Mowito raises $3M pre-seed to teach factory robots by demonstration
The AMW Read
Incremental funding for a new entrant in Robotics/Physical AI; small round with no structural impact beyond sub-segment.
Mowito raises $3M pre-seed to teach factory robots by demonstration
Physical AI startup Mowito has raised $3 million in pre-seed funding led by Version One Ventures, with participation from All In Capital, Unisol, iSeed, and angel investors including Soumith Chintala. Founded in 2024 by Puru Rastogi, Adityanag Nagesh, and Safar V, Mowito builds software that enables standard industrial robotic arms to learn tasks via human demonstration rather than traditional programming. The company operates in Bengaluru and Detroit, serving automotive and electronics manufacturing clients.
Why it matters: Mowito's approach targets a persistent bottleneck in manufacturing automation: the time and cost of reprogramming robots when production lines change. By replacing hand-coded control logic with learning from demonstration, the startup aligns with the broader shift toward Physical AI, where models perceive and act in the real world. This fits the 'fastest ARR ramp' pattern in robotics, where software-defined flexibility can unlock faster deployment cycles in structured environments like factories.
Grounded expert take: At $3M pre-seed, this is a small bet on a well-defined vertical — manufacturing robotics — but it signals growing investor appetite for Physical AI applications that bypass the complexity of general-purpose humanoids. The involvement of Soumith Chintala (PyTorch creator) adds technical credibility. The real test will be whether Mowito can achieve repeatable cross-customer deployment without custom integration, a classic challenge for industrial AI startups. If successful, it could contribute to the 'context-engineering moat' pattern by building proprietary demonstration datasets.