This Week in Permitting Tech June 1, 2026: AI Is Driving the Deals, and the Bill Lands on Local Government

Share
This Week in Permitting Tech June 1, 2026: AI Is Driving the Deals, and the Bill Lands on Local Government

This is an AI-flavored edition, because the week's biggest permitting-tech stories are all the same story. CoStar paid $800 million for a development-data platform built on public records. The records boom burying local clerks turns out to be the AI industry feeding itself. And GovWell expanded again, with a number worth an asterisk. The thread tying them together: the value AI creates from public permitting data keeps getting captured privately, and the cost keeps landing on the governments that hold the data.


CoStar pays $800M for Zonda

CoStar Group, the S&P 500 real estate data company behind LoopNet, Apartments.com, and Homes.com, agreed on May 29 to buy Zonda for $800 million in cash. Zonda is the dominant new-home-construction data platform, with more than 3,000 customers and, at its core, a proprietary lot-level database covering land development, construction status, and home sales, as well as the NewHomeSource and Livabl marketplaces. CoStar is buying the data and the builder workflows wrapped around it.

Look at where that data comes from. That kind of lot-level development data originates substantially in public records: plats, permits, construction filings, the paper trail every project leaves at a county office. Structure it, wrap software around it, embed it in how builders underwrite and plan, and the result is an asset CoStar will pay $800 million to own. The records are public; the structured, joined, software-wrapped version is private and increasingly expensive. An $800M deal is the market putting a number on how valuable it is to be the company that holds the structured copy of the public's own development data.

That value didn't appear from nowhere, and it didn't accrue to the counties that generated the records. It's the through-line for the rest of this issue: AI has made structured permitting data worth real money, and the systems built to extract and resell it are running well ahead of any decision about what the public keeps. The broader market's cooling doesn't seem to be slowing it. Q1 gov-tech deal volume was $1.9 billion, down from $3.5 billion in the prior quarter, per investment bank Shea & Co., a kind of pullback that usually accelerates consolidation rather than easing it.

CoStar / Zonda · Gov Tech Biz Q1 Roundup (Jeff Cook / Shea & Co.)

The FOIA boom is partly an AI data grab, and now there's AI to handle that too

CivicPlus published a piece this week on the strain that rising public records requests are placing on local governments, and the numbers are real. Federal FOIA requests topped 1.5 million in fiscal 2024, a 25% jump in a single year, with backlogs up a third and processing costs at $723 million. Local governments are drowning in the same wave: New York City's records requests doubled between 2018 and 2024, and San Antonio's open-records team grew from two people to twelve to keep up, processing more than 86,000 requests last year. The vendor's framing is a demand for transparency, and the suggested fix is, naturally, its own AI module to flag and redact records faster.

Here's the part the post leaves out, and it connects straight back to the Zonda deal. A meaningful share of that request volume isn't journalists or curious residents. It's the AI permitting and property-intelligence industry feeding itself, the same industry whose structured output CoStar just valued at $800 million. The tools run on local development data, permits, plans, inspections, and code-enforcement histories, and, for several of these companies, the competitive edge is getting that data before a jurisdiction has digitized or published it. Shovels, a venture-backed permit-data startup, advertises coverage of more than 1,800 jurisdictions and 180 million permits, and a product called Decisions that surfaces council and planning-board approvals months before a permit is ever filed. Being early is the pitch. The way you get the records a portal hasn't posted yet is to ask for them; at volume, that means records requests. So the same buildout that sells cities on AI to move permits faster is generating the FOIA load that's burying their clerks, and the answer on offer is a second layer of AI to process the requests. AI tools manufacturing work for AI tools, with an overworked records officer stuck in the middle.

It's another version of the abundance paradox. The data exists, but because nobody built the governance to share it openly, in standard formats, on a public cadence, it has to be pried loose one FOIA request at a time. Automation gets bolted onto both ends of a broken pipe, intake on one side, redaction on the other, while the thing that would actually relieve the pressure, a decision to publish the data as a matter of course, goes unmade. This is a thread worth more than one section, and the companion AI in Permitting column out today lays out the whole map: the nine ways companies actually get permitting data, and why the contract, not the method, decides who keeps it. Right now the market is selling everyone the more expensive option.

CivicPlus · Brechner FOIA Project · Shovels

GovWell expands to 35 states, and that 95% claim needs an asterisk

GovWell, fresh off the $25 million Series A this newsletter covered two weeks ago, now says it's running in more than 130 municipalities and counties across 35-plus states. The platform bundles an AI community assistant, an application portal built to catch errors before they reach staff, and a review engine that checks submissions against local code. It also claims it can cut processing times by up to 95%.

So: 95% of what? Processing time could mean the clock from submission to a staffer's first look, from submission to a final decision, or just the slice of review that one automated check handles. Knock 95% off the intake-and-error-flagging step, and I believe it. Knock 95% off the full permit lifecycle, multi-department coordination and all, and you're making a far bigger claim than the marketing copy is careful to specify. The part I'd actually buy is the error-catching portal. A huge share of permit delay traces back to incomplete applications ping-ponging between applicant and reviewer, and catching those at the door is a real win that doesn't require pretending software can do a planner's judgment for them.

GovWell Permitting · Series A Release (PR Newswire)

CEQ Permitting Innovators deadline: tomorrow

The CEQ Permitting Innovators Call for Solutions closes June 2 at 6 p.m. ET. Up to 50 selected solutions demo at the Expo in Washington this July and get featured in a Solutions Catalog distributed to federal agencies. If you're building in this space and haven't submitted, today is the day.

CEQ Call for Solutions


Permitting Tech is an independent news site covering investment, products, and policy in permitting technology. Written by Boon Sheridan.to flag and redact

Read more