A Startup Is Betting AI Can Draft Engineering Plans From Land Surveys

Mach9 announced a $34 million funding round on Wednesday, October 21, 2026. TechCrunch The round is called a Series A, which usually comes after a company's early seed money.
First Round Capital and Bessemer Venture Partners led the round together. Prudence and Quiet Capital also took part. Quiet Capital was already an investor in the company.
Mach9 has dropped "Robotics" from its name. It now makes software for civil engineering, the work of designing and building public projects such as roads, bridges and water systems. The software is a type of CAD, short for computer-aided design, which engineers use to draw plans on a computer.
Alex Baikovitz co-founded the company in 2021. His goal was a faster, cheaper way for engineers to inspect sites. Mach9 first built surveying equipment to do that. Then it concluded that the equipment was not the real problem. The slow part came afterward, when survey measurements had to be turned into usable drawings.
The new software takes survey data and produces a first draft of a plan. Engineers then refine it, much as an editor works from a writer's rough draft. Mach9 says this makes survey-to-plan work 10 times faster and greatly cuts costs. Construction work needs positions accurate to within about a centimeter, so the company built its own AI models, a effort it has focused on since 2024.
About 100 customers tested the software in a private beta. The company said it has used the software on more than 2,500 projects. TechCrunch
The company's earlier funding was a $12 million Series Seed round led by Quiet Capital, reported in November 2024, to scale its AI mapping work. GeoWeek News Another company, CivilGrid, raised $26 million in August 2026 for a mapping product it calls "Google Maps for the underground." TechCrunch
In my view, the round is a bet on where the real delay lies. Measuring ground conditions faster helps little if the drawing afterward is still done by hand. Mach9 has concluded that drafting, not measuring, is where most time and money are lost. The 10x claim will get the closest examination, and it will be tested by how many of those 2,500-plus projects end with approved drawings and only manageable edits.
The broader context here is that survey, mapping and design tools have usually come from separate companies with separate teams. If a draft plan can come straight from survey data at construction-grade accuracy, moving from field measurement to design review becomes mostly a software question. Whether those drafts hold up under professional review at scale is what the next year of customer data should show.


