Impulse programme entrepreneurs boost AI and robotics innovation
Dr Markus Hellenbrand, a postdoctoral Research Associate working for the Department of Materials Science and Metallurgy at the University, is gracing the programme by combining processing and memory in the same place.
The approach is said to be unlocking a revolutionary way of computing. Dr Hellenbrand says ‘in-memory computing’ will make AI applications, in particular, even faster and more efficient than has previously been possible.
Dr Hellenbrand says: “The emergence of AI is one of the most transformative developments in human history. However, to train AI models currently requires unbelievable amounts of data and energy. This is because they are being implemented in conventional, all-purpose computing architecture, reliant on a separation of data storage and processing, which is not very suitable for AI-specific tasks.
“Shuttling huge amounts of data back and forth between the two components results in valuable energy being lost on the data transfer instead of being invested in the actual compute operation, and the transfer additionally constitutes a speed bottleneck.”
In-memory computing presents a more promising approach in his view. In this emerging architecture, data storage and computing occur physically in the same place, removing both the bottleneck and the energy sink.
Dr Hellenbrand explains: “In-memory computing requires a different type of memory element with multiple, ideally a continuous range of states, instead of binary memory operating with 1s and 0s. We have developed such memory elements in our lab, and we are now looking to spin out our technology into a startup to develop revolutionary AI hardware.”
Fellow Impulse 2025 participant, Zhuo Chen is focused on transforming robotic capabilities and, in particular, a growing need to develop robot bodies and dexterity, which has not kept pace with the breakthroughs in AI transforming robotic abilities.
Zhuo Chen is regarded as revolutionising robotic interaction with an advanced next-gen electronic skin (e-skin) for robots, enabling real-time, high-resolution 3D force and temperature sensing inspired by human tactile perception.
In stark contrast to the current offering of tactile sensors, which are typically bulky and limited to normal force sensing, Zhuo Chen’s miniaturised, high-resolution sensor array delivers human-like precision as well as multidimensional feedback.
Applications of the technology range from enhancing surgical robotics with delicate tissue handling to improving efficiency in precision manufacturing and reducing food waste in automated agriculture.
He explains: “Our innovation uses a multiscale structured composite with adaptive electrical conductivity, significantly improving sensitivity, reliability and dynamic range. A unique rectangular pyramid structure distinguishes normal from tangential forces, enabling precise tactile feedback and slip detection that is critical for delicate tasks previously impossible for robotic systems.
“Robots have the potential to ease critical labour shortages but to do so, we must close the gap between what robots and humans can handle. Without the flexibility, speed and precision of human manipulation, robots are unable to undertake the difficult and dangerous tasks they are most needed for.”
Impulse is specifically designed for high-potential individuals (early-stage entrepreneurs, PhDs, postdocs, researchers) and organisations (engineers, managers, scientists of large corporations) across the different technology fields of Engineering, Physics, Biotechnology, Computer Science, Mathematics, Chemistry and Medicine.
The action-learning initiative is designed to accelerate high-potential innovations into commercial propositions, guided by successful innovators and mentors from the Cambridge innovation cluster.
Applications for Impulse 2026 will open on Monday 10th November 2025. The Round One application deadline is 2nd February 2026. Round Two application deadline is 2nd March 2026. Fellowships are available for certain technology and research areas. To find out more visit https://www.maxwell.cam.ac.uk/apply-impulse
Impulse is specifically designed for high-potential individuals (early-stage entrepreneurs, PhDs, postdocs, researchers) and organisations (engineers, managers, scientists of large corporations) across the different technology fields of Engineering, Physics, Biotechnology, Computer Science, Mathematics, Chemistry and Medicine.
The established action-learning programme exists to help unique individuals keep focus and turn their exciting ideas or ventures to the point where it can become a viable business.
The programme is driven by a network of passionate mentors and top-tier experienced entrepreneurs, the key players of the celebrated Cambridge Cluster – and, over three months, participants are given the unique opportunity to explore directions and build deeper relationships with their mentors, as well as with investors and other key experts.
Impulse alumni include the founders of Riverlane, Abselion, Nu Quantum, Xampla, Echion Technologies, Wave Photonics, AllFocal Optics (formerly Lark Optics), Semarion, and About:Energy.