Mowito’s physical AI helps standard industrial robot arms learn new factory tasks by observing people instead of relying on manual programming. The Bengaluru startup has raised a $3 million pre-seed round (₹28.6 crore) led by Version One Ventures, with participation from All In Capital, Unisol, and iSeed. The company aims to reduce the time and cost factories spend on rigid setups and custom programming for inspection and assembly tasks. It also takes too much specialist labor. Founded in 2024 by Puru Rastogi and Adityanag Nagesh, Mowito is building AI software for manufacturing lines in India and the US.
What does Mowito physical AI actually do?
At the center of the product is NeuralPick, Mowito’s software stack for industrial robot arms. Customers configure the application through a web interface and upload site-specific data. The model then learns the parts and conditions on that manufacturing line. Then it deploys the robot with ongoing cloud updates instead of starting another long programming cycle from scratch. It’s built to run on standard robot arms, not custom hardware that forces a buyer to rebuild the whole cell.
And the details matter here. NeuralPick combines computer vision and force sensing. This lets robot arms see, feel, and handle parts without relying on fixed jigs or rigid fixtures. The company pitches rapid line reconfiguration and browser-based training. It also promises operator-friendly controls instead of a setup that only an outside robotics integrator can maintain. Mowito says its system can achieve ±200 micron accuracy for inspection and assembly. It also adapts to moving conveyors, making it more practical than generic AI factory solutions.
The use cases are concrete, not theoretical. In one automotive case study, Mowito says a robot with one arm-mounted camera replaced eight fixed cameras and reduced inspection costs by more than 60%. It checks for defects, missing parts, and dimensional errors across multiple product variants. In another case study, Mowito says its software helped a partner achieve sub-100 micron accuracy on a live smartphone assembly line in India.
That gives you the “before versus after” pretty fast. Before Mowito, a factory might need multiple camera stations and rigid part presentation. It might also need manual recoding every time a component changed. After Mowito, factories can use one adaptable robot cell to inspect, assemble, tend machines, and handle wires or connectors with less custom rework. It’s still hard robotics. But it’s a cleaner software pitch than most automation vendors manage.
Who built Mowito physical AI and what’s the early traction?
Mowito was founded in 2024 by Puru Rastogi and Adityanag Nagesh, and it now operates across Bengaluru and Detroit. The company’s thesis is that industrial robots should be taught on the floor by operators and demonstrations, not treated like brittle systems that need months of specialist integration every time the job changes. That’s why its early focus has landed in automotive and electronics manufacturing, where even tiny variations in parts or positioning can wreck throughput.
Why Puru Rastogi fits the technical side
Rastogi’s background is about as robotics-native as you’d want for this kind of company. His public profiles link him to Carnegie Mellon University’s robotics program, the Robotics Institute, and Near Earth Autonomy, where he worked on motion planning, mapping, and localization for aerial systems. He’s also been linked to CleanRobotics and the TrashBot project. This experience shows he has built robots for real-world environments, not just controlled demonstrations.
Where Adityanag Nagesh brings edge
Nagesh looks like the more commercial cofounder. His public profiles and company records link him to Insieve Technologies and Sensara Technologies. His LinkedIn profile also highlights his robotics and manufacturing background at Mowito. That mix matters because startups like this don’t fail only on tech. They fail when nobody can translate the tech into a buying decision inside a real factory.
What traction is visible already
The startup isn’t selling a science project. Mowito says its robots are already running on manufacturing lines at a Fortune 500 automotive company and at one of the world’s largest electronics contract manufacturers. It also says it supports customers from Bengaluru and Detroit. That’s a smart split: India gives it engineering depth and a growing manufacturing base, while Detroit puts it close to US auto buyers and suppliers that actually spend on automation.
How the Mowito funding round came together
The new Mowito funding round is a $3 million pre-seed led by Version One Ventures, with participation from All In Capital, Unisol, and iSeed. Angel backers include Thinking Machines Lab CTO Soumith Chintala, Foundry Robotics founder and CEO Adarsh Kulkarni, Coformer.ai cofounder and CEO Ashish Kulkarni, and Better Capital founder and CEO Vaibhav Domkundwar. The money is earmarked for a US expansion, engineering and go-to-market hiring, broader deployments across automotive and electronics manufacturers, and continued development of the physical AI platform.
