An F1 Aerodynamicist Raises $55M to Teach Factory Robots Using Footage of People Doing Chores
Bercan Kilic left Red Bull Racing at the peak of its success. His startup, microagi, offers a unique solution to the robot data problem by filming 20,000 people performing tasks like mopping floors.
Bercan Kilic’s Journey and Startup Success
Kilic secured his dream job in 2023, designing aerodynamics for Red Bull Racing during their winning streak. However, he found the engineering work thin. His Munich startup, microagi, has now secured a massive $55 million in funding, the largest seed round for a German company. Hummingbird led the investment round, with Northzone, LocalGlobe, Village Global, and redalpine also participating.
microagi’s Niche Approach
microagi does not build robots or models; instead, it focuses on data collection. It uses cameras and sensor-equipped gloves to record workers performing tasks and then teaches existing robotics models to execute specific jobs within a customer’s factory using this footage.
"We provide the labs with data, they provide us with models, and then we layer on proprietary data to make our customers happy," said Kilic.
A Detour Leading to Success
Initially, microagi planned to only offer deployment services but realized that most robotics models were not advanced enough to start from scratch. Kilic uses the analogy of training a new hire: an adult can learn a factory job quickly, while a child might never master it. Existing robot models are akin to children in this context.
shift: From Viral Cleanings to Private Chefs
microagi’s solution is called shift, which gained popularity this year by offering free apartment cleanings in New York and San Francisco in exchange for filming cleaners doing dishes, mopping, and folding laundry. This week, shift expanded its services to include free private chefs.
shift operates in 15 countries and pays over 20,000 people to record themselves performing physical tasks, selling the footage to labs developing robot brains. It competes with companies like Scale AI, Turing, and micro1, all addressing the same data gap.
The Asymmetry and Future of Robotics
The reason for this business model lies in a genuine asymmetry: language models were trained on the internet, but robots have no access to it—a gap that Berkeley roboticist Ken Goldberg estimates at 100,000 years. Hence the need for cameras to collect data.
Kilic’s approach also caters to the global labor shortage, as Europe and the US face worker scarcity, while China is rapidly automating regardless. In 2024, China installed a staggering 295,000 factory robots, accounting for 54% of the world’s total, compared to just 34,200 in the United States.