Methodology

What each dataset contains

242 readings record one measurement for a named operator, resource and date. The 49 directory listings describe operators, data providers, software and study firms. Each dataset has its own download: readings JSON / CSV, and listings JSON / CSV.

The observation date describes the snapshot, completion cohort or reported event. It retains the source’s precision: a year or month is not converted into an invented day. The accessed date says when the source was checked. Readings are dated snapshots, not a live feed.

Resource MW and hybrid projects

ISO/RTO and US active capacity comes from the regional/resource summary in LBNL and GridTracker’s Queued Up 2026. GW are multiplied by 1,000 to store MW. Wind combines onshore and offshore. Storage combines standalone and hybrid storage, including the summary’s imputed hybrid MW. The year-end 2024 readings use the historical series in the 2026 workbook, so they can differ from an earlier edition.

Major utility readings and ISO nuclear readings sum the project sheet’s reported component MW for active requests in the named entity group. Battery and other-storage components count as storage. Missing component MW stay missing and are excluded from the reported sum; the note states how many matching components lack MW. No imputation is added. A zero means no reported matching MW in that sample. Utility footprints are the workbook entity groups, including grouped Duke companies and Dominion’s non-ISO sample.

Resource components are not net export limits. A hybrid project can contribute to both solar and storage, but request counts count its workbook row once. The regional total includes resources beyond the displayed solar, storage, wind, gas and nuclear columns. Displayed MW round to whole MW; stored calculations round to three decimal places.

Completion and withdrawal denominators

Outcome readings divide operational or withdrawn MW by all MW in the source’s requests-submitted-2000–2020 cohort, observed at end-2025. The denominator includes active, operational, withdrawn and suspended capacity. Sample counts are the corresponding request counts, not the weights used in the percentage. The source’s regional completion sheet covers operators that supply comprehensive status information. These historical proportions are not forecasts for a new application.

Measured wait times

Waits are the workbook’s median months from the interconnection request to actual commercial operation, grouped by the year projects reached operation. This is elapsed development time, including activity after interconnection studies. It is not the median age of today’s queue or a guaranteed study duration. Proposed CODs are never substituted for actual CODs.

The all-region annual table and the regional table have slightly different sample counts in some years. Each reading retains its own sheet’s sample and locator. An absent regional/year reading stays absent; the snapshot shows a dash when no measured 2025 regional cohort is supplied.

Large-load requests and stages

Generation queue MW exclude load requests. Large-load readings come from operator reports and utility filings: study requests, delivery-point requests, conditional classifications and binding contracts retain those labels. Operator and utility footprints can overlap; load stages and generations of a queue can also overlap. No national load total or inferred data-center share is calculated.

ERCOT’s September Batch Zero report gives a studied-load total of 127.9 GW in its summary and 125.4 GW on its chart. Both sourced readings are retained with conflict notes. They describe one population and must not be added. Base load also includes some energized facilities. Exclusion from Batch Zero is not treated as withdrawal from every future process.

Reform events and scheduled milestones

Each milestone separates its observation date from the event or scheduled target date. A filing is described as filed, partial compliance as partial, and a future schedule as scheduled. Historical orders are not presented as a complete statement of current legal status. Primary orders and operator announcements are linked on each record.

Workbook retrieval and reproducibility

The official workbook download challenged automated shell retrieval. Extraction used a public mirror pinned to a commit. The mirror matches the official attachment’s listed size (14.85 MB), and its national totals match every headline on LBNL’s Queued Up page: over 2,060 GW active, 773 GW solar, 749 GW storage, 220 GW wind, 253 GW gas and a median above 5 years to operation for 2025 builds. Both URLs appear on each affected reading.

Each reading includes a workbook sheet/cell reference or a report locator and the calculation applied. The repository retains a compact CSV extraction ledger, the import script, the workbook hash and the builder for operator readings. Rebuilding the JSON from the checked ledger does not require a network connection.

Directory verification

Directory fields use first-party queue pages, tariffs, product pages and documentation. A missing fact stays blank. Vendor claims are labelled where their scope matters. Fees and security reflect the cited source on the access date. Acquired companies sit under the surviving provider where the existing directory records the change.

Attribution and reuse

LBNL-derived readings are adapted from Lawrence Berkeley National Laboratory and GridTracker, Queued Up 2026, under CC BY 4.0. Adaptations include unit conversion, component aggregation, cohort percentages and rounding. Retain that attribution when reusing these readings. Operator and regulatory readings retain their own cited authorship; source documents are not relicensed by this site.

Corrections and placement

Send corrections with a public source to hello@interconnectionqueues.com. A substantive correction updates the reading and access date. Paid featured placement is labelled and does not change measurements or sorting.