Mach9: bringing civil engineering into the AI era

Every road, utility line, and drainage system starts with a survey. Field crews capture the physical world in LiDAR, but turning that data into a usable deliverable is still done by hand. Drafters trace curbs, paint lines, poles, and drainage structures point by point, in desktop software designed decades ago. Mach9 is automating this work, which is why Bessemer is thrilled to co-lead the company's $34M Series A and partner with a team that knows this problem well.

A buildout that outpaces its workforce

The U.S. is in the middle of an infrastructure buildout cycle akin to the railroad buildout that took place when Bessemer was founded. Federal and state programs are funding repairs to aging roads and bridges, data centers are straining power and water systems, and utilities are racing to extend and harden the grid. Broadband, electrification, and network rebuilds add more projects, and each one can only start by capturing and measuring the current reality.

That first step is where the cycle stalls. The engineers and drafters who turn field data into design-ready deliverables are in short supply, and the pipeline of work is far outpacing how quickly they can draft, which is a problem made worse by the legacy software they rely on. The stakes are high, too. Design, permitting, construction, and asset management all begin with the survey, so any delays and errors have a domino effect on every subsequent step.

While capture has gotten cheaper and faster, with better sensors and more ways to collect scan data, extraction hasn't kept up. Skilled drafters fill the gap between what a crew collects and what an engineer can use, but their hours cap how much infrastructure gets mapped, designed, and maintained. That labor sits inside project budgets, which are far larger than any software budget these firms and agencies manage. By automating this entire workflow, Mach9 helps firms take on more projects and improve margins without scaling drafting hours at the same rate.

Automating the work that comes before every build

Mach9 ingests LiDAR scans and imagery and uses AI to extract the features a surveyor would otherwise draw manually. The interaction feels familiar to anyone who has used AI coding tools. The model proposes a first draft, and the surveyor edits and refines.

The bar for this product is exceptionally high because survey deliverables feed directly into design and construction, so an error on the page becomes an error in the field. A tool that is mostly right isn’t usable, which is why so few have tried it. Mach9 spent its early years crossing the threshold where professionals could rely on the product. Today, customers run it in production, and they tend to expand their use quickly once it becomes part of how they deliver projects.

Bessemer’s conviction in Mach9

Three key factors went into our decision to back Mach9 and the team:

1. Mach9 earns from the project budget, not the IT budget. Pricing is usage-based, per mile of road or acre surveyed, which mirrors how engineering firms bill their own clients. Mach9 takes on labor spend rather than competing for a line item in software spend. Its revenue grows as customers do more work with it, and customers can see the return per project.
2. The data compounds and the moat deepens. Each project improves the models, and each survey stays in the platform as a system of record. Over time, that gives Mach9 a dataset of scan data paired with survey-grade deliverables that no one else can easily assemble. Incumbents sell seat-licensed desktop tools and have little incentive to automate the work their pricing depends on.
3. Surveying underpins civil engineering. Accurate survey data is the starting point for designing, building, and maintaining infrastructure. By automating the process of turning field data into usable CAD deliverables, Mach9 addresses a critical bottleneck across these workflows. We believe this gives the company a strong foundation to become an enduring AI-native software platform for civil engineering.

A team built for what’s next

Co-founders Alexander Baikovitz (CEO), Haowen Shi, Michael Mong, and Zachary Sussman met at Carnegie Mellon while building robots. The team witnessed firsthand the challenges around surveying. Their robotics backgrounds gave them deep expertise in 3D perception, and they’ve been building Mach9 together since.

With this financing, Mach9 will expand its go-to-market team, deepen the product across new data sources and feature types, and bring their AI CAD software to more of the engineering firms, public agencies, and utilities that maintain the built world.

We've been backing vertical software for decades, including Procore for construction, EvenUp and Legora for legal, Abridge for healthcare, and much more. Mach9 fits neatly into our idea of a deeply vertical, AI-native product making skilled labor more productive to achieve goals faster than ever. The U.S. and beyond are in a buildtime boom, and Mach9 is helping surveyors and civil engineers keep pace. We're grateful to Alex, Haowen, Michael, Zach, and the entire Mach9 team for the chance to build alongside them.

Learn more at mach9.ai, or join the team.