Construction · Deep dive

TerraFirma

Two ex-SpaceX engineers building a vertically integrated robotic earthworks contractor — retrofitted semi-autonomous excavators run from a mission-control room, sold to customers as ordinary construction bids, not as robots.

emerging

The question that decides it: TerraFirma booked venture money to be the contractor, not the vendor: it bids earthworks jobs and keeps the productivity gain from one operator supervising 3-5 retrofitted machines. Does that operator-leverage ratio actually convert into winning bids at better-than-industry margins fast enough to outrun Bedrock Robotics — which raised $350M+ at a $1.75B valuation (Feb 2026) to sell fully operatorless excavation to every incumbent contractor — or does vertical integration just mean venture capital subsidizing a thin-margin, working-capital-hungry dirt business that scales like a subcontractor, not like software?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

TerraFirma is a two-year-old Austin construction company that does excavation, grading and site work with retrofitted heavy equipment its own engineers converted into semi-autonomous robots, supervised three-to-five machines per operator from a mission-control room. Founded in 2024 by two ex-SpaceX engineers — both named Noah — it announced roughly $115 million in total funding on July 14, 2026, including a $100 million Series A led by Kleiner Perkins, one of the larger Series A rounds in construction tech. The wedge claim: one operator moves at least 300 percent more dirt per shift. The strategic claim is sharper — TerraFirma is not selling the technology at all. It bids and executes construction projects itself, which means its revenue is construction revenue, its data is proprietary, and its scaling problem is a contractor’s, not a software company’s.

Founding story

Noah Schochet and Noah McGuinness met on the first day of engineering classes at Princeton, where they were robotics lab partners; by their own count they have worked together for over a decade. Both went to SpaceX — McGuinness onto Starshield, the classified government variant of Starlink; Schochet first as a lead engineer on the Starlink Wi-Fi router, where he built out an automated line of some 30-40 robot cells and, in his phrase, automated himself out of a job, then onto Starship as one of three-to-five engineers on the rocket’s payload volume.

The company was conceived at Starbase — a rented circus tent for a factory, javelina hogs on the production floor, an office in a moldy shipping container (MCJ podcast, Jun 2026). The formative contrast: SpaceX built skyscraper-sized rockets in a month or two while contractors took six months to add a parking lot, and GCs quoted three to four years to build out the Starbase site. The origin moment, as Schochet tells it: at a post-launch party, Elon Musk, asked what goes inside a Mars-bound Starship, answered without hesitation — construction robots, some two trillion dollars of infrastructure. Schochet asked to run that program, was told no, and left to attack the Earth-side version. The founders bootstrapped from a crumbling 1900s ranch house renting at $600 a month before raising institutional money.

How it works

TerraFirma buys or rents standard heavy equipment — excavators, dozers, loaders, rollers, skid steers — and clamps on a retrofit kit (Schochet’s own metaphor: the facehugger from Alien) that converts the machine for remote and semi-autonomous operation. Operators sit in a command center Schochet compares to a real-time strategy game like StarCraft: each has a multi-screen battle station and is assigned three to six machines depending on job complexity.

The core interface is called click-to-dig: an operator sketches the required work as a 3D model in about 60 seconds — dig here, dump there — and the machine executes autonomously for 20-plus minutes. Repetitive sequences become macros: a roller can be told to follow the dozer at a six-foot standoff and compact in three passes overlapping 1.5x. The operator’s role shifts from joystick work to something closer to a CNC programmer. For edge cases and un-automated tasks, the fallback is deliberately mundane: an Xbox controller for direct teleoperation. Safety runs on a digital heartbeat — each machine continuously receives a signal from TerraFirma’s safety system, and any interruption (a dropped network link, for instance) triggers an emergency stop that only a manual reset clears (Inc, Jul 2026). The claimed labor math: 100 operators in 100 cabs becomes 33 people orchestrating 100 machines, then multiplied across three shifts because remote operators in Austin or Dallas can hand off around the clock.

Product and business overview

The company describes a full stack of three named layers (Business Wire, Jul 2026): AI-enabled pre-construction software (estimating and planning that lets it bid faster and cheaper), the remote command-and-control center, and the retrofitted semi-autonomous fleet. But the product a customer actually buys is none of these — it is finished earthwork: excavation, grading, land clearing, demolition and site preparation, currently across housing, energy, transportation, manufacturing and education projects in Texas. TerraFirma operates as a subcontractor to general contractors — Schochet has explicitly corrected interviewers who called it a GC. A second, quieter line of business is government work: the company says it is executing mission-critical international infrastructure and logistics projects for the U.S. government in challenging operating environments, and cites wildfire cleanup, radioactive-site remediation and disaster response as categories where remote operation has no viable alternative. The Series A cap table — angels from SpaceX, Anduril and Hadrian — signals the defense-adjacent ambition. The stated long-term product is construction on Mars, which functions today mostly as recruiting narrative.

