Updated and consolidated — October 10, 2026. This page now combines our two overlapping queue explainers. It uses the 2026 Berkeley Lab release covering the end of 2025 and separates proposed capacity, connection agreements and commercial operation.
Read the 2026 snapshot at its actual boundary
Berkeley Lab’s Queued Up release reports about 8,200 active US projects at the end of 2025, comprising 1,312 GW of generation and 749 GW of storage. The rounded components total 2,061 GW; the headline is more than 2,060 GW. The collected data cover regions representing about 98% of US installed generating capacity.
The same release identifies 549 GW with draft or executed interconnection agreements that were not yet operating. For projects built in 2025, the median time from request to commercial operation exceeded five years in regions with available timing data. These are proposed capacity and historical timing measures, not a forecast of next year’s available electricity.
Separate milestones before making a claim
| Status | What it tells you | What remains to be checked |
|---|---|---|
| Queue entry | A project requested an interconnection process. | Whether it will be built, its latest size and its route to service. |
| Draft or executed agreement | The process reached a stated contractual milestone. | Network upgrades, financing, construction and operating status. |
| Commercial operation | The project reached an operating milestone. | Measured generation, availability and subsequent performance. |
| Withdrawn | The request left the recorded queue. | The reason and whether a different project or request replaced it. |
A queue total is useful for understanding developer interest and workload. It is unsuitable as a direct addition to installed capacity. A headline that says “2,060 GW is coming online” skips the work between an application and an operating plant.
Use the withdrawal statistic with its cohort
For capacity entering queues between 2000 and 2020, the release reports 13% operating by the end of 2025, 75% withdrawn and 10% still active. Keep the entry cohort and observation cutoff in the sentence. Do not turn that historical completion share into a precise probability for a project joining in 2026; newer requests have different exposure time and conditions.
Berkeley Lab also says it is too early to measure the full impact of FERC Order 2023 reforms. Our earlier discussion overstated what could be concluded about reform results. A procedural change and an observed causal improvement require different evidence.
Avoid double counting a hybrid project
Consider a fictional application for 500 MW of generation and a 250 MW battery. Adding the two equipment ratings gives 750 MW, but does not establish 750 MW of simultaneous delivery or operation. If the battery’s stated duration is two hours at that rating, its nominal energy is 250 MW × 2 h = 500 MWh. MW and MWh answer different questions.
A dataset may represent several technologies or phases in separate rows. Before adding capacities, identify the project ID, resource type, point of interconnection and any duplicate or replaced requests. Use the dataset’s own field definitions; do not infer them from the column name alone.
A practical reading worksheet
| Question | Record before publishing |
|---|---|
| When is the snapshot? | End-of-year cutoff and dataset release date. |
| Which regions are covered? | The operator coverage and any missing regions or timing fields. |
| Which milestone? | Active request, agreement, operating or withdrawn. |
| Which unit? | Generation MW, storage MW or storage MWh; preserve each separately. |
| Which cohort? | Entry years and observation cutoff for a completion or delay statistic. |
| What is the inference? | Separate source-reported counts from your interpretation and avoid a causal claim without causal evidence. |
The useful conclusion from a large queue is that proposals and network integration need to be evaluated together. For an individual project, request the latest milestone documents and dependencies. A national backlog number cannot replace that project-specific check.
Sources and scope
- Berkeley Lab, Queued Up (2026 release; end-2025 data): Supports the counts, milestone definitions, historical cohort and limits on assessing reforms. Dataset credit: Lawrence Berkeley National Laboratory and Grid Tracker; data licensed CC BY 4.0.
- Creative Commons Attribution 4.0: Licence for the credited queue data; our worksheet and hypothetical hybrid example are original.
AI-assisted drafting. The linked material supports the stated facts; the examples and worksheets are Luminesca’s own editorial work. We did not conduct facility measurements, commercial project due diligence or model benchmarking for this article. Send a correction with the specific passage and supporting source.