Polish robotics startup Nomagic secures €8.3 million to scale Physical AI operations in the US

Warehouse automation is entering a new phase where “Physical AI” is moving from pilots to scaled, real-world deployments, and Nomagic’s latest funding is a clear sign of that shift. As the ai world organisation, we’re tracking this momentum closely because it’s exactly the kind of applied innovation the ai world summit and wider ai world organisation events are designed to spotlight through ai conferences by ai world, including ai world summit 2025 / 2026.

Nomagic, a Warsaw-based warehouse robotics company focused on general-purpose Physical AI for warehouse operations, has announced an €8.3 million (about $10 million) Series B extension to accelerate its commercial push in the United States. The extension round is led by Cogito Capital Partners, and with this new capital, Nomagic’s total funding is now reported to exceed €70 million (about $84 million). In a market where warehouse leaders are constantly balancing speed, accuracy, labour availability, and rising service expectations, this round positions Nomagic to expand its footprint in one of the world’s most competitive logistics environments.

Kacper Nowicki, CEO and Co-Founder of Nomagic, framed the investment as a validation of the company’s belief that intelligent autonomous systems can finally connect digital optimisation with real-world execution on warehouse floors. Sylwester Janik, Managing Partner at Cogito Capital Partners, echoed that view by emphasising the urgency of modernising warehouse and logistics operations with solutions that bring intelligence, adaptability, and autonomy into everyday workflows. From our lens at the ai world organisation, that alignment between investor conviction and operational urgency is what typically accelerates category shifts, turning “interesting technology” into “default infrastructure.”

A Series B extension built for U.S. scale

Nomagic’s announcement isn’t just about the size of the round; it’s about what that capital is intended to unlock in the U.S. market. The company has stated that this Series B extension is aimed at accelerating commercial operations in the United States, signalling a focus on go-to-market capacity, enterprise deployments, and long-term customer expansion. For warehouse robotics, the “U.S. scale” conversation is often less about a single facility and more about repeatability: multi-site rollouts, predictable integration timelines, and the ability to keep systems stable through seasonal surges.

What stands out in this deal is the framing around Physical AI rather than “robots” in the traditional automation sense. In modern fulfilment and logistics, automation buyers increasingly want systems that can handle variability: mixed SKUs, changing packaging, different item textures, shifting demand curves, and process changes that happen faster than traditional industrial engineering cycles. Physical AI, as described in the report, is positioned as the next step in that evolution, blending AI compute with embodied systems (robots and machines) to solve physical-world problems like complex object manipulation.

For readers who follow the ai world summit, this is the same narrative we see across multiple industries: AI is moving from assisting decisions to taking responsibility for outcomes in real environments. That’s a big reason ai conferences by ai world consistently highlight not only model capabilities, but also deployment realities: integration, reliability, safety, and measurable business impact. It’s also why ai world summit 2025 / 2026 programming increasingly connects “what’s possible” with “what’s operational.”

Why “Physical AI” is becoming the logistics advantage

Warehouses are high-pressure systems where small inefficiencies compound quickly, and the “physical” constraints are unforgiving: items slip, boxes deform, lighting changes, barcodes get damaged, and stock positions are not always perfectly predictable. Traditional automation can work brilliantly when conditions are fixed, but fulfilment environments—especially e-commerce and omnichannel—tend to be dynamic by design. Physical AI, as described, aims to tackle that dynamic reality by letting systems adapt as they learn from ongoing operations.

In the EU-Startups report, Nomagic positions Physical AI as a way to bridge the gap between digital optimisation and real-world execution, which captures a key pain point for operators. It’s one thing to optimise a picking route on a dashboard; it’s another thing to reliably pick a soft apparel item, handle it carefully, place it correctly, and maintain throughput at peak demand. When AI becomes part of the control loop—perception, reasoning, and action—the system can, in principle, improve its own performance as conditions evolve.

This also explains why funding is flowing into the broader robotics stack, not just into “robots that move.” In 2025, adjacent infrastructure startups like Neuracore raised funding to build unified robot-learning infrastructure that standardises data and training pipelines for robotics teams—essentially strengthening the foundations that make Physical AI scalable. In parallel, companies like Filics raised funding in warehouse-automation hardware, showing that the market is investing across the full spectrum: learning infrastructure, autonomy layers, and physical platforms.