Business model and pricing

Revenue is booked as construction contracts won through competitive bidding, not licenses or subscriptions; there is no pricing page because the price is the bid. Schochet is blunt about why (MCJ, Jun 2026): a robotics company is a bad business, and nobody in construction wants to buy a robot — they want to buy an outcome. Contractors have no R&D budgets and no tolerance for debugging. By keeping the interface to the industry identical — a bid, a schedule, a finished site — TerraFirma pockets the spread between industry-standard pricing and its own lower internal cost. Asked directly whether it underbids rivals and keeps better margins, Schochet answered yes, but no margin figures have been disclosed, and he separately concedes construction’s profit margins are very slim. The structural catch of subcontracting is unchanged by robots: subs front the working capital and get paid when the GC gets paid.

Traction over time

No revenue, fleet-size or absolute headcount figures have been disclosed. What is on the record:

The unverifiable claims to flag: Schochet says TerraFirma has beaten the timeline and cost on every project it has taken (Jun 2026), and that it likely has the most revenue and largest robotic fleets of any construction-robotics player — plausible given rivals sell kits rather than book construction revenue, but no third party has confirmed either.

Market analysis

The macro argument is unusually well-documented. U.S. construction labor productivity has declined about 0.6% a year since 1965 while the broader economy gained roughly 1.6% annually — a divergence the company (citing industry research) prices at about $1 trillion over the past five years. Associated Builders and Contractors estimates the industry must attract 349,000 additional workers in 2026 just to balance supply and demand, after ~454,000 in 2025; 87% of contractors report difficulty finding qualified equipment operators, up from 68% three years earlier (industry surveys, 2025). The BLS counts only ~417,000 construction equipment operators nationally (2026), while data centers, reshored factories and grid buildout inflate demand. The global earthmoving equipment market was ~$75.5B in 2025, headed to ~$141B by 2034 (Fortune Business Insights) — and equipment is only the proxy; the earthworks services market TerraFirma actually bids into is a large slice of the roughly $2 trillion U.S. annual construction spend. Structurally, the shortage is the forcing function: automation is arriving because the cabs are empty.

Competitive intel

See the competitor table for detail; the shape of the field matters more than any one rival. Bedrock Robotics (Waymo alumni, $350M+ raised, $1.75B post-money as of Feb 2026) is betting the opposite architecture — full autonomy, sold as a kit to existing contractors, operatorless deployments targeted for 2026. Built Robotics ($137M raised since 2016) tried the broad kit model and retreated to a narrow, repeatable niche (solar piling), which is the strongest available evidence that TerraFirma’s own-the-project model may be the right lesson learned from a decade of construction-robotics disappointments. Teleo, architecturally TerraFirma’s twin, raised only ~$30M and was absorbed by defense startup HavocAI (Feb 2026). Caterpillar and Komatsu already sell remote operation and have run autonomous mine haulage for a decade; their dealer networks and warranty terms are a standing threat to any third-party retrofit. And on every actual bid, the competition is a local excavation sub with paid-off iron and 40 years of GC relationships. Schochet himself concedes the retrofit kit is replicable — several startups already do it — locating the moat instead in the operating company: sites, fleets, reputation, and data from real jobs.

History and evolution

2024: founded in Austin after the Noahs leave SpaceX; bootstrapped period in the ranch house. 2025: first retrofits and first commercial Texas earthworks; early government work begins. Early 2026: category heats up around it — Bedrock’s $270M Series B (Feb), Teleo’s acquisition (Feb). Jun 25, 2026: Schochet records the MCJ Inevitable podcast, the fullest public account of the model. Jul 14, 2026: $115M announced; hiring plan of ~300, Texas factory and mission-control center disclosed. No pivots yet — the notable evolution is emphasis: the early website led with Mars; the Series A materials lead with critical infrastructure and government work. Too young for a crisis; also too young to have survived one.

What people say

The case for. Investors supply the loudest praise, but with specifics. Kleiner’s Josh Coyne says the company is succeeding at real-world scale, proving the business model, and securing government and commercial contracts (Jul 2026). BCV’s Ajay Agarwal, an early investor, says he personally watched one operator run three machines in parallel (Inc, Jul 2026). Trade press (The Robot Report, Construction Equipment, Equipment World, Jul 2026) covered the raise as validation that the operator-out-of-cab model has arrived. The recruiting pitch appears to be landing: 60%+ of engineering hired from SpaceX/Tesla-tier companies.

The complaints. TerraFirma is too young for G2, Capterra or a meaningful Glassdoor record — a gap in itself, since every operational claim currently traces to the company or its investors. The sharpest criticism comes from Schochet’s own mouth: customers are largely indifferent (“a local minimization… everyone is kind of happy” with agreed prices and timelines, Jun 2026), meaning demand for 3x-faster earthworks must be created, not harvested; the retrofit kit is admittedly not defensible; and full autonomy is, in his words, something no one is capable of today — a direct shot at Bedrock’s 2026 operatorless promise, but also a concession that TerraFirma’s labor leverage is capped near 3-5x. The vertically integrated model inherits everything ugly about contracting: crews, deadlines, weather, working-capital cycles, slim margins, bid-by-bid growth (a point Inc made gently, Jul 2026). Labor is a live risk: the IUOE chose partnership with Built Robotics, but trade unions have grown publicly alarmed about blue-collar automation (Futurism, 2026), and TerraFirma’s non-union Texas footprint will not travel unexamined to union states. The 300%-more-dirt figure and the beat-every-bid record remain unaudited marketing until a GC says them on the record.

