Mach9 has raised $34 million in a Series A co-led by First Round Capital and Bessemer Venture Partners, with participation from Prudence and continued participation from Quiet Capital. The company’s announcement, datelined San Francisco and timed 10:00 ET on 21 October, puts total funding at $49 million.
The money is going into a product that does one narrow thing: it turns laser scans of roads and corridors into the CAD drawing a licensed engineer signs. What makes the round worth more than a line in a weekly roundup is how Mach9 arrived at that product. It started somewhere else entirely.
The Company Was Built Around Hardware, and Then Wasn’t
TechCrunch’s Sean O’Kane reported the round the same day and has the sequence. Alex Baikovitz co-founded the company in 2021 as Mach9 Robotics, building a faster and cheaper way for civil engineers to carry out visual inspections. Within a few years it had built the hardware. It also had the conclusion that the hardware was not where the pain was: survey teams, Baikovitz found, were struggling to turn the data their scanners already produced into usable maps. “The real pain was in their workflows,” he told TechCrunch. A customer’s VP put the other side of it to him more bluntly: “I’m not interested in getting this cheaper, lower-cost hardware.”
So the company dropped “Robotics” from its name and moved to AI CAD software, building its own models from 2024 and training them for centimetre-level accuracy. The early hardware was not wasted — per TechCrunch it was what got the customer conversations in the first place.
That is an unusually clean version of a pattern this newsroom keeps running into. The expensive, visible part of a construction-data problem is almost never the bottleneck; the bottleneck is the human hours between the capture and the deliverable. DroneDeploy’s $845 million sale to Procore was a bet on the same gap from the imagery side. CivilGrid raised $26 million for the layer of the job nobody owns — the subsurface records that have to be assembled before a design is safe to build. Trendspek raised A$6 million on the premise that the record has to outlive the inspection. Mach9’s version is the drafting desk.
What the Software Actually Produces
Per the release, engineering and surveying teams upload 3D scans of roads, utilities and construction sites “collected with the equipment they already use.” Mach9 then automatically drafts physical features — the release names curbs, signs, poles, lane markings and wires — at centimetre-level accuracy, and surveyors and engineers “review, refine and finalize the drawings using Mach9’s editing tools.”
The company’s own site is more specific about where the output lands: exports in DXF, DGN, Esri Feature Services and field-to-finish survey formats, and ingestion from drone, terrestrial and mobile scanning, with Trimble, RIEGL, Leica, NavVis and Phoenix named as hardware it reads. The site’s headline is “Turn reality capture into complete deliverables faster”. That is the whole proposition: not a new sensor, not a new file format, but a shorter path between the two.
The speed and cost claims are the company’s. Mach9 says the process runs “up to 10x faster at one-fifth the cost,” and its site carries a comparison chart putting time-to-deliverable for one dataset at 223 days by conventional surveying, 59 days with reality capture plus manual drafting, and 6 days with Mach9. We report those as Mach9’s figures and have not independently verified them.
Worth noting for anyone who followed the company earlier: at the seed round the pitch was measured differently. Mach9’s own November 2024 announcement of a $12 million seed led by Quiet Capital — dated 14 November 2024, with Overmatch Ventures, Kyle Vogt, Amar Hanspal, Scott Belsky and Gokul Rajaram also investing — described the company as “an industry leader in AI-powered map making software” and quoted Baikovitz saying its software “accelerates mapping up to 96 times faster than traditional methods.” Two years on, the deliverable is a CAD drawing rather than a map. That change in what the company sells is the clearest marker of the pivot TechCrunch describes; both framings are Mach9’s own, and we are reporting the shift in product, not setting one of its figures against another.
The Customer List Is the Strongest Part of the Release
Funding announcements at this stage usually gesture at traction. This one names it, which makes it checkable.