At the ai world organisation, we view this trend as a signal that the robotics sector is increasingly converging with the AI sector, rather than operating as a separate industrial niche. That convergence is a recurring theme at the ai world summit because it changes how enterprises evaluate technology: they start asking about data flywheels, model updates, remote operations, and long-term roadmaps—not just mechanical specs. And for anyone attending ai world summit 2025 / 2026, it’s a prime example of why “AI strategy” now includes warehouses, supply chains, and physical operations.

Nomagic’s approach: real data, always-on operations, and VLA models

Nomagic’s reported advantage is tied to how its deployed robots learn from “massive scale of real operational data,” built over millions of tasks in 24/7 environments, which trains an adaptable Physical AI platform for different warehouse tasks. That emphasis on 24/7 operational learning matters because it suggests the system is trained on real warehouse variance rather than only on lab conditions. In practical terms, the more diverse and continuous the operational dataset, the more opportunities the models have to generalise across item types, edge cases, and workflow changes.

The company also highlights its next-generation VLA (visual language action) models, which it says integrate automatically into its fleet of AI-powered robots and can accelerate autonomy while reducing deployment time. This matters because VLA-style systems aim to connect perception (what the robot “sees”), intent (what it is supposed to do), and action (how it physically executes), reducing the fragmentation that historically slowed robotics deployments. While the industry is still early in standardising these approaches across vendors, the direction is clear: robotics autonomy is becoming more model-driven, and updates are becoming more software-like.

On Nomagic’s own site, the company positions its solutions as AI-powered robotics designed to streamline warehouse operations for leading retailers, manufacturers, and logistics companies. The site also describes a support model built for “lights out” warehouses, including a 24/7 monitoring system and a remote operations platform intended to keep automated workflows running continuously. This operational layer is important because in real fulfilment settings, reliability and rapid issue resolution often matter as much as raw picking performance.

Nomagic also states it is backed by Khosla Ventures and references leveraging advances like large language models and vision-language-action models to enable high picking accuracy and careful handling at scale. That investor-and-technology positioning is consistent with a broader industry theme: investors increasingly reward robotics companies that present themselves as scalable AI platforms, not as one-off automation projects. For those following ai conferences by ai world, this is a familiar pattern across sectors: the winners tend to be the teams that combine strong engineering with a credible, evolving AI roadmap.

The 2025 funding context and what it says about 2026

Nomagic’s €8.3 million Series B extension sits within a wider 2025 European funding narrative in warehouse robotics and automation. The same report notes that earlier in 2025 Nomagic raised €41.5 million to scale AI-driven deployments across Europe, showing that the company has been building scale on its home continent before pushing harder into the U.S. In that same 2025 landscape, Neuracore’s €2.5 million and Filics’ €13.5 million rounds are presented as part of a broader movement toward standardised learning infrastructure and warehouse-automation expansion.

Taken together, EU-Startups describes these disclosed rounds as roughly €65–70 million flowing into European robotics, physical AI, and warehouse automation during 2025. Even without treating that number as an exhaustive market total, it’s a useful signal of investor attention: Physical AI is no longer a “future bet” only, but an active build cycle. And crucially, the same report indicates Nomagic plans to use the momentum from 2025 traction and breakthroughs to accelerate U.S. operations and continue development of VLA models through 2026.

From the ai world organisation perspective, this is precisely where the conversation becomes actionable for operators, founders, and investors. In 2026, the competitive edge will likely come from how quickly robotics platforms can be deployed, updated, and expanded across facilities—without long downtime windows or bespoke engineering each time. That’s why we expect Physical AI and robotics to keep gaining stage time in ai world organisation events and panels at the ai world summit, alongside more familiar GenAI use cases.

This also connects naturally with the way the ai world summit frames real-world AI adoption: practical learning, measurable outcomes, and cross-industry collaboration. The AI World Summit 2025 was positioned as a gathering of AI leaders, innovators, and practitioners, held on 17–18 January 2025 at Chitkara University in Rajpura, Punjab. Looking ahead, the AI World Summit 2026 Asia & Global AI Awards is presented on your site as taking place on May 28, 2026 at Singapore EXPO (1 Expo Drive).

For our community, stories like Nomagic are not just “funding news”; they are case studies in how AI becomes embedded into critical infrastructure, turning warehouses into living AI environments where data, autonomy, and operations improve continuously. If you’re building, buying, or investing in applied AI, the ai world summit and other ai conferences by ai world are built for exactly this kind of insight—what worked, what failed, and what scaled. And if your organisation is exploring warehouse automation, robotics, or Physical AI partnerships, the ai world organisation events ecosystem is the right place to compare approaches, meet operators, and learn what it takes to deploy at enterprise scale.