Production homebuilding is manufacturing trapped in computer-aided design (CAD) drafting software. Builders configure, price, permit, sell, and build homes thousands of times a year, but the underlying software treats homes like a one-off custom drawing. On top of that, the people working within homebuilders and at service firms supporting homebuilding are converting those drawings into calculations, estimates, and other outputs in a structurally inefficient way. In what other industry would you do critical cost estimation math to define a purchase order of stone by reviewing a drawing?
That gap is why AI has not yet shown up at the job site in any meaningful way, and why we just led a $95 million Series C with Insight portfolio company Higharc.
The round, with participation from Wellington Management and Fifth Wall alongside existing investors, brings Higharc’s total funding to more than $170 million. Alongside the financing, Higharc launched AI estimating for building-materials distributors, debuting with US LBM, one of the largest privately owned full-line specialty distributors in the U.S. We have been tracking Marc Minor and the Higharc team for seven years. The company had matured from an architectural thesis into a platform infrastructure builders rely on every day into a system the supply chain plugs into.
Higharc’s result: more efficient homebuilding, enabling affordability, and producing homes that delight the people living in them every day.
An industry running on disconnected files
Homebuilding is one of the most economically important industries in the U.S. Single-family housing starts were running at roughly 1 million on a March 2026 run rate, and the National Association of Homebuilders (NAHB) estimates that custom-built homes accounted for only about 20% of single-family starts in 2025. The rest is repeatable production building, driving scale efficiency to build homes affordably, with last-mile customization to suit a consumer’s preferences.
But the software has not caught up. Builders run on a patchwork: CAD (or Revit) for drafting, spreadsheets for estimating, separate rendering tools, enterprise resource planning (ERP) for purchasing, and construction management platforms for the field. A single design change ripples through every one of those systems through phone calls, markups, and re-checks. Traditional CAD was built drawing-first. Production builders need a system that can keep blueprints, takeoffs, permits, sales collateral, and purchase orders in sync from one source.
That is also why leveraging horizontal technology alone has not solved this. In homebuilding, approximate answers lead to delayed starts, material overages, and change orders. To make AI useful here, the model needs a reliable representation of the home underneath it.
Higharc starts with the home as data
Higharc’s core move: Represent the home as structured spatial data, then generate every downstream artifact — blueprints, estimates, permit documents, 3D renderings, sales configurators — from the same model. The company is not layering AI on a legacy drawing workflow. It is rebuilding the workflow around a data model that AI can reason over.
With Higharc, when a builder changes a wall, an elevation, or a finish option, the change flows through the drawings, the estimate, the 3D model, the showroom experience, and the construction documents at once. The reconciliation work that has sat between design, purchasing, sales, and the field for decades collapses into the platform.
Why this team saw the problem differently
Cofounder and CEO Marc Minor came to homebuilding through advanced manufacturing, with experience at Carbon3D and Desktop Metal, and he experienced his own frustration trying to get a home built. His view: Generative design had already transformed how physical products get made, and homes were overdue.
His cofounder Michael Bergin brought the technical and product brain to match. Before Higharc, he led Autodesk Research’s generative design work for the built environment, after graduate research on mass-customized housing. The broader leadership team combined CAD, generative design, and interactive 3D backgrounds with operators from inside homebuilding: a team of drafters, estimators, and product leads who knew what actually happens in the field.
They were backed by a team that combined CAD, generative design, and real-time 3D with operators who knew what actually happens in the field — including Cofounder and CTO Emeritus Peter Boyer, who built CAD and real-time rendering on the web, and Cofounder Thomas Holt, who leads visualization from a background in 3D art and gaming.
Describing “AI for homebuilding” is easy. Encoding how a home is actually designed, configured, estimated, sold, permitted, and built is the hard part.
Why now
We led the Series C because Higharc hit an inflection across three dimensions.
First, the product became core infrastructure for the workflow. Builders report compressing product development from months to weeks, opening communities faster, and improving their margins and ability to scale.
Second, the platform compounds. As each builder’s standards, options, and construction logic get represented in the model, the system becomes harder to replace. It is no longer a drafting tool. It is the builder’s operating layer.
Third, the same architecture is now extending into the supply chain. Distributors face a version of the builder’s problem: fragmented tools, manual plan review, slow estimating, and inconsistent takeoffs. Higharc’s AI estimating product, powered by AutoTranslate AI, turns a 2D floor plan image into a 3D spatial model and generates structured takeoffs aligned to purchasable materials. The US LBM partnership is not a side product. It is evident that the home data model becomes useful to more participants across residential construction.
A platform, not a CAD replacement
Day one was replacing CAD. Today, Higharc is building the data backbone for residential construction: a single representation of the home that builders, distributors, manufacturers, trades, and buyers all operate from. Higharc’s AI compounds with faster configuration, more accurate estimates, earlier issue detection, more personalized buyer experiences, and less waste across the system.
That is the kind of vertical AI company we look for at Insight Partners, driving superior customer outcomes. Not a horizontal model retrofitted onto a legacy workflow, but a workflow rebuilt so AI can deliver production-grade, build-ready outcomes.
As Josh Fredberg, Managing Director at Insight Partners, put it:
“Most AI in legacy industries is being layered onto systems that were never designed to support it. Higharc took the harder path and built the data foundation first. That is why we believe it has the potential to become the defining platform for residential construction in the AI era.”
What comes next
Higharc was named to Fast Company’s Most Innovative Companies of 2026 and ranked #28 on Deloitte’s 2025 Technology Fast 500. The next phase is scale: deeper AI capabilities for builders, broader supply-chain participation, and more workflows on the same data model. US LBM is a public proof point of that direction. We believe more will follow.
Homebuilding will not become AI-native by adding a chatbot to a drafting tool. It will become AI-native when the home itself is represented in a way software can understand, reason over, and keep synchronized from design through construction. That is what Higharc has built, and why we are excited to partner with Marc, Michael, and the team as they build the AI operating system for residential construction.