Where Mowito sits against rivals and old-school alternatives
Mowito isn’t entering an empty category. The legacy alternative is still the same old mix of fixed camera inspection and rigid fixtures. It also includes system integrators and robot programming that can take weeks or months to modify for each new part family. The newer overlap comes from several directions at once: Wandelbots made its name by letting non-programmers teach robots new tasks, Vention sells code-free cloud programming and digital-twin deployment, Standard Bots pushes AI-native industrial robot arms with real-time recording, and Mujin offers a no-code robotics platform that works across brands. Mowito’s differentiation is tighter and more specific: it wants standard industrial arms to learn from operator demonstrations, then adapt on the line with vision and force feedback for inspection, assembly, and machine tending rather than selling a whole replacement robot stack.
Why Mowito’s physical AI round matters now
A $3 million pre-seed round doesn’t sound huge in a market that loves giant robotics headlines. But for this kind of company, it’s enough to matter. Mowito already has live industrial proof points, so this round is less about inventing the category and more about proving it can repeat deployments, ship reliably, and shorten the sales cycle in the US. That’s a different problem. And it’s the one that separates interesting robotics startups from actual businesses.
The use of funds tells you what investors think the bottleneck is. Mowito isn’t talking about building flashy new humanoids. It’s putting money into US expansion and engineering depth. It’s also hiring for go-to-market roles and funding more rollouts inside automotive and electronics factories. That suggests Version One and the angel group are backing a practical thesis: there’s room for a software layer that makes existing robot arms more useful before factories go shopping for entirely new machines.
Rastogi summed up the pitch neatly: “We believe robots should learn the same way people do: by observing and repeating. This funding allows us to accelerate that vision, expand globally, and bring Physical AI to more manufacturing environments.”
How big is the industrial robotics market Mowito is chasing?
It’s big enough that even a narrow slice could build a serious company. Grand View Research estimates the global industrial robotics market was worth about $33.96 billion in 2024 and projects it will reach roughly $60.56 billion by 2030. The International Federation of Robotics has also said more than 4 million robots are now operating in factories worldwide, with annual installations above 541,000 units in 2023 — the second-highest level on record.
Why now? Because the enabling stack is finally catching up. Deloitte’s 2026 outlook points to better compute and newer robot AI models such as vision-language-action systems and world models. It also points to heavier investment from large tech and robotics players as the forces likely to accelerate adoption through 2030. That doesn’t mean factories suddenly want “AI” as a slogan. It means buyers are more willing to pay for robots that can handle variation without a fresh integration project every quarter.
Final take on Mowito physical AI
Mowito physical AI is interesting because it’s chasing a boring problem in the best possible way. No grand robot mythology. Just a hard claim that existing industrial arms should be easier to train, faster to redeploy, and less painful to own.
If Mowito can turn those first automotive and electronics wins into repeatable US deployments, it won’t just look like another robotics startup from Bengaluru. It’ll look like a credible factory software company with hardware-level consequences.
Read how Norm AI raised a $120M Series C led by Khosla Ventures to build AI-native legal and compliance systems that combine agentic AI with attorney oversight through its affiliated law firm.
FAQ
- What funding did Mowito raise? Mowito raised $3 million in a pre-seed round, which the source article values at about ₹28.6 crore. Version One Ventures led the round, and the cap table also includes All In Capital, Unisol, iSeed, and angel investors such as Soumith Chintala and Vaibhav Domkundwar.
- How does Mowito’s physical AI work on industrial robot arms? It works by letting factories configure a job in a browser and train the model on site-specific data. Then they deploy it on standard robot arms with ongoing software updates. Mowito’s NeuralPick stack combines vision with force sensing so the robot can inspect, assemble, and handle parts with less dependence on jigs, rigid fixtures, or manual recoding.
- Who are the founders of Mowito? Mowito was founded in 2024 by Puru Rastogi and Adityanag Nagesh. Rastogi brings a deep robotics background shaped by Carnegie Mellon and Near Earth Autonomy, while Nagesh adds company-building experience and a business-side view of how industrial automation gets sold and deployed.
- Is Mowito a manufacturing robotics company or a warehouse robotics startup? Right now, it’s best described as a manufacturing robotics and physical AI startup. Its visible traction and case studies are in automotive and electronics production, especially inspection, precision assembly, and machine tending, even though some older profile pages around the company also reference warehouse automation roots.
