Sereact, a Stuttgart-based AI company, has built software that lets robots pick, sort, and place physical objects in warehouses and factories. Its core product, called Cortex, runs on picking robots, dual-arm stations, and humanoid platforms. The company was founded in 2021 by Dr. Ralf Gulde, now CEO, and Marc Tuscher, CTO. Their starting premise was straightforward: a robot trained only in a lab will not hold up in a real warehouse.
The global warehouse robotics market was valued at $14.7 billion in 2024 and is projected to grow at over 23% annually through 2034, pushed by the growth of online shopping, rising labor costs, and tighter fulfillment timelines. Within that market, robotic picking, the task of identifying, grasping, and placing individual items, remains technically difficult. Most systems handle familiar objects well enough. When something unexpected arrives, many of them fail.
Sereact has raised $110 million in a Series B round led by Headline. Bullhound Capital, Daphni, and Felix Capital joined as new investors. Air Street Capital, Creandum, and Point Nine, all of which backed Sereact in earlier rounds, returned. Creandum led the company's 2025 Series A. Total funding to date now exceeds $140 million.
Most robots in production today were trained on simulated data or controlled test environments. When the real world does not match those conditions, performance drops. A box arrives crushed. A label is missing. An item looks nothing like any object the model has seen before. When a robot misses a grasp, many systems simply repeat the same failed motion.
Sereact has spent five years trying to fix that by training on real operations rather than simulated ones. Every pick completed across its deployed systems feeds back into a central model. Every failure and every recovery goes in too. The model is retrained continuously and pushed back out to the fleet. Sereact calls this a data flywheel: the more robots it deploys, the more real-world data it generates, and the better the model gets.
"The physical AI opportunity is one of the largest we've seen in a generation, and we believe it will rewire global supply chains and manufacturing. Behind great opportunities and great companies are great founders, and Ralf and Marc are building into that opportunity the right way: real deployments, real data, and a model that compounds and gets better with every single pick. Customers love the product, which leads to continued expansion, only accelerating the data flywheel — this is why we are so excited to back Sereact." — Trevor Neff, Growth Partner, Headline.
The new capital goes toward two priorities: building out Cortex 2.0, the next version of Sereact's AI software, and entering the US market. The company is opening its first US office in Boston and hiring commercial, engineering, and applications staff there.
Cortex 2.0 is designed to work differently from the current version. Today's system sees a scene and acts. The new version generates several possible outcomes before the robot moves, scores each one for stability, risk, and efficiency, then commits to the best option. If conditions change mid-task, it updates its plan in real time. The company describes this as moving from try-and-see to plan-and-try.
That difference matters most for tasks where a single wrong move has consequences. Fitting a component under tension, placing a windshield wiper on glass without scratching it, or positioning a part so the next station in the line can process it correctly. These are the applications Sereact is targeting with Cortex 2.0.
More than 200 Sereact systems are operating across Europe. The company says it has completed over one billion real production picks across its customer base. Of those, roughly one in every 53,000 required a human to step in remotely. Customers include Daimler Truck, Mercedes-Benz, BMW, Active Ants, Austrian Post, bol., DeltiLog, Monta, MS Direct, PepsiCo, and Rohlik Group.
Cortex runs on single-arm picking cells, dual-arm returns stations, humanoid robots, and Sereact Lens, a 3D perception system used for inventory and quality control. The software is designed to run across different robot types without needing to be rebuilt for each one.
The model on a Sereact robot is not static. It updates with every shift. New data comes in, the model is retrained, checked against automated tests, and redeployed across the fleet. Tuscher has said the company ships one thing: the model that runs on any robot, whether single-arm, dual-arm, humanoid, or fixed cell. His view is that robot hardware is becoming less differentiated, and the software is what matters.
"We bet early that you can't build real robotics AI in a lab. You build it with a data flywheel fed by real deployments — shipping into production, living with the failures, and letting the model learn from what actually happens on the floor. The numbers show it worked. Two hundred systems. One billion picks. One intervention per 53,000. Nobody else is close." — Dr. Ralf Gulde, CEO and Co-Founder, Sereact.
The Series B was led by Headline. Bullhound Capital, Daphni, and Felix Capital participated alongside returning investors Air Street Capital, Creandum, and Point Nine.
"After looking at a deluge of humanoid robotics companies, my fellow Partner Alon Kuperman and I were delighted to meet Ralf and Marc, the co-founders of Sereact, who have built an AI operating system that seamlessly retrofits into the world's vast fleet of industrial robots already in action." — Per Roman, Founding Partner, Bullhound Capital.
