Every dollar going into construction autonomy assumes the machine knows what’s in front of it. Underground, nobody does. That gap is what CivilGrid just raised $26 million against.
The round
Spark Capital led the Series A, with participation from Energy Impact Partners, Afore, A*, Ford Street Ventures and SNR. It was the largest disclosed construction technology round of the week ending 31 August, out of roughly $58 million across six companies. Founder Josh Mackanic spent about a decade at PG&E before leaving in 2020, which is a useful credential here: he spent that decade on the utility side of the records problem, watching what happens when the as-builts and the trench disagree.
CivilGrid pulls utility asset records, property ownership and environmental constraints into a single preconstruction view of a site. Mackanic calls it a Google Maps for the underground, which is his phrasing and not a technical description. The company points to roughly 200,000 utility strikes a year in the U.S. as the problem it’s aimed at. That figure comes from industry damage-reporting data and should be read as an order of magnitude rather than a census.
Records beat robots this week
The second-largest round in the same week went to Digs, and Builders FirstSource took the whole thing. Both deals funded records. That’s a shift after five weeks in which more than $2 billion went into construction autonomy, and it points at the same conclusion from two directions: the machines are further along than the data they need.
Locate tickets and 811 markings are the current control, and they’re a control on the surface. What sits three feet down is frequently a utility nobody has a drawing for, installed by a company that no longer exists, under a street that’s been repaved four times. Excavation crews already know this. What they haven’t had is a way to price the uncertainty before mobilizing.
Where this gets tested
Urban heavy civil is the proving ground, because that’s where the density of unknowns is highest and the cost of finding one with a bucket is worst. Rehabilitating structures in built-up corridors, like the Howard Street Tunnel clearance work under downtown Baltimore, turns on knowing what’s around the excavation before anyone breaks ground.
The honest caveat is that a records aggregator is only as good as the records, and the reason strikes happen isn’t that the data is hard to find. It’s that it doesn’t exist. Aggregation gets you a better prior. It doesn’t get you certainty, and no funding round changes that. TechCrunch first reported the round.