Most robotics demos are filmed in a lab, where the lighting is good and nothing goes wrong. Nomagic’s pitch is that its robots have spent years in the unglamorous reality of actual warehouses, millions of tasks, 24/7 environments, the messy edge cases that break tidy systems. The Warsaw-based company just added an €8.3 million Series B extension, led by Cogito Capital Partners, bringing its total funding past €70 million and pointing the next phase squarely at the United States.
The framing the company uses is “Physical AI”, the idea that the next leap in AI isn’t another chatbot but intelligence wired into machines that manipulate the physical world. In Nomagic’s case that means robots handling the genuinely hard problem of picking and moving varied objects in a warehouse, trained on a large base of real operational data rather than simulation. CEO Kacper Nowicki calls it bridging “the gap between digital optimisation and real-world execution,” which is consultant-speak for “the robot actually grabs the right box.”
The hard part of warehouse robotics was never the arm. It was the judgment, knowing what to grab and how. Nomagic’s edge is the millions of real tasks its system has already failed and learned from.
The money is for commercial expansion, not invention: scaling US operations and continuing to develop its next-generation vision-language-action models. That’s the right use of capital for a company at this stage, Nomagic isn’t trying to prove the technology works, it’s trying to sell it into the world’s largest logistics market while the demand for warehouse automation is acute.
The competitive field is crowded and increasingly well-funded, and “Physical AI” is a phrase a lot of companies are suddenly using. But there’s a specific satisfaction to a Polish robotics company, founded in 2017, crossing €70 million in funding and expanding into the US on the strength of robots that learned in real European warehouses. The talent didn’t leave to build this. It built it in Warsaw and is now exporting it.