study of all qualified projects of 75 MW and above. ERCOT says that 75 MW is enough electricity to power 18,750 homes during peak hours.
The timeline is set out in ERCOT's June explainer. Projects wanting to be part of Batch Zero, the first group, had to submit required technical studies and documentation by July 10, 2026. ERCOT planned to notify applicants of their classification in August, and to tell each Batch Zero customer in spring 2027 how much electricity can reliably be allocated to its project. A final transmission plan covering the batch is expected in fall 2027.
The July 29 presentation gives the first view of how much of the queue clears that initial screen. About 205 GW of large load is eligible for inclusion in Batch Zero based on existing studies, with final determinations due by August 7. In other words, less than half of the 474 GW queue had studies and model data good enough to enter the first batch. The presentation also notes tougher financial terms, including financial security of $50,000 per megawatt.
What share gets built
Nobody knows, and ERCOT is careful not to claim to. Its June release noted that not all interconnection requests result in built projects, while also saying that ERCOT data show the majority of applicants expect to be operational by 2030. Developer expectations are not the same as construction. The binding constraints are physical: transformers, turbines, transmission lines and the time it takes to build them.
The financial security requirement is the most direct tool for narrowing the gap between requests and reality. At $50,000 per megawatt, a 1 GW request would require $50 million of security. That is small relative to the cost of a hyperscale campus, but large enough to discourage developers who file in multiple locations to see which one moves first.
The policy context
Texas has moved faster than most states to set rules for large loads. Senate Bill 6, signed on June 20, 2025, set standards for loads above 75 MW, including disclosure of substantially similar requests elsewhere in Texas, a flat screening study fee of at least $100,000, and provisions for curtailment or backup generation deployment in emergencies. The batch process is ERCOT's implementation of the planning side. The political pressure is rising, too. Legislators are hearing from residential customers concerned about prices and reliability, and from developers who see Texas as the fastest place in the country to connect.
The 474 GW figure is best understood as a measure of interest rather than a forecast. It shows that Texas has become the default destination for AI infrastructure capital. Whether even a fifth of it reaches the grid in this decade will depend on the batch study results, the generation that gets built, and the security requirements that make developers commit.