The Texas queue is moving again, one company filing at a time

ERCOT paused its first batch of big power requests in August; this week the companies in it began reporting which of their gigawatts got a conditional place.

· 4 min read

What happened

In July the Texas grid operator ERCOT reported approximately 474 gigawatts of large-load connection requests, about 90 percent of them data centres. On 3 August, after the governor ordered a verification process, ERCOT paused the first batch, called Batch Zero. Galaxy's release explains what happened next: on 20 August the state utility commission let ERCOT hand out conditional classifications while it audits the data centre projects in its queue, a process Galaxy says ERCOT expects to take several months. The classifications are now appearing through the companies.

Galaxy said on 8 September that five of its Texas projects, approximately 4.2 gigawatts, received conditional classifications: 800 and 830 megawatts at its Helios campus as Base Load, capacity it says ERCOT approved in January 2026, and 700, 900 and 1,000 megawatts as Studied Load. It had submitted the 700 megawatt Caspian project as Base Load and received Studied Load. Core Scientific's filing of 10 September says 431 megawatts at Hunt was conditionally approved as Base Load, that 300 megawatts of existing load at Pecos was Base Load with a further 300 megawatts as Studied Load, and that it "will not receive the final allocation and load ramp until the conclusion of the Batch Zero process". Soluna said on 8 September that its full 166 megawatt Kati campus was conditionally classified as Base Load.

Why it matters

The labels decide who is exposed to the study and who is not. Galaxy's release says ERCOT's process "groups qualifying projects into a single, system-wide study that allocates available transmission capacity", that "Base Load projects are modeled in the study and are not subject to further capacity allocation, while Studied Load projects are subject to capacity allocation as part of the study". Soluna's filing describes Base Load the same way: a project's connection capacity "is preserved on the basis of prior study work" and it "is not subject to further reliability evaluation or megawatt allocation within the Batch Zero study". So Galaxy's two Helios projects and Soluna's Kati campus have their capacity preserved, conditionally, while Galaxy's other 2.6 gigawatts and Core Scientific's second 300 megawatts at Pecos go into the allocation.

A worked example. Galaxy's 4.2 gigawatts is less than 1 percent of the approximately 474 gigawatts ERCOT reported in July, our arithmetic. Soluna's filing cites a different ERCOT figure, more than 438,000 megawatts as of mid-2026; we do not know why the two figures differ. The three companies in this story account for a little over 5 gigawatts between them.

The US Energy Information Administration reported on 3 September that hourly load in ERCOT averaged a record 74.5 gigawatts for the week ending 22 August, and noted that ERCOT is auditing data centre proposals following the governor's moratorium.

From FERC's order to the first conditional labels we can see, on the site's own numbers.
  1. 6 RTOs: FERC gives grid operators 60 days on large-load rules (stated: the federal push to connect large loads faster, and the queue it met in Texas)
  2. 474 GW: ERCOT's large-load queue about 90% data centres (stated: the queue was paused for verification)
  3. Batch Zero: Texas pauses first batch after governor's order (our reading: the labels now visible through company filings)
  4. approx 4.2 GW: Galaxy's conditional labels five projects, two Base Load

What would change it, and when

ERCOT's audit, which Galaxy says is expected to take several months, is what turns conditional into final. Core Scientific says its allocation and load ramp come only at the end of the process. ERCOT's own news releases carry nothing on the classifications; if it publishes an aggregate list, that document replaces the company filings as the source.

What we do not know

We do not know how many projects across the whole batch got each label, because ERCOT has not published a total. We do not know how much of the Studied Load capacity will be allocated when the study ends, or when the audit ends. And conditional means conditional: every classification in this story can change.

What this changes for you

If you pay a Texas electricity bill, this is the machinery deciding how much new data centre demand joins the grid you draw from, which just set a weekly load record. Nothing on your bill moves this week.

Sources

Everything above is written from these. Each line says what that document proves.

  1. prnewswire.com: Galaxy's release, 8 September: the five projects, approximately 4.2 GW, the labels and their definitions, Helios approval in January 2026, Caspian submitted as Base Load, the 20 August commission exceptions and "several months".
  2. sec.gov: Core Scientific's 8-K, 10 September: Hunt 431 MW Base Load, Pecos 300 MW Base and 300 MW Studied, no final allocation until the process ends.
  3. sec.gov: Soluna's 8-K exhibit, 8 September: the 166 MW Kati campus as Base Load, its definition of Base Load, and its citation of ERCOT's 438,000 MW queue figure.
  4. eia.gov: EIA, 3 September: the record 74.5 GW weekly average load and the note that ERCOT is auditing data centre proposals.
  5. ercot.com: ERCOT's own approximately 474 GW figure and the 90 percent share, already on the site.

Strata, the whole AI stack, explained simply. Every number carries a source and a confidence label.

Today · The Stack · Latest issue · Archive · Glossary · Break the Chain · Focus

Privacy · Terms · Corrections · Report an error

Also from Dheeco: Dheeco · AP Fact Check

Published by dheeco.com · © 2026 Dheeco