Medical logistics firm Apian has announced plans to develop a foundation world model for pathology laboratory logistics and automation, collaborating with service provider Synnovis, and incorporating advanced robotics and physical AI.
Building on photorealistic digital twins constructed for major clinical facilities like Great Ormond Street Hospital (GOSH) and Guy’s and St Thomas’ NHS Foundation Trust, Apian recently closed the loop between simulation and reality. In live supervised trials at the Synnovis hub laboratory in Blackfriars, Apian’s autonomous ground robots (pictured) delivered more than 1,400 live patient samples, connecting drone landing pads with the diagnostic laboratory through autonomous ground transport. Now, Apian and Synnovis are taking the next step: moving from modelling where healthcare happens to teaching AI how healthcare laboratories operate.
Much of the physical work behind the two billion pathology tests carried out in the UK each year depends on manual ‘human middleware’: highly trained staff physically carrying samples between landing pads, centrifuges, and analysers. That physical movement is critical to the entire clinical pathway. A delayed sample does not just slow down a laboratory; it delays diagnosis, postpones treatment, and keeps hospital beds occupied longer than necessary.
Pathology logistics is therefore an ideal proving ground for physical AI. A single medical sample moves between people, software, autonomous robots, transport systems, and laboratory equipment before producing a diagnostic result. Apian’s orchestration platform connects these disparate elements into one intelligent system, with the aim of understanding the entire pathway as a unified physical workflow.
Apian and Synnovis are combining observations of real-world laboratory workflows with data from laboratory equipment, software and the physical environment, using NVIDIA’s physical AI stack to develop the foundation world model. The goal is to teach AI not just where objects move, but the spatial dynamics and operational decisions behind complex laboratory operations.
To bring this intelligence layer to life, the initiative leverages NVIDIA's advanced physical AI stack. NVIDIA Omniverse and NVIDIA Isaac Sim provide high-fidelity 3D digital twins where robotic agents practice navigating complex laboratory layouts, lifts, and dynamic obstacles in simulation before live deployment. Then, using NVIDIA Cosmos foundation world models to explore real-world operational telemetry and visual data can help AI understand and predict physical interactions and spatial dynamics inside complex laboratory environments.
This creates a new class of training data for physical AI. As more sample deliveries are performed, telemetry feeds back into the simulation, making each deployment smarter and safer. The ambition is not pathology laboratories without people, but pathology laboratories where automation takes more of the repetitive physical movement of sample logistics, giving specialists more time to focus on vital diagnostic care.
Alexander Trewby, CEO at Apian, commented: “Getting robots to move thousands of real patient samples was an important milestone. But it also gave us something potentially much more valuable: real-world experience of how physical healthcare actually works. Now we want to turn that experience into intelligence. We started by teaching robots how to navigate using NVIDIA-powered digital twins. With Synnovis and NVIDIA, we are building a foundation world model for pathology laboratory logistics so healthcare professionals can spend less time moving samples and more time on the diagnostic work that requires their expertise.”
Mark Dollar, CEO at SYNLAB UK & Ireland, added: “Our scientists perform cutting edge medical diagnostics every day, providing answers for patients and healthcare professionals when they’re most critical. Utilising autonomous ground robots and AI to transport samples through our laboratories means our people can spend more time doing what only they can do – supporting better decision making and better outcomes for patients."