microagi

Robotics · Industrial AI · Training Data

Funding

Microagi raises $55M seed, largest ever in Germany, for Atlas platform

The startup skips building robots or AI models entirely, selling only the factory-floor data layer that fine-tunes other companies' machines.

• Munich-based microagi raised $55 million in seed funding led by Hummingbird, with Northzone, LocalGlobe, Village Global and Redalpine participating, in what the company calls Germany's largest-ever seed round.
• microagi doesn't build robots or AI models; its Atlas platform captures factory-floor data via cameras and sensors to fine-tune existing robotics models for a specific customer's plant-specific tasks.
• The company buys hardware from Chinese robot makers like Unitree and UBTech, leases it to factories, and layers its own software and training data on top before deployment.
• Founded roughly 10 months ago by former Formula 1 engineers from Red Bull Racing and Mercedes-AMG Petronas, microagi already has five companies collecting data through Atlas, with one preparing a factory deployment.
• microagi's data-collection arm, shift, went viral offering free apartment cleanings in New York and now private chefs in San Francisco in exchange for filming people doing chores, competing with Scale AI, Turing and micro1 for robot training footage.
• Funds will go toward scaling Atlas across large industrial organizations, expanding engineering and research teams, and building out a new research hub in Zurich alongside operations in Munich, London and New York.
• The deal shows investors betting big on the 'data layer' between robot hardware and robot brains, a less glamorous but potentially higher-leverage bet than the hardware-focused megarounds dominating robotics funding.