CivilGrid has raised $26 million in a Series A led by Spark Capital, with Energy Impact Partners, Afore Capital, A*, Ford Street Ventures and SNR Ventures participating. The San Francisco company sells a preconstruction platform that pulls utility, environmental, geotechnical and property data into a single map, and the round was announced on August 27.
The founder-market-fit story is unusually literal. Co-founder and CEO Josh Mackanic spent roughly a decade as an engineer at Pacific Gas and Electric before co-founding the company in 2020, and told TechCrunch the origin was a job that stopped because of a pipe nobody knew was in the excavation.
The Problem Is a Data Problem, Not a Digging Problem
Everything buried under a street belongs to somebody different. Electric, gas, water, sewer, telecom — separate operators, separate records, separate formats, separate levels of accuracy. Layer on parcel ownership, environmental constraints and geotechnical reports and you have a design input that no single party holds and no single party is accountable for.
The consequence is a well-documented failure mode. CivilGrid’s announcement puts US utility strikes at roughly 200,000 a year and the associated cost of damage, delay and rework at about $30 billion annually. Those are the company’s framing figures rather than independently audited ones, and should be read that way — but the direction is not in dispute, and the industry has been measuring some version of this number for years.
What CivilGrid sells is the aggregation step. Its own site describes centralising utility, environmental and geotechnical data, generating utility composites, downloading shapefiles, and capturing field observations back into the same record — so that the scoping conversation happens against one map rather than six PDFs and a phone call.
The PG&E Number Is the One That Matters
The strongest evidence in the announcement is a case study with Mackanic’s former employer. CivilGrid says that across 1,600 planned gas distribution projects, PG&E identified more than $60 million in avoidable paving conflicts, cut field research activity by about 40%, and pulled project timelines forward by up to four months.
Treat the precise figures as vendor-supplied — they come from CivilGrid’s release, not from a PG&E filing. But the shape of the claim is worth sitting with, because it is not a productivity story. It is a scope story. Paving cost is not incurred because the crew worked slowly; it is incurred because nobody knew, at design time, what the trench would have to cross. Finding that out earlier does not make anyone faster. It makes the project smaller.
That distinction is why this kind of software can clear a utility’s procurement bar. Efficiency claims get discounted. Avoided capital does not.
The Customers Are Checkable
CivilGrid’s announcement names PG&E, Atmos Energy, and engineering firms Mark Thomas and GHD. Its own site carries named testimonials from Mark Thomas, GHD, Underwood & Rosenblum, Civiltec and San Jose Water.
These are vendor claims rather than independent confirmations, and we have not sought comment from any of the named firms. But they are checkable institutions with checkable people attached, which puts them a category above an anonymous logo wall — and assembling investor-owned utilities and large engineering consultancies from a 2020 start is the substantive signal in this round.
Christine Cowsert, a senior vice president at PG&E, is quoted in the announcement saying tools like CivilGrid “help our teams identify risks earlier, build more efficiently.” Alex Finkelstein, a partner at Spark Capital, frames the bet more expansively, calling it “a new standard for how America builds.”
The Investor Mix Says Where This Sells
Spark Capital is a generalist early-stage firm — it led Twitter’s Series B in 2008 — and its presence here reads as a view that preconstruction data is a large, boring, under-served market rather than a niche one.
Energy Impact Partners is the more diagnostic name. EIP is built around a coalition of utility and industrial companies as limited partners, which is a different kind of validation than a venture markup: it suggests the buyer set has looked at this and wants it to exist. For a company whose flagship reference is a gas distribution programme, that is the investor you want on the cap table.
The Honest Caveats
This is not an AI company, and does not claim to be. CivilGrid’s own site does not mention artificial intelligence or machine learning. It is a data aggregation and workflow product. We cover it here because the constraint it addresses — that the built environment’s site data is fragmented across owners and formats — is the same constraint that limits every AI tool aimed at design and preconstruction. Models cannot reason over records that were never assembled. But nobody should read this as an AI round.
Valuation and total funding were not disclosed. Where the $26 million leaves CivilGrid’s cumulative raise, and at what price, is unknown.
The metrics are the company’s. The 200,000 strikes, the $30 billion, the $60 million and the 40% all originate in CivilGrid’s announcement. We have verified that the announcement makes those claims and that the named customers appear in it. We have not independently audited any of the figures.
What to Watch
Whether this data becomes a permitting input rather than a design convenience. Right now the platform’s value proposition is internal: better scoping, fewer surprises, less avoidable paving. The larger version is regulatory — if utilities and municipalities begin requiring a consolidated subsurface record as part of the permit file, the product stops being a way to work faster and becomes a condition of starting at all.
For a related bet on the compliance side of infrastructure delivery, see our coverage of Dili’s $15 million Series A.
