Silicon Valley Keeps Funding ConTech Founders Who’ve Never Set Foot on a Job Site

On April 13, Primepoint closed a $10 million seed round. Navitas Capital led the larger tranche, with participation from NextView, Penny Jar, GS Futures, Aglaé Ventures, and angel money from Yann LeCun. The company’s CEO, Lubomir Bourdev, is a founding member of Facebook AI Research, holds more than 100 patents, and built the object recognition system that ran across every photo on Facebook and Instagram. His co-founder, Hamid Palo, was employee number five at Trello. Between them, the construction experience is basically zero.

They hired someone for that. Kamran Azarbal, the VP of Strategy, spent over a decade at Webcor, progressing from field engineer to project director on multi-billion-dollar commercial jobs. Bourdev and Palo on the product side, Azarbal on the industry side. That structure, more than the pedigrees, is the interesting bit. Because if the debate is really “outsiders vs. insiders”, the most successful ConTech companies of the last decade keep quietly proving that the question is framed wrong.

The industry still argues about it anyway. Every time a former big-tech engineer raises money to fix construction, someone on a job site rolls their eyes. Every time an insider-led startup takes five years to ship v1, a VC mutters that construction people are too close to the problem to solve it. Both sides have receipts.

The Case for Outsiders

Construction’s biggest software unlocks have rarely come from people with a hard hat

Silicon Valley’s argument is simple. The hardest parts of modern ConTech are not construction-specific. They are AI, systems engineering, distribution, and pricing. Those skills are in short supply inside the industry, and long-tenured construction professionals rarely have them. If the moat is technical, the founder has to be technical.

Primepoint’s product thesis backs this up. Bourdev, who worked on computer vision at Facebook before co-founding a video compression company that Apple acquired in 2023, was upfront on the Bricks & Bytes podcast that reading construction drawings is genuinely harder than the problems he solved at Meta. Large language models fall apart on technical drawings because they were trained on captioned photos of the natural world, not linework and symbols. Building the knowledge graph that links every drawing element to its corresponding schedule, spec, and RFI requires a particular type of research talent, and that talent is not sitting in a site trailer waiting to be recruited.

The wins back this up. Tooey Courtemanche grew up around construction and started as a carpenter, but he was also a software founder in Silicon Valley before Procore. His category-defining advantage was the technology background, not the jobsite time. Mo Fawzi at HoloBuilder was a computer science graduate with a basketball coaching side hustle. He built a reality-capture product that became one of the most adopted on construction sites in the world, and Faro acquired it. Altaf Ganihar at Snaptrude came from computer graphics research, not architecture. Bricks & Bytes literally wrote up his story under the headline “The Outsider Redesigning Architecture”.

When BCG Senior Partner Matthias Tauber talks about hiring for his construction and building materials practice, he explicitly deprioritises deep sector expertise. He hires for culture, performance, and connection to the industry. His reasoning, lifted from Foundamental University’s recent masterclass series, is that industry knowledge can be learned but the traits that make a team function together cannot be retrofitted later. Translated to founders: domain knowledge is an asset you can buy. Technical and product instinct usually is not.

The Case for Insiders

Construction keeps eating outsider founders who think it is just another vertical

Now the other side of the ledger, and it is ugly. Katerra raised over $2 billion from SoftBank, Soros Fund Management, and others. Its founder, Michael Marks, had been CEO of Flextronics, one of the largest electronics manufacturing companies in the world. His thesis was that construction could be streamlined the way manufacturing had been. Applied Silicon Valley lessons, acquired architects and contractors, built $150 million factories ahead of demand. By 2021, the company had filed for bankruptcy and employees were told there would be no severance.

The post-mortem was brutal. Architect Magazine’s coverage quoted industry figures describing Katerra’s leadership as “naive” and “hubristic” about a sector they did not understand. The Spokane cross-laminated timber plant cost $150 million and closed two years after opening, operating at a fraction of capacity. Developers wanted to specify products they knew, not Katerra’s alternatives. Buildings, it turned out, are not mass-produced electronics. They are one-offs with local codes, soil conditions, and client preferences that electronics supply chains never have to solve.

