Robots are installing solar panels at 4x the human rate, and a stealth startup just raised $34M to scale it

Gritt's robots install solar panels at four times the human rate.

The world is building solar at a record pace. Factories are churning out panels. Prices keep falling. And yet the pace of actual deployment keeps hitting a wall that has nothing to do with silicon or subsidies: there are not enough trained workers willing to haul hundred-pound glass panels onto metal frames in the desert sun.

Gritt, a startup founded by Carnegie Mellon-trained roboticists Puneet Puri and Vishal Dugar, exited stealth on Tuesday to attack exactly that problem. The company has built AI software that controls off-the-shelf construction equipment to automate solar panel installation. CEO Puri says a typical eight-person crew installs about 800 panels per day. The same crew, augmented with Gritt's systems, can install 3,000 to 4,000 1.

That is roughly four times the throughput, from the same headcount.

The implication extends well beyond one company's revenue. If those numbers hold at scale, the binding constraint on utility-scale solar deployment shifts from labor availability to equipment rollout. The energy transition timeline stops depending on recruiting enough people for dangerous, repetitive physical work and starts depending on something capital and engineering can actually solve.

Renting the arms, owning the brains

Gritt's approach differs from competitors in a way that matters for cost structure. Rather than designing custom robots from scratch, the company uses rented skidders and robotic arms built by companies like Kawasaki, then layers its own AI models on top 1. The systems unload large glass panels, carry them toward installation frames, and position them with sub-millimeter accuracy so human workers can fasten them in place.

Rivals like Luminous Robotics, Cosmic, and China's Trinabot are each building proprietary hardware instead 1. That distinction could determine who scales faster. Off-the-shelf components are cheaper and available now, but custom hardware might eventually deliver tighter integration. Gritt's bet is that speed of deployment matters more than optimization, at least in the near term.

The $26 million Series A was led by Obvious Ventures, with participation from Union Square Ventures and Active Impact Investment 1.

2.8 gigawatts already on the books

Gritt currently has two systems deployed in the field, collecting data to improve their performance 1. The company says it is contracted to help install 2.8 gigawatts of solar capacity over the next 18 months, with customers that include three of the top 10 U.S. power construction companies. It plans to have 48 systems operating within six months.

A customer who spoke to TechCrunch on condition of anonymity said the system should make it easier to operate at remote sites where attracting workers is difficult, and should reduce injuries by eliminating the need for crews to repeatedly lift 100-pound panels overhead 1.

Why AI changes the math for construction robotics

Industrial robots have long been reliable on factory floors, where every variable is controlled. Construction sites are the opposite: uneven terrain, weather, shifting layouts, unpredictable materials. Previous generations of construction robotics struggled with that unpredictability.

Puri argues that newer AI models have changed the equation. He notes that training the system to stack cinder blocks took weeks, while training it to tie rebar took a single day, because the same underlying software pipeline could be reused across tasks 1. The founders envision their systems eventually performing tasks like drilling posts, building racks, and tying rebar before concrete pours. They also see the sensor arrays as a management layer: detecting open trenches before storms arrive, flagging missing inventory, feeding real-time data back to project managers 1.

The broader bet

Solar installation is Gritt's entry point because the task is physically demanding, the labor shortage is acute, and demand is guaranteed by policy and market forces. But the founders are not building a solar company. They are testing whether general-purpose AI, layered onto commodity hardware, can finally make construction robotics viable in messy, unstructured environments where every prior generation has struggled.

The 2.8 gigawatts of contracted work signals that customers are willing to run it under real conditions. Whether Gritt's off-the-shelf approach outpaces competitors building custom hardware is an open question. But the problem it is solving is not hypothetical. The solar buildout is underway. The open variable is whether robots can meet the demand that humans cannot.