Post by Robin Louvet

As a data scientist who started my career building software for power systems, I love seeing AI tackle physical, real-world challenges over sandbox benchmark environments. Today, Tapestry announced initial results from the first phase of our collaboration with PJM Interconnection, North America’s largest grid operator, and the metrics show a step change in operational velocity. PJM’s engineers used HyperQ, Tapestry’s AI-powered grid interconnection tool, to streamline the site control portion of PJM’s newly reformed Cycle 1. Evaluating an unprecedented 811 applications for new energy generation projects — representing 220 gigawatts of potential generation capacity — HyperQ successfully executed 9,312 parallel site control assessments with a median computational runtime of just 6 minutes and 15 seconds (a 20x+ acceleration over traditional workflows). Establishing site control for any energy project requires navigating complex, unstructured property deeds and lease amendments. PJM’s operators are expert power systems engineers, not real estate lawyers, and until recent reasoning models matured, automating this kind of unstructured legal analysis was largely out of reach. Our team at Tapestry was able to get into the weeds with PJM and make it possible. This collaboration puts macroeconomic theory into practice. Drawing on Stanford economist Chad Jones’s research on how systemic bottlenecks constrain growth, we targeted the administrative intake phase: deploying parallel AI agents to clear this primary upstream queue and make PJM's engineering expertise far more valuable. HyperQ acts as a powerful backend “super assistant,” surfacing exact evidence and page-level citations so PJM's experts are freed up to make the final, critical determinations. While we certainly aren't declaring victory on the entire interconnection queue, today’s milestone is an important first step toward accelerating what has historically been a painstaking, manual, sequential process. Stay tuned for more results in the coming months. Read the full performance report here: https://lnkd.in/gvRqYv3K #TapestryEnergy #AIforEnergy

Video Content