PlanGrid is the other bookend. Tracy Young majored in construction management at Sacramento State, worked as a construction engineer at Rudolph and Sletten and Turner Corp, wore hardhats and safety boots, and got shocked by the inefficiency of paper blueprints on site. She and her co-founders, also from construction management backgrounds, built the product as an answer to their own daily pain. They shipped, iterated with customers they already knew, and Autodesk paid $875 million for the company in 2018. Young’s Sequoia profile makes the whole thing sound inevitable in hindsight. It was not. She just understood the buyer, because she had been the buyer.

Foundamental University makes the same point from a different angle. Bricks & Bytes recently summarised 13 masterclasses from the VC’s founders and operators, and one takeaway landed harder than the rest: vertical software companies that skip domain immersion keep losing credibility with construction buyers. Credibility is the scarce resource here. You cannot fake it with a deck.

The Actual Pattern: Hybrid Teams Win, Pure Plays Do Not

Look at the cap tables, not the founder bios

When you line up the winners side by side, the story is not about who led. It is about who was in the room within the first twelve months. Primepoint is outsider-led but insider-advised, with a Webcor project director as VP of Strategy and a founding-team structure that puts construction expertise at the strategy layer. PlanGrid was insider-led but technical-heavy, with co-founders who came from Pixar rendering and high-frequency trading engineering. Procore was both: Courtemanche had construction exposure through family and carpentry, and software founding experience before Procore.

Katerra had none of this. Its top ranks were dominated by outsiders who, per former structural engineer Brian Potter’s post-mortem in Construction Physics, assumed construction-specific problems would yield to the same playbook that had worked in electronics. They hired industry people, but late, and they often did not listen to them. The new ONX Homes venture, founded by much of the same Katerra leadership team, is now openly described as “Katerra 2.0” and runs into similar questions about how much has actually been learned.

Dominik von Achten, CEO of Heidelberg Materials, ran one of the most significant carbon capture retrofits in the industry, integrating a new chemical process into a cement plant that had been operating for over a century. His assessment of why it worked, surfaced in Foundamental’s masterclass series, was that the technology accounted for maybe 30 to 40% of the outcome. The rest was people. The specific trick he credited was blending experienced internal operators with project managers recruited from oil and gas, where carbon capture had already been proven at scale. Cross-industry transfer plus deep sector operators. Neither alone.

What This Means for Buyers, Founders, and LPs

The filter that actually works is cap-table literacy, not founder-background grading

For contractors and owners evaluating ConTech this year, the question is not whether the founder has swung a hammer. It is whether there is credible construction expertise somewhere in the first ten people. Ask to see the team page. If the only person with a hard hat in their bio is in a customer success role, that is data. If there is a named VP of Strategy or Industry who spent real time at a GC, that is different data.

For founders, the playbook is now clear enough that ignoring it is a choice. If you come from construction, your first hire in the first six months should be a senior engineer who has shipped production ML or product at scale. If you come from big tech, your first hire should be a project director from a top-20 GC who will tell you when your product ideas are nonsense, and be in the room when you say them. Both hires are expensive. Both are non-negotiable. The teams that try to cheap out on this half of the founding squad are the ones that spend year two rebuilding product after the first enterprise pilot fails.

For limited partners watching where ConTech capital flows, the filter has to move beyond pedigree-counting. Primepoint’s round was partly underwritten by Navitas Capital, which had backed PlanGrid through the exit and recognised the pattern. Navitas is not funding the Facebook pedigree. It is funding the Facebook pedigree combined with Webcor domain coverage, which is a different animal. That discipline matters more now than ever, because 68% of ConTech venture capital in Q2 2025 went to AI-specific funding. The dumb money will keep chasing founder bios. The smart money will keep reading the cap table.

Outsider vs. Insider vs. Hybrid: The Scorecard

Who won, who lost, and what the team actually looked like on day one

Frequently Asked Questions

For operators, founders, and investors trying to calibrate

Related Articles
Primepoint Raises $10M Seed to Build AI That Actually Reads Construction Drawings
The Outsider Redesigning Architecture
The VC Firm Behind Facebook Just Told Construction: Your Software Is Next