Sereact bets warehouse robots fail on software; $135M later, 200+ systems run

Sereact, founded in Stuttgart in 2021 by robotics researchers Ralf Gulde and Marc Tuscher, bet that warehouse automation was stuck because of software, not robots. Pick-and-pack is roughly 55% of warehouse operating cost, yet traditional automation took weeks to program, flinched when product ranges changed, and needed constant human oversight. Their answer, PickGPT, let an operator simply chat with a robot arm: no code, no retraining for every new object.

The bet funded itself in steps. A $5M seed from Point Nine and Air Street Capital in 2023 backed a product claiming 1,500 picks per hour and up to 77% savings per pick. In January 2025, Creandum led a €25M Series A with Daimler Truck, Bol, and BMW among customers. In April 2026, Headline led a $110M Series B — with Bullhound, Daphni, and Felix Capital — to scale Cortex 2.0 and open a Boston office for the US push.

As of September 2026, Sereact reports over 200 systems live in European production environments and more than one billion picks completed on its Cortex platform. Its zero-shot approach means robots handle objects they were never trained on, learning from every real pick. The company is still scaling and remains software-first in a market full of hardware bets.

The case is not settled: Sereact must prove its software can generalize across robot brands and hold margins against hardware giants moving into AI. But its trajectory shows a coherent thesis — the brain, not the arm, is where value accrues.