Hey everyone! We’re Douglas, Raphaël, Maëlic and Tristan, the founders of Asim (YC F26).
TL;DR
We’re building standardized benchmarks for physical intelligence: how well robots handle objects, coordinate movements and adapt when conditions change. We test real robots to help engineering teams measure their capabilities, understand their limitations and identify safety risks before deployment.
https://www.youtube.com/watch?v=mwA0V_8gfnU
The problem
A robot picks up a glass perfectly in the lab. Then the glass is heavier, the surface is slippery, or someone reaches into its workspace. Does it still behave safely?
Completing a task in one setting doesn’t tell us how reliably a robot will perform when conditions change. A success-or-failure score alone doesn’t explain what caused a failure or how to prevent it.
As robots increasingly operate around people, understanding those limits is critical to safety.
Our solution
At Asim, we systematically vary objects, tasks and environments in the real world to find the conditions where robot behavior breaks down.
Our testing examines three things:
Robotics teams get actionable feedback on their robots’ limits for safer deployment around people.
Why us
We’re four founders, including three PhD researchers, two specializing in robotics. Our backgrounds include Sony and NASA JPL, world-class research in machine learning and human–robot interaction, and two RoboCup@Home championship wins.
Our shared mission: enable robots to operate safely in the real world.
Our asks
Asim. Benchmarking physical intelligence.