Robot AI Learns From 1 Billion Real Picks, Expands to US

Imagine a robot brain that gets smarter every time it picks up an object, learning from real mistakes instead of simulated ones. That's exactly what Sereact has built, and the results are stunning: their robots now need human help only once every 53,000 picks.

The Stuttgart-based company just secured $110 million to bring its Cortex 2.0 "robotic brain" to the United States. They've already opened an office in Boston and are hiring American teams to help scale the technology that's already working in warehouses and factories across Europe.

What makes Sereact different? While competitors spend billions training robots in simulations and labs, this company has spent five years teaching its AI in actual warehouses during real shifts. Every successful pick, every failure, and every recovery feeds data back into the system, making it smarter with each interaction.

Major companies like Mercedes-Benz, BMW, and Daimler Truck already trust Sereact's technology. The AI works across different types of robots, from single-arm pickers to humanoid robots, proving the same brain can adapt to any hardware.

The newly upgraded Cortex 2.0 brings something remarkable to the table: imagination. Instead of trying a motion and hoping it works, the robot now thinks through several possible approaches first, ruling out the risky ones before moving.

Dr. Ralf Gulde, the company's co-founder and CEO, calls it a "data flywheel." Real deployments create real data, which trains better models, which handle more picks successfully, which generates even better data. The cycle keeps improving.

Co-founder Marc Tuscher explains it simply: "The robot dreams in latent space. We give it a form of imagination, the ability to anticipate how the world will respond before it moves."

This matters for delicate work like placing windshield wipers without scratching them or assembling components under tension. The AI can adjust how much planning it does based on how expensive a mistake would be.

The Ripple Effect

This breakthrough reaches far beyond warehouse efficiency. As physical AI becomes more capable, it could help solve labor shortages in manufacturing while making dangerous or repetitive work safer for humans. The technology's expansion to the United States means more American companies can access robots that actually learn and improve over time.

Investment firms are betting big on this approach. Headline led the funding round, with multiple previous investors returning because they've seen the results. Trevor Neff from Headline calls it "one of the largest opportunities we've seen in a generation" for rewiring global supply chains.

The gap between Sereact and competitors widens with every shift their robots work, every night they operate, and every messy, unusual item they successfully handle.