Recap of Q2. Concrete targets for Q3 and Q4. Our quarterly business review with HÖRMANN Intralogistics. The partnership works because it runs on live systems, not concepts. AutoStore installations that were built for storage now handle autonomous goods-in, picking and returns. HiLIS orchestrates the process. Cortex handles the objects, including the ones the robot has never seen before. No teach-in per SKU. Q2 showed what that looks like in production. Q3 is about repeatability. More joint sites, shorter integration cycles, broader process coverage. Thanks to Anna Sophia Kehmstedt, Robert Heinz, Sebastian Behrens, and the HÖRMANN Intralogistics team for a direct and useful session. #Intralogistics #AutoStore #PhysicalAI #WarehouseAutomation
Info
Sereact's mission is to drive your economic growth by closing automation gaps in your intralogistics. Our AI software enables machines to perceive their environment and develop solution strategies on their own, thus qualifying them to become autonomous skilled workers. By leveraging embodied AI for robotics, we ensure that our systems not only think but also physically interact with their surroundings in an intelligent manner. Sereact's AI software for autonomous robotics fully automates pick-and-place processes, making them more efficient, reliable and resilient. Our goal is to optimize your supply chain with minimal integration effort to increase productivity in your warehouse from the first pick while significantly reducing costs.
- Website
-
https://sereact.ai
Externer Link zu Sereact
- Branche
- Softwareentwicklung
- Größe
- 51–200 Beschäftigte
- Hauptsitz
- Boston(US) / Stuttgart (DE)
- Art
- Kapitalgesellschaft (AG, GmbH, UG etc.)
- Spezialgebiete
- Robotics, Industrial Automation, Deep Learning, Artificial Intelligence, Logistics und Embodied AI
Orte
-
Primär
Wegbeschreibung
Boston(US) / Stuttgart (DE), DE
Beschäftigte von Sereact
Updates
-
Rohlik Group has ordered 27 additional Sereact pick-and-place stations, bringing the total deployed across their fulfillment centers to 51, on the joint path toward 100 stations for automated e-grocery fulfillment. Our AI-powered stations handle single-item picking at Rohlik's sites, helping keep same-day delivery fast, accurate, and profitable. This expansion builds on the successful operation of the first 24 stations, with our partner Veloq helping scale the solution commercially and operationally across Rohlik's network. 51 stations live. 100 on the horizon. #AIRobotics #Ecommerce #SupplyChain #Automation #Sereact
-
-
New market. New category. New deployment. Sereact and Monta are launching an AI-powered dual arm returns program for fashion. We are deploying multiple dual arm stations, beginning at Monta's fulfillment center in Nieuwveen, with the remaining stations rolling out across the wider Monta network. Fashion carries the highest return rates in e-commerce. High SKU variety. Mixed packaging formats. Quality standards that leave no room for error. The category that has resisted automation the longest. Powered by Cortex 2.5, our dual arm stations receive, identify, assess, fold and sort returned items across high-volume operations. Monta serves more than 3,000 online retailers and is part of DHL Supply Chain. This brings autonomous returns handling into one of Europe's most dynamic fulfillment ecosystems, and marks our entry into the Dutch market. Every station handling live returns feeds our learning cycle. Every deployment makes the next one stronger. Phase one starts now. The blueprint scales from here.
-
-
Three sites. Two continents. One Brain Together with Arvato, we are scaling AI-powered picking across their fulfillment network. Four robots in Dortmund. One in Gütersloh. One in Memphis. The systems pick directly from AutoStore bins into cartons and totes, and pick from and place onto conveyor lines. Powered by Cortex 2.5, they handle a wide range of items with human-like dexterity and adapt in real time to changing products, packaging, and order profiles. No pre-programming. No fixed item catalogs. Dortmund becomes an anchor location for high-throughput automated picking. Gütersloh and Memphis extend the technology across Europe and North America. One AI. One consistent automation standard across the network. From pilots to live operations. Thank you, Arvato, for the trust.
-
-
25 dual arm robot stations. All running on live fashion returns. All learning at the same time. Together with MS Direct AG, Switzerland's market leader in e-commerce logistics, we have launched the world's first AI-powered dual arm returns program for fashion at their St. Gallen facility. The problem: returns handling is hard because every item is different. Varied SKUs, mixed packaging, high quality standards. Powered by Cortex 2.0, our dual arm system receives, identifies, assesses, folds and sorts each one. The breakthrough is the structure. This is not one robot in a lab. It is 25 stations processing real production returns in parallel. More stations, more data, faster path to full autonomy.
-
-
We're excited to welcome Zalando as a strategic investor in Sereact's Series B, now at $116M. This is what makes the partnership meaningful: Zalando is one of Europe's leading e-commerce platforms, built on operational excellence at real scale. Their decision to back Sereact is a strong signal of confidence in physical AI for the warehouse - and in the technology we're building together in Europe. Our robotic brain, Cortex, runs as a universal AI operating system across single-arm picking cells, dual-arm returns stations, and humanoid robots. With Zalando on board, we'll scale Cortex and expand into new markets - working alongside a partner who understands fulfillment from the inside. Thank you, Zalando, and all our investors for believing in the future of physical AI.
-
-
We've trained Cortex 2.5, and we're excited to share what it can do on a completely new task without any finetuning. A genuinely new task is where a model normally needs a training cycle: a longer sequence, a different workflow, several steps that have to happen in the right order. That's where the cost of teaching a robot something new really bites. But what if you could just show it? Cortex 2.5 learns from demonstrations the way a language model learns from a few examples in a prompt. Record the task a couple of times, and the robot does it. No retraining in between. A few things that make it work: 🦾Demonstrations are video. The clips are what specify the task. 🦾We keep them whole and retrieve the relevant moment as the task unfolds. 🦾The robot can tell how well the demonstrations cover its current situation, so it knows when it's on familiar ground and when to think harder. Early results on related tasks: 92.6% success from a handful of demos and no training. Full write-up in the comments.
-
-
Berlin. Inside a knuspr.de fulfillment center. We sat down with Rohlik Group to talk about what we're building together. Grocery is one of the hardest picking domains there is: Tens of thousands of SKUs. Chilled, heavy, fragile. Throughput windows that don't forgive mistakes. This is where Cortex runs. In production. Real stations, real volume. And Rohlik is scaling with us. More stations in Berlin and up to 100 stations across the network. Every pick trains the model. Every new station widens the gap. Thanks to Vincent Meyer, Anastasiia A., and the teams at knuspr.de and Rohlik Group for the partnership
-
Are humanoid robots ready for your warehouse yet? It's a question we hear a lot, and one we dug into with the logistics and factory planning team from LOGSOL GmbH when they joined us for a Consulting Day at Sereact. Instead of the easy answer, we opened up the topics that actually decide whether a robotics project works on the floor: - what today's AI-driven picking and manipulation can (and can't) do - how we design robotics cells and where they create the most value - how our systems connect to warehouse technologies like AutoStore™ Days like this go both ways. The better our partners' planning teams understand what Physical AI makes possible today, the better the robotics projects we scope and deliver together tomorrow. Thank you to the LOGSOL GmbH team for the open exchange and the questions that pushed the conversation further. We're looking forward to what's next.
-