One industry’s productivity has been falling for fifty years. Origin went there.

When Yogesh Ghaturle was in eighth grade in Hinganghat, a teacher moved to his town from Hyderabad. Her husband had been transferred to a manufacturing plant nearby. Before she arrived, nobody at the school, not in anyone’s memory, not in anyone’s ambition, had gone to IIT. She asked the class what colleges they were hoping to attend. Everyone named colleges in Nagpur.

She asked if they knew about IITs.

They didn’t.

That question didn’t just open a door. It told Yogesh something he’s been running on ever since: the ceiling you see is the ceiling of the people around you. The moment someone showed him a different one, the old one stopped working.

The front page

Yogesh used to read newspapers. What he noticed was that the people who made the front page were always one of four types: politicians, cricketers, Bollywood celebrities, or entrepreneurs. He had represented Vidarbha at the under-16 Ranji level, cricket was real, but grassroots cricket from a small town was “filled with politics, not very meritocratic.” Politics was out. Bollywood was obviously out.

Entrepreneurship was left. “Very meritocratic,” he says. “The destiny is in your hand.”

He got to NIT Karnataka for computer science, started his first company in third year, raised from Kalaari Capital, built POS solutions for offline retail, pivoted once, and sold to Amazon in 2021.

What the Amazon years gave him was a checklist. Perpule’s edge had been speed of execution and pricing, not technology. Anyone could replicate it. The next company would have to be deeply technical, something that couldn’t be copied by running faster. Second: Perpule was India-only until too late. The pricing power outside India is at least five times higher. The next company would be global from day one. Third: Perpule had built a slice of what Square does. Square built all of it and became a $30 billion company. The next company would go full-stack.

He took those three criteria and went looking for an industry where they all applied. Construction robotics came out of it.

The only industry going backward

Construction is the only major industry where productivity has actually declined over the last fifty years. Not stagnated. Declined. Manufacturing, agriculture, mining, all mechanized, increasingly automated. The best car plants in the world run at 90% robot labor. US agriculture feeds 99% of the country with 1% of its workforce. Construction: barely mechanized. Excavators and cranes for the exterior shell, and that’s roughly where it ends.

Interior construction: wall finishing, painting, plumbing, tiling, flooring, is still almost entirely done by hand. This is where most of the labor in construction sits, and in the US it’s where the shortage is sharpest. A younger workforce that won’t enter skilled trades. Labor costs rising 4-6% a year. Hundreds of thousands of unfilled openings. The problem gets structurally worse as each cohort ages out, and it’s happening exactly when the AI buildout, data centers, power plants, manufacturing facilities, is generating the largest infrastructure construction cycle in decades.

Origin builds robots for interior construction. The first product finishes walls: compound application, sanding, priming, painting, one robot, four tasks, on a single modular platform. It deploys at $20,000 a month per robot, robot-as-a-service, to general contractors in the US. Four robots are active in New York. Over 100,000 square feet delivered. 5,000 hours of construction work completed.

He didn’t rush to announce it

The day Perpule’s Amazon acquisition was announced, Yogesh didn’t post about it on LinkedIn until late that evening. His co-founders had to nudge him. His explanation, direct: “I knew about it. The people I care about also knew about it. So who cares?”

Somewhere around 2014-2015, when he was in college, he’d written a list, things he wanted to do in the next four or five years. Starting a company. Getting featured in YourStory. Forbes 30 Under 30. He made the list, set it aside, and didn’t look at it again. “Somehow those things happen,” he says. He doesn’t explain it as luck. He explains it as the subconscious running a program he’d set.

The Forbes 30 Under 30, the Amazon exit, has a single line for both of them: “Forbes 30 Under 30 would hopefully be a footnote in my current list.”

The 100-year company framing isn’t rhetorical. It’s why he spent two years after Amazon figuring out what to build next instead of starting fast. He wanted something technically deep enough that the moat isn’t execution speed. Something that prices globally. Something where controlling the full stack means controlling the full margin. He was looking for the industry where those three things intersected, and where the problem had been waiting longest for someone to go and find it.

5,000 interviews

Origin’s two other co-founders came from Yogesh’s network before they came from a pitch. Tushar Makkar, who leads software and AI, is a college senior from NIT Surathkal. They’ve known each other for twelve years. Ashish Daga, who leads hardware, is a school friend who went to IIT Kharagpur. Both are people Yogesh had kept in touch with for years before they built anything together.

That’s the first principle. The second is the volume.

To build a 45-person full-time team in deep-tech robotics in India, Origin has run approximately 5,000 interviews. Yogesh says this without complaint: “It’s like a one-person conversion rate.” The roles are niche: hardware engineers who understand construction conditions, AI researchers who understand physical environments, people who can work across simulation and job sites. The bar is high because the problem demands it.

The 5,000 interviews is not a number he offers as a hardship story. It is his answer to the question of how you build a team for something genuinely hard. You run the process until you find the people who can do it. In a market where most of the hiring infrastructure is built for software, and where the category you’re building is contested globally by companies with 50 times your funding, this is what the work actually looks like.

The ceiling, again, is only the ceiling of the people around you. You find different people.