Ryhla, a Y Combinator Fall 2026 startup, has launched models it says replicate the distribution grid, aimed at helping solar, storage and EV charging developers pick sites likely to win utility approval for projects of 1 to 50 megawatts. The company's claims have not been independently validated, which matters because interconnection is a bottleneck and, as this analysis argues, the tools meant to ease it are multiplying faster than the evidence that they work.
What Ryhla says it does
Ryhla's website describes its market as distribution and subtransmission interconnection, spanning solar, batteries, EV charging and small data centers. It is not a transmission-queue business.
"Our proprietary models replicate the distribution grid, and we use them to take projects through interconnection end to end," the founders wrote in their joint YC launch statement, which was accessed November 6, 2026.
They also said they grew up in Morocco, have been friends since middle school, studied at École Polytechnique and researched grid intelligence at Stanford. Their first deployments, they said, were with utilities. No utility names or deployment dates accompany that account.
The company's public materials promise faster screening and better application selection, but they do not disclose independently verified reductions in the days between application and energization. Its approval statistic, which concerns 2025 applications, should not be read as a completed-connection rate.
Approval is not energization
Berkeley Lab's September 2026 distribution research draws a line between executing an interconnection agreement and receiving permission to operate, the latter of which includes construction and inspection. Approval is one milestone among several.
The California Public Utilities Commission made a similar point in an August 17, 2026 article on transportation electrification. Charging equipment can be delivered and installed in months, it said, while localized grid upgrades push energization later. The commission also described flexible connections, which allow partial-capacity service while customers wait for full energization. Smarter site selection or faster studies do not eliminate those construction intervals.
The data gap
National queue figures do not describe Ryhla's market. Berkeley Lab's Queued Up: 2026 Edition leaves out distribution-connected generation, behind-the-meter projects and load-interconnection requests.
The lab's distribution report has limits of its own. It found that 97.3% of applications were below 30 kW and that California accounted for 65% of applications. It also cautioned that recent medians for completed projects can skew low, because slower projects remain unfinished. The figures, in other words, are not representative national wait-time estimates.
A crowded field
Ryhla is entering a busy market for grid-capacity and interconnection tools, and the competitors offer a useful yardstick for what evidence looks like.
GridCARE raised a $64 million Series A, reported May 14 and 15, 2026 by Latitude Media and Data Center Dynamics, led by Sutter Hill Ventures. It is an adjacent power-acceleration business, not a valuation comparable for Ryhla. At a May 29, 2026 Portland General Electric presentation hosted by Oregon's Department of Energy, the partnership was said to have uncovered more than 80 MW of capacity in existing infrastructure. In an October 2, 2026 announcement, Shinjini Menon, Southern California Edison's senior vice president of system planning and engineering, said: "We are looking for innovative solutions that can help us get more out of our existing grid." The announcement describes identified potential, not 20 MW already energized.
Paces said EDF evaluated more than 90 assets for data-center co-location and triaged them within weeks into a ranked pipeline. That does not show that every ranked project connected. Paces also said an investor weighing more than $750,000 in interconnection-study spending received an initial utility indication of 400 MW, while its power-flow assessment put the figure closer to 150 MW. The customer is unnamed, and the claim is company-reported.
Elsewhere, Enverus acquired Pearl Street Technologies in March 2025. GridUnity, which said on October 13, 2026 that it had been honored by the Alliance for Climate Transition, sells into a similar problem. "Our mission is simple: make it faster and easier to connect new generation and large loads to the grid," said Keith Laska, its chief growth officer.
Site choice is only one lever
IREC, a participant in the technical process, reported regulatory changes in Minnesota and Massachusetts. Its Minnesota article suggested that application design and operating controls, not only site choice, can affect outcomes.
IREC also reported on May 14, 2026 that Maryland and Oregon had created limited-operation pathways while upgrades are completed. Those emergency provisions were set to expire in early September and October 2026 unless made permanent. Their current status was not verified.
What to watch
The founders' academic work offers a technical clue, though a modest one. One of them co-wrote a December 20, 2025 Stanford CS224W course project on graph neural networks for optimal power flow, using an IEEE 118-bus benchmark. It is a course project, not evidence of Ryhla's production architecture or commercial accuracy.
The open questions are practical: how the production models enforce electrical constraints, quantify uncertainty and undergo utility validation. Ryhla has not published answers. Until it names utility partners and reports elapsed-time results, its pitch remains a claim rather than a benchmark.