Outlook: the open question

TerraFirma works if winning bids compounds — if every job makes the software better, the bids sharper and the margins fatter, so that by the time rivals’ autonomy kits mature, TerraFirma is a large, data-rich contractor rather than a small robotics vendor. For that to be true, three things must hold: the 3-5x operator leverage has to survive contact with messy, varied job sites at consistent quality; construction revenue has to grow faster than the contractor cost structure underneath it (Texas factory, 300 hires, owned fleets); and government/hazardous-environment work has to mature into the high-margin anchor the commercial market’s admitted apathy cannot yet provide. It fails if the vertical integration is the trap — if earthworks stays a lowest-qualified-bidder market where TerraFirma’s edge prices away into customer savings, while Bedrock’s $350M sells operatorless capability to every incumbent sub, turning TerraFirma’s differentiator into an industry default it paid to pioneer. The honest tell will be visible within 18-24 months: either named GCs and agencies start awarding it large repeat contracts outside Texas, or the fleet stays small, the projects stay Starbucks-pad-sized, and the Series B pitch quietly becomes a technology-licensing story — the exact business Schochet says is a bad one.

How a challenger would attack it

Arm the incumbents TerraFirma has to outbid. TerraFirma’s moat, by its own CEO’s admission, is not the retrofit kit — “several startups already do it” — it is the operating company. So the attack is Bedrock’s, executed faster: sell the operator-leverage capability to the thousands of local earthworks subs who already own paid-off iron and 40 years of GC relationships. Every sub equipped with a $100K kit that lets one operator run three machines erases TerraFirma’s bid advantage on that job, and the sub keeps its working-capital relationships, bonding capacity, and union standing — the things a two-year-old Austin subcontractor lacks. Price it as revenue-share on labor savings so contractors with no R&D budget carry no risk. The second front is labor politics: TerraFirma is non-union Texas, and the IUOE has already shown (via Built Robotics) that it partners with vendors who train operators rather than replace them — a challenger that signs union training agreements locks TerraFirma out of union states and public megaprojects before it arrives. Third, hit the concentration: TerraFirma’s disclosed commercial record is Starbucks-pad-sized Texas jobs plus opaque government work; a rival that publishes audited per-yard cost data against the unverified “300% more dirt” claim forces TerraFirma to defend marketing with numbers it hasn’t shown.

Same playbook, new buyer

TerraFirma’s real invention isn’t the excavator kit — it’s the model of venture-funded vertical integration: be the contractor, keep the productivity spread, sell outcomes to an industry that won’t buy robots. That playbook transfers to adjacent dirt-adjacent trades with the same empty-cab problem and even less automation attention: utility trenching and boring for grid buildout, aggregate hauling, mine-site earthworks outside the OEM-automated megamine tier, and demolition — each a lowest-qualified-bidder market where operator leverage converts directly into margin. The sharpest single shift is the one TerraFirma’s own materials gesture at but underweight: hazardous-environment work as the primary business, not a category mention. Wildfire cleanup, radioactive remediation, and disaster response are markets where remote operation isn’t 3x cheaper — it’s the only option — meaning pricing power instead of bid-sheet parity, and government buyers instead of apathetic GCs. A challenger that builds the FEMA/DOE/DoD-first version captures the margin anchor TerraFirma treats as a sideline, and TerraFirma can’t easily pivot: its 300-hire, Texas-factory, commercial-fleet buildout is committed to winning ordinary earthworks bids at scale.

Sources and further reading

• TerraFirma Raises $115M to Accelerate Construction on Earth and Beyond — Business Wire, Jul 14, 2026
• SpaceX alumni building remote-controlled construction equipment land $115 million round — CNBC, Jul 14, 2026
• This Startup Says 1 Worker Can Do the Work of 5 — Inc., Jul 15, 2026
• Faster, Safer, Cheaper: How TerraFirma Is Rebuilding Construction — MCJ Inevitable podcast (Noah Schochet), recorded Jun 25, 2026, published Jul 28, 2026
• TerraFirma raises $115M to build robotic infrastructure for construction — The Robot Report, Jul 2026
• Former SpaceX engineers raise $115m for tech that removes operators from cabs — Construction Briefing, Jul 2026
• Bedrock Robotics Raises $270 Million in Series B Funding — PR Newswire, Feb 2026
• Teleo acquired by HavocAI — PitchBook profile, Feb 2026
• ABC: Construction Industry Must Attract 349,000 Workers in 2026 — Associated Builders and Contractors, 2026
• Earthmoving Equipment Market Size — Fortune Business Insights, 2025