From the research lab to industry: DLR spin-off Entrox Systems brings Edge AI to market readiness
- Four DLR scientists founded Entrox Systems, a start-up that develops edge AI predictive software based on reservoir computing
- The technology is based on physics-inspired AI methods researched at DLR and runs directly on existing hardware, without the need for a cloud connection
- The first pilot customer is using the software as a soft sensor in an artificial heart to calculate vital parameters in real time
- The target audience consists primarily of machine and plant manufacturers, who can integrate the technology directly into their products
Can artificial intelligence heal hearts? The DLR spin-off company Entrox Systems is working on it - albeit indirectly. This spin-off from the German Aerospace Center (DLR) is developing edge AI predictive software that can perform complex calculations directly on existing hardware without requiring high-performance servers or a cloud connection.
From a research idea to setting up a business
The idea for the start-up originated with mechanical engineer Tamon Nakano. Together with physicists Daniel Köglmayr, Sebastian Baur and Christoph Räth, he conducted research at the DLR into physics-inspired AI methods. In 2026, the team jointly founded Entrox Systems.
This is an example of how DLR research gives rise to marketable technology.
The Challenge: AI usually requires data centres.
AI applications typically require large, computationally intensive models, a stable cloud connection, and a great deal of energy. This becomes problematic whenever there is no reliable internet connection, machines need to make real-time decisions, or sensitive data must remain within the system.
The solution: Edge AI without expensive servers
This is precisely the area in which Entrox Systems specialises: the team has developed a local AI application that addresses these issues. Customers do not need any additional costly servers – the software runs on existing hardware. This saves energy, costs and space.
First application: soft sensor in artificial heart
The potential of this application is particularly evident with the first pilot customer, where the AI is used as a soft sensor in an artificial heart to calculate vital parameters directly within the device in real time. In this application, neither a full-scale AI data centre nor a cloud connection with the associated latencies would be practical – Entrox's edge AI approach is virtually without alternative.
Looking ahead, Entrox Systems is targeting primarily machinery and plant manufacturers, who can integrate the soft sensors directly into their products to deliver smarter machines without the need for additional hardware.
Further information on the fields of application and the transfer of technology from DLR research is available from DLR Innovation & Transfer and on the Entrox Systems website.