Problem

Spark addresses the slow, fragmented, and economically consequential permitting and site-diligence process for solar, battery storage, and data-center developers. Critical information is spread across zoning codes, permits, local-government records, maps, and community channels, forcing teams to spend weeks researching sites, miss promising opportunities, or advance projects with fatal flaws. The economics are especially painful in power-constrained infrastructure: Spark cites an estimated $98 billion in projects blocked or delayed in Q2 2025 and estimates that a one-month delay on a 60-MW data center can erase $10 million in value.

The killer use case is early-stage site screening: before a developer commits substantial capital or time, Spark helps determine whether a project can realistically be permitted at a particular parcel or in a target county. It is designed to surface zoning restrictions, moratoria, permitting obstacles, opposition, and other risks early enough to avoid pursuing unviable sites and to improve speed to power.

Product / Service

Spark is domain-specific workflow software for energy infrastructure and data-center developers. Its AI system scans and interprets millions of PDFs, maps, land-use codes, local websites, town-hall materials, social-media activity, and other project documents, then produces evidence-backed outputs such as permit matrices, zoning and permitted-use analyses, process steps, risk flags, community-sentiment intelligence, alerts, heatmaps, and data-room diligence. Users can analyze an individual parcel, screen entire markets, or monitor more than 3,000 U.S. counties for changes in zoning, moratoria, or sentiment.

The delivery model combines a recurring permitting, zoning, and sentiment intelligence platform with expert-curated permit matrices. A single site-specific matrix is priced at $3,500 and includes state, local, and federal requirements, an authority-having-jurisdiction contact directory, expert outreach, and a standard seven-day turnaround; the platform includes two matrices per month for customers on its permitting-intelligence plan. The benefit is a faster, more verifiable alternative to manual research and a complement to GIS tools: Spark aims to answer not merely where a site looks attractive, but whether the project can actually be built there.

Market

Spark competes in the emerging market for AI-enabled permitting intelligence, energy-infrastructure site selection, and project-development diligence. Its primary customers are solar, storage, and data-center developers, with broader applicability to infrastructure such as wind and natural gas. The competitive set includes adjacent GIS and site-selection platforms such as RatedPower, LandGate, Latapult, Pivvot, dvlp.energy, and Spottitt; Spark differentiates itself by emphasizing permitting interpretation, persistent citations and timestamps, local-government and community-sentiment monitoring, and evidence-backed risk analysis rather than GIS-based suitability scoring alone.

Spark is a live, customer-facing company rather than an announced pre-revenue concept. Its public materials say it is backed by Y Combinator’s Winter 2024 batch, has been trusted by dozens of developers representing more than 90 GW of project pipeline and $30 billion in financing, and reports coverage of 3,000-plus counties, 100-plus projects screened per customer on average, and a 60% reduction in evaluation time. It announced a seed round after YC but did not disclose the amount or investors in the cited announcement, and no public revenue figure or explicit pre-revenue designation was found.