2.6 terawatts of proposed solar, wind and battery projects are waiting in US interconnection queues, the line to plug into the grid. That is roughly twice the country's installed generating capacity, and the wait for a study result already runs past four years in most regions.
AI-driven study software is compressing months of engineering into hours. Pearl Street's SUGAR cuts power-flow analysis time by up to 200x and is contracted to automate interconnection studies across more than 500 GW of capacity.
Consolidation is moving fast. Enverus bought Pearl Street, ISO New England picked GridUnity, and the U.S. Department of Energy put $30 million behind its AI4IX program, turning a back-office bottleneck into a genuine software market.
The queue is the process a new power plant must survive before it can sell electricity. Every solar farm, wind plant and storage unit needs an interconnection study: a power-flow analysis that decides which grid upgrades are required and who pays. Each study depends on every other proposal in the same cluster, so operators process them in groups, and the paperwork has become as binding as the physics.
TIMELINE: The US interconnection queue
─────────────────────────────────────────────────────────────
2021 ─────── 2023 ───────── 2024 ─────── 2025 ─────── 2026
📈 ⚖️ 🏛️ 🤝 ◉ NOW 🔥 NEXT
1.4 TW FERC Order DOE AI4IX Enverus ~2.6 TW
in queue 2023 mandates offers buys backlog,
(LBNL) cluster $30M for Pearl 80% of
studies AI studies Street; projects
ISO-NE withdraw
picks
GridUnity
Queue growth, regulation, and the software push, per LBNL, FERC, DOE, Enverus and GridUnity (2021–2026).
Why the queue became the bottleneck
The queue was not always the problem. At the end of 2021 it held roughly 1.4 terawatts. By the end of 2023 it had jumped 27% in a single year to 2.6 terawatts, more than the entire existing fleet of US power plants, with solar and storage accounting for the bulk of the growth, according to Lawrence Berkeley National Laboratory (LBNL) data compiled by Utility Dive. Requests have kept climbing since: around 8,200 projects were actively seeking interconnection at the end of 2025.
The delay is not the study itself, which computers can now finish quickly. It is the serial structure around it. Under the old first-come, first-served model, one project's study could not start until the one ahead of it finished, and every analysis assumed the projects behind it had also been built. A single flawed assumption forced a restudy that delayed the whole line. Grid operators such as MISO, PJM and CAISO simply do not have enough power-flow engineers to chew through the backlog.
The cost of that backlog is concrete. In a single PJM capacity auction, the failure to connect cheaper new generation cost consumers an estimated $7 billion. Roughly 80% of projects that enter the queue withdraw before completion, and the interconnection studies they did trigger, up to a third of total project cost for the ones that failed, are sunk. The line keeps growing because applications are cheap, and the penalty for squatting on a queue position was, until recently, nil.
That is what the reformers and the software vendors are attacking.
Software that runs a month of engineering in minutes
The leading tool is SUGAR, the Suite of Unified Grid Analyses with Renewables, built by Pearl Street Technologies, a 2018 spin-out from Carnegie Mellon University in Pittsburgh. It is a power-flow engine designed for the specific chore of interconnection studies: model the grid, identify the network upgrades a cluster requires, allocate the cost, validate the fix. MISO has run it on Phase 1 studies since its 2022 cluster cycle.
The number the company quotes is 200x: engineering analysis that took months now takes hours. At the time of its acquisition in March 2025, SUGAR had processed more than 300 GW of queued projects for transmission providers, with another 100 GW in the pipeline. The vendor's current site claims 500+ GW contracted and more than 2,000 projects modeled.
Analysis speedup, SUGAR vs legacy tools
Power-flow simulation engine used by MISO for interconnection Phase 1 studies since 2022 · Pearl Street / Enverus, 2025
Bringing new generation capacity online faster is the greatest challenge facing the energy industry today.— David Bromberg, co-founder and CEO, Pearl Street Technologies
The company also sells Interconnect, a developer-facing platform that runs custom studies on demand and returns an injection capacity map and a risk assessment in hours. For a developer deciding whether to sink capital into a site, that is the difference between guessing and knowing.
The consolidators and the federal push
Software this valuable does not stay independent for long. In March 2025, Enverus, the energy data giant, acquired Pearl Street Technologies, folding SUGAR into a platform that already serves more than 7,700 customers in 50 countries. "We are excited to integrate Pearl Street's and Enverus' data, analytics and platforms to provide an unprecedented end-to-end solution to all power market participants," said Manuj Nikhanj, its CEO.
A parallel consolidation is running in queue administration. In December 2025, ISO New England selected GridUnity and its GridInterConnect platform to automate the region's generation and large-load interconnection processes. With that win, GridUnity claims to serve the majority of ISOs and regional transmission organizations (RTOs) coast to coast, and its customers cover more than half the US population. GridUnity's pitch is FERC Order 2023 compliance by software: cluster-based studies, standardized data, automated workflow, the exact reforms regulators made mandatory in July 2023.
The public sector is funding the same shift. The Department of Energy's AI for Interconnection program, AI4IX, opened in November 2024 with up to $30 million to pair software developers with grid operators and project developers, managed through the ConnectWERX partnership intermediary. Federal money does not build the products; it buys the adoption risk that private capital was too nervous to price in.
At the distribution scale, where community solar projects connect to utility-owned lines rather than regional grids, the same automation wave is arriving through smaller players. Canary Media tracked MeanderX, a startup that scrapes utility interconnection data and gives mid-size developers live queue tracking: which projects dropped out, where disputes are clustering, which feeders are saturated. One Illinois developer told the outlet his projects had been stuck in Ameren's queue for 18 to 19 months. That is the long tail of the same problem.
Turning points
Whether queue volume actually falls once cluster reforms land. PJM ran its first cycle under the new first-ready, first-served process in August 2025, and CAISO launched its own prioritization push the same month.
Whether AI study software reaches true national coverage. GridUnity already claims most US ISOs and RTOs; MISO is the reference deployment for SUGAR.
Whether faster studies translate into faster power. The skeptical view, from grid analysts, is that software optimizes within the constraint: the physical transmission deficit, not the paperwork, may ultimately cap how much the queue can shrink.
What it means for investors
The interconnection bottleneck is now a bottleneck economy. One market-research firm, MarketIntelo, sizes interconnection queue solutions at $1.4 billion in annual revenue in 2025 and projects $5.8 billion by 2034, a 17.1% compound rate, with the AI-driven queue optimization segment growing faster still at over 20%.
The opportunity is narrower than those headline numbers. The grid software market has room for platforms, not point products: the three most interesting names are each attached to a bigger balance sheet. Its maker now belongs to Enverus, GridUnity is the platform of record across most US markets, and the DOE is deliberately spreading federal dollars across dozens of pilots rather than anointing one winner.
The honest read is that adoption is real but early. AI tools can cut study time from months to days, yet they cannot add transmission capacity or shorten permitting. The developers who will clear the queue are the ones with the best data on which projects actually survive it, and that is the durable advantage the software vendors are selling.
The sooner we can speed up interconnection, the better off we'll be.— Rob Gramlich, president, Grid Strategies
