The ERCOT Queue Cannot Distinguish Captive Gas from Speculation
For anyone screening ERCOT gas by interconnection status: a dedicated plant can be commercially committed to a named customer and still appear in the queue as unsigned and in study. Project Kilby shows why captive generation needs a separate diligence framework.
For hyperscalers · For developers · For infra funds · ercot · gas · data-centers · interconnection · behind-the-meter
Kris Narayanan · Tafel Power · July 13, 2026 · 3 min read
ERCOT's interconnection queue is the clearest public read on gas supply, and a signed interconnection agreement is its best public maturity signal, but that signal has a blind spot.
What the queue shows
Search the June 2026 GIS Report for Energy Forge One and Project Kilby is there, four blocks near Pecos in West Texas: a 371 MW simple-cycle unit due 2027, a 560 MW battery due 2028, a 950 MW simple-cycle block due 2028, and a 1,056 MW combined-cycle block due 2029. About 2.38 GW of gas plus 560 MW of storage; Chevron describes the delivered project at about 2.67 GW. Every block carries the same status: study underway, no interconnection agreement. On the queue's own best signal, Kilby is indistinguishable from the roughly 55 GW of unsigned gas that status-based screens heavily discount.
What the queue does not show
None of the off-queue evidence supporting Kilby's commercial maturity is visible in the queue. In June 2026 Chevron and Microsoft signed a 20-year agreement for the plant, to be built by Chevron's Energy Forge One and dedicated to one Microsoft AI campus. Chevron targets a final investment decision by the end of 2026 and first power in 2028. Kilby's TCEQ air-permit application (181895), at roughly 2,595 MW, is advancing through the state's permitting process. The turbines are named: GE Vernova frames, with Caterpillar's Solar Turbines for the balance. The project has a named customer, a long-term agreement, an active permit, and identified equipment, and none of it is visible in the interconnection-agreement field.
Why the distinction matters
A captive project can advance commercially through a private customer agreement while remaining unsigned in the interconnection queue. A completion model based only on queue status, like the one we use to weight the gas queue, may underweight a project like Kilby when applied mechanically at the individual-project level. That model is still appropriate for ordinary merchant projects, where interconnection status is the best available signal. It is insufficient on its own for captive generation.
The broader signal inside the GIS
Kilby is not the only data-center-linked gas in the unsigned queue. Two other entries are filed as "Bullock Data Center, LLC" and "Liberty Data Center I, LLC," each with 1.4 GW of gas, each unsigned. Their commercial backing is not publicly established, so we draw no conclusion about their maturity. But their names show that data-center-linked gas is already embedded in the unsigned queue, where a purely status-based screen would treat it as speculative.
The practical takeaway
Screen merchant gas through interconnection status; it is the best public signal for projects seeking a broad market. Screen captive or dedicated gas through customer commitment, permitting, and equipment evidence instead. The queue can tell you which merchant projects are maturing, but not which captive plants are already committed.
Methodology
Kilby's queue detail is from the ERCOT June 2026 GIS Report (Energy Forge One LLC, four blocks totaling 2,937 MW including a 560 MW battery), reconciled against an archived snapshot of that source; the Bullock and Liberty entries are from the same report. Deal terms, timeline, and turbine suppliers are from Chevron's and Microsoft's announcements. The permit reference is TCEQ air-permit application 181895 (PSDTX1684, GHGPSDTX260), described in the record as advancing through TCEQ permitting. The completion-rate model referenced is the gas-queue weighting method in How Much of ERCOT's 69.7 GW Gas Queue Is Credible?. This piece establishes generation commitment and project structure, not firm fuel supply. Figures reflect the June 2026 snapshot and may have changed since.
All data compiled by Tafel Power from public sources. Framing informed by the firm's transaction advisory work in ERCOT and cross-ISO markets.
For advisory work involving power transactions, large-load strategy, infrastructure investment, or cross-market diligence: kris@tafelpower.com
More from Insights
If You Can't Get a Turbine Until 2029, What Powers the Load?
Firm power · Turbines · Time to power
July 12, 2026 · 4 min read
For hyperscalers, developers, and infra funds: the firm-power options that can actually connect before the gas-turbine backlog clears, ranked by how fast they deliver.
What Actually Powers a 24/7 Data Center?
Renewables · Firm power · Storage
July 12, 2026 · 8 min read
For hyperscalers and infrastructure investors: the interconnection queue contains nearly 2 TW of proposed capacity, most of it solar, wind, and storage. The scarce product for a 24/7 load is firm capacity, not annual clean energy.