Gritt launches with $32.4M for construction robotics retrofits

Why it matters

Gritt's retrofit model could scale construction robotics without forcing contractors to replace their fleets. Its next test is expanding from two field systems to dozens while preserving jobsite reliability.

Puneet Puri and Vishal Dugar brought Gritt out of stealth on July 21 with $32.4 million in funding and a practical approach to construction robotics: add robotic arms and AI to the equipment contractors already own.

The San Francisco-based Gritt mounts its systems on common jobsite machines, including skid steers and forklifts. Gritt is starting with the repetitive work of moving and placing panels on utility-scale solar farms, then plans to expand into assembly, material transport, verification and site planning.

The funding consists of a $26 million Series A led by Obvious Ventures, with Union Square Ventures and Active Impact Investments participating, according to Gritt's launch announcement on Business Wire. Previous investors First Round Capital, Climactic, Congruent Ventures and VSC Ventures also backed the financing.

Gritt's announced total implies $6.4 million of earlier capital before the Series A. Gritt did not disclose a valuation, revenue, pricing or the value of its signed contracts.

Two roboticists take autonomy outdoors

Puri's argument starts with the environment. Factory automation works inside spaces designed for machines, where lighting, floors and production sequences can be controlled. Construction crews contend with dust, mud, snow, uneven ground and site plans that change while work is underway.

"You can't tell a construction site to be less muddy or to stop snowing," Puri said in the announcement.

Puri came through Carnegie Mellon University's Robotics Institute and Stanford. Gritt's team page describes him as a serial robotics founder who developed inspection drones for industrial systems. The claim that those were the world's first such drones comes from Gritt and is not independently established in the launch materials.

Dugar is a Carnegie Mellon Robotics Institute alumnus who later worked at autonomous vehicle developer Aurora. Gritt says he has deployed autonomous trucking fleets and self-piloting aircraft.

Their backgrounds explain Gritt's chosen technical problem. Dugar is applying motion planning and autonomy developed for vehicles and aircraft to machines that need to manipulate heavy objects precisely, around workers, on changing outdoor sites. Gritt says its arms can pick, place, assemble and transport materials with millimeter precision.

The founding group extends beyond the two co-founders. Gritt lists Anirudh Vemula, a Carnegie Mellon robotics researcher, as head of AI and MIT mechanical engineer Felipe Roz as head of mechanical engineering.

Gritt had also spent time testing its solar thesis before the public launch. The U.S. Department of Energy named Gritt Robotics one of two winners of the American-Made Solar Prize's seventh round in September 2024. The award description identified the same core idea: turning off-the-shelf construction equipment into intelligent robots for utility-scale solar work.

The retrofit model is the commercial bet

Construction robotics companies face a hardware problem alongside the autonomy problem. Purpose-built machinery takes capital to design and manufacture, and contractors must learn how to operate and maintain an unfamiliar fleet. Gritt is trying to remove that adoption hurdle by mounting its arms, sensors and software on machines already present at a jobsite.

That choice could let Gritt add deployments faster and with less manufacturing risk than developers of dedicated solar-installation robots. It also ties Gritt's reliability to hardware from outside suppliers and requires the autonomy stack to work across equipment, panel and tracker configurations.

Business Wire says Gritt has autonomously placed tens of thousands of solar panels with zero breakages. TechCrunch reported that Gritt is contracted to install 2.8 gigawatts of solar capacity over the next 18 months, while Gritt's site claims a fourfold efficiency gain with no additional labor.

Those operating figures are self-reported. Gritt has not named the contractors behind them or published the methodology behind the efficiency comparison. Gritt says top-10 U.S. solar engineering, procurement and construction contractors are customers, while its public testimonials identify customers only by job title.

TechCrunch reported that Gritt currently has two systems in the field and plans to operate about 48 within six months. TechCrunch also spoke with an unnamed customer who expected Gritt to make remote projects easier to staff and reduce injuries caused by repeatedly lifting heavy panels.

Moving from two field systems to roughly 48 would be the first large test of the retrofit strategy. Solar sites can contain hundreds of thousands of repeated placements, giving Gritt a concentrated task and a large stream of training data. Each successful placement can improve perception and motion planning, while failures carry the risk of damaged panels, stalled crews or safety incidents.

Gritt says it has placed tens of thousands of panels without breakage. That record remains Gritt's claim, although the disclosed deployment numbers show the machines have moved beyond laboratory demonstrations.

Solar is the wedge, site intelligence is the larger product

Gritt is initially automating panel pick-and-place. Puri and Dugar want the same system to fasten panels, assemble supports, transport components and eventually handle construction tasks such as rebar work. Gritt also plans to use the data collected by its machines to track completed work, material movement and site conditions.

That data layer could become as important as the robotic arm. Construction supervisors coordinate workers, equipment, deliveries and changing schedules with an incomplete view of what has happened across a large site. Machines already performing the work can create a live record of progress and recommend what should happen next.

Gritt still has to prove that intelligence learned from a repetitive solar installation transfers to concrete, rebar, data centers, bridges and roads. Those jobs involve different materials, tools, tolerances and safety requirements. The retrofit architecture gives Puri and Dugar a way to enter each market without designing an entirely new vehicle, but the manipulation software must still master each task.

The Series A gives Gritt room to pursue that expansion while scaling its solar deployments. The immediate benchmark is simple: turn a small field fleet into dozens of working systems without losing the reliability that contractors require. For Puri and Dugar, the hardest part of physical AI begins after the demonstration, when a robot has to perform the same movement thousands of times in mud, heat and wind while the rest of the jobsite keeps moving.