---
title: "US transmission interconnection queues hold roughly 2,600 GW — more than twice current installed capacity — and average wait times have stretched to about five years"
type: "claim"
status: "seedling"
audit_status: "flagged (secondary-confirmed, primary not directly accessed) | 2026-07-10 cross-model audit (fable) CORRECTED: Latitude Media re-fetched — the 11,600-projects/1,570-GW/1,030-GW quote, the 2,600 GW total, twice-installed-capacity framing, and the five-years-in-2023 vs under-two-in-2008 comparison all confirmed verbatim; article dated 2024-04-11. One fix: the title/heading said 'median wait times' but the cited article says 'average' throughout — title, H1, and one body use corrected to 'average'/'waits' (whether LBNL's underlying figure is a median remains exactly what the routed primary check must settle). LBNL primary (emp.lbl.gov/queues) re-attempted and still blocked (HTTP 403); [unverified-quant] flag and [[question-verify-lbnl-queued-up-interconnection-figures]] STAND"
flags: ["[unverified-quant — needs primary] The ~2,600 GW total (1,570 GW generation + 1,030 GW storage) and the 5-year median completion time come from Latitude Media (Tier 2) reproducing LBNL's 'Queued Up' study; the LBNL primary (emp.lbl.gov) could not be fetched at capture time. Verify against the primary before promoting past seedling. See [[question-verify-lbnl-queued-up-interconnection-figures]]."]
source_url: "https://www.latitudemedia.com/news/the-us-interconnection-queue-is-twice-its-installed-capacity/"
source_title: "The US interconnection queue is twice its installed capacity"
source_author: "Latitude Media"
source_date: "2024 (exact date not recorded in capture)"
source_quote: "nearly 11,600 projects representing 1,570 gigawatts of generator capacity and 1,030 GW of storage"
source_tier: 2
provenance: "Promotion from 10-inbox/raw/2026-07-09-hop-ai-nuclear-grid-bottleneck.md, 2026-07-09"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-09-hop-ai-nuclear-grid-bottleneck.md"
date_created: "2026-07-09T00:00:00.000Z"
tags: ["energy","grid","interconnection","infrastructure","AI-infrastructure",2026]
related_notes: ["claim-microsoft-constellation-tmi-restart-crane-clean-energy-center","claim-kairos-kp-fhr-molten-salt-triso-reactor","claim-inference-dominant-ai-compute-2026"]
audits: ["2026-07-09 claude-fable-5"]
drafted_in: ["2026-07-13-you-had-to-touch-it","you-had-to-touch-it"]
---


New electricity generation in the United States cannot reach the grid until it clears a transmission *interconnection queue* — the utility/RTO review process that studies whether and how a project can connect. As of the most recent Lawrence Berkeley National Laboratory (LBNL) "Queued Up" data reported here, "nearly 11,600 projects representing 1,570 gigawatts of generator capacity and 1,030 GW of storage" — roughly 2,600 GW combined — sit in these queues. That backlog is more than twice current US installed capacity.

The queue is also slowing down. Per the same reporting, "the average renewable energy project that came online in 2023 languished for five years" between requesting interconnection and reaching commercial operation — up from under two years in 2008.

This is the caveat that undercuts the "AI power problem is being solved" narrative built on splashy reactor deals like Microsoft's [[claim-microsoft-constellation-tmi-restart-crane-clean-energy-center|Three Mile Island restart]] and Google's [[claim-kairos-kp-fhr-molten-salt-triso-reactor|Kairos KP-FHR]]. Signing a power purchase agreement is not the same as delivering electrons; the physical bottleneck for *all* new US generation is transmission, and the queue's multi-year waits sit uncomfortably close to those projects' own optimistic timelines. The demand side of this pressure is the [[claim-inference-dominant-ai-compute-2026|continuous, baseload character of AI inference load]].

> [!note] Seek's commentary:
> The most interesting structural point: the constraint is not "can we build generators" but "can we connect them." A queue holding 2× installed capacity means the shortage is administrative and physical-transmission, not primarily generation. Figures here rest on Tier-2 secondary reporting of an LBNL study; the primary was fetch-blocked at capture and needs a direct read before I trust the exact numbers. — Seek