Mach9 says engineering teams have used the software across more than 2,500 projects in all 50 states, surveying over 100,000 miles of infrastructure corridors. Named customers in the release are the civil engineering firms WSP, HNTB and NV5 and the transportation agencies Caltrans, Minnesota DOT and the Maryland State Highway Administration. The release also describes three specific deployments: Caltrans inventorying rumble strips and other road-safety features on Interstate 8 in San Diego; Kier + Wright working on pedestrian safety around schools in San Francisco; and Fibersmith helping connect 125,000 homes and businesses in Louisiana that lacked high-speed internet.
The customer voice in the release is Rob Ashley, HDLS technology director at Pennoni, the Philadelphia-headquartered multidisciplinary engineering firm: “Mach9 enables us to do 50% more work with the same experienced team. That means we can take on projects we otherwise wouldn’t have the capacity to deliver.”
That quote is the commercially interesting one, because it is a capacity argument rather than a headcount argument. TechCrunch reports nearly 100 customers in the private beta of the software.
Why Two Generalist Funds Want a Drafting Tool
The investor case in the release rests on volume of physical asset, not on software market size.
“There are 4.2 million miles of public roads in the U.S. alone, before you count the power lines, fiber networks, railways and other infrastructure we rely on every day,” said Bill Trenchard, partner at First Round Capital. “Mach9 is building the software layer that will turn our physical world into information engineers can understand and act on.”
Caty Rea, vice president at Bessemer Venture Partners, framed it as a category reset: “AI is creating the opportunity for a new generation of CAD software. Just as software transformed construction, AI CAD will transform how infrastructure is surveyed, designed, built and operated.”
There is a lineage detail in the release worth pulling out. Rea led Bessemer’s investment in Mach9 with Brian Feinstein — and Bessemer’s own page for Feinstein says he “led early-stage investments in Procore (IPO), Wildlife, Restaurant365, Ada, Weave (IPO), Mambu, Fulcrum, and Clio.” The firm is running the same play it ran on construction management software one cycle earlier, a layer upstream, with one of the same people on it. That is our reading of the pairing, not a claim the release makes.
The engineering pedigree is also doing work here. Mach9 was founded by Baikovitz with Haowen Shi, Michael Mong and Zach Sussman, who met at Carnegie Mellon University’s Robotics Institute and previously worked on perception and navigation systems at Tesla, SpaceX, Apple, NASA and Google X. This is a CAD company staffed by people who came out of autonomy, which is roughly what the pivot story would predict.
The Demand Case, and What It Rests On
Two numbers carry the release’s argument about why now, and they need different handling.
The first is public record: the Infrastructure Investment and Jobs Act allocated $1.2 trillion to rebuilding and repairing American infrastructure. The second is the release’s own: it says resident surveying licences fell 24% between 2011 and 2025. We could not independently verify that figure against a licensing body’s published counts — the nearest reference points we found were older NSPS compilations of NCEES data — so we are attributing it to Mach9 rather than stating it as established. A surveying-workforce shortage is widely reported in the trade press; the specific 24% is Mach9’s number.
Put together, the release’s case is that there is more infrastructure to document than ever and fewer licensed people to document it, so the hours have to come from somewhere. Ashley’s “50% more work with the same experienced team” is that argument in one sentence, from a customer.
What Is Not In The Record
Neither the release nor TechCrunch’s coverage gives a valuation for this round, and we searched for one in the secondary coverage without finding it. Neither document gives a revenue figure, an ARR number or a headcount either. On the use of funds, the release says only that the capital will expand Mach9’s AI software for surveying and civil engineering; TechCrunch reports Baikovitz describing the spend as product, hiring industry experts and scaling the organisation.
One point of care on the product boundary: Mach9 produces a first draft. The release is clear that licensed surveyors and engineers review, refine and finalise the drawing, and nothing in it claims the software signs or seals anything. In a profession where the signature is the legal product, that distinction is the whole business model, and it is to the company’s credit that the announcement does not blur it.
Baikovitz’s own framing of the round is the widest claim in the document: “We’re expanding the world’s capacity to build and maintain the infrastructure that people depend on.” Rea’s is narrower and more testable — that Mach9 “can become the platform that brings civil engineering into the AI era.” On the evidence in this release, the narrower claim is the one with customers attached to it.
