The Energy-Saving Potential of Neuromorphic Computing

UK researchers are pioneering neuromorphic computing – hardware and software systems designed to mimic the neural structure and function of the human brain – to overcome today’s energy bottlenecks. In May 2025, an EPSRC-funded UK Multidisciplinary Centre for Neuromorphic Computing was announced, led by Aston Institute of Photonic Technologies at Aston University. This £5.6 million initiative brings together top teams from Aston, Oxford, Cambridge, Southampton, Queen Mary London, and others to co-design energy-efficient computing systems inspired by the human brain. The centre plans to integrate neuroscience, photonics, nano-electronics and new algorithms so that computer hardware can mimic neuronal networks with the aim of achieving a far higher compute per watt than conventional chips.

Technical Details

The centre's innovation lies in combining human neuroscience with photonic hardware development. Researchers study how human neurons, derived from stem cells, compute information, then use these insights to design photonic chips – systems that use light rather than electrons to process data. Photonic neuromorphic systems promise orders-of-magnitude improvements in energy efficiency for two reasons: they reduce the heat loss inherent in electronic circuits, and light’s ability to carry multiple signals simultaneously allows hardware to physically replicate the brain’s highly parallel, spike-based processing model. The centre brings together neuroscientists, photonics engineers, materials scientists, and computational researchers to co-design these systems, with industry partners including Microsoft Research, Thales, BT, and major defence contractors guiding development toward real-world applications.

Why This Matters for Organisations Today

Conventional AI hardware consumes kilowatts of power, whereas the human brain performs complex tasks on roughly 20 watts. The expansion of AI and the infrastructure it requires has exposed energy costs and sustainability as critical constraints for companies. The new UK centre directly targets this opportunity. Its research could lead to ultra-efficient AI accelerators and sensors that operate on minimal power, transforming data-center design and enabling smart devices with continuous learning. For industry and government, this work offers a path to next-generation computing infrastructure.

---

Source: Aston University press release – UK Multidisciplinary Centre for Neuromorphic Computing (EPSRC £5.6M centre launch, May 2025).

Author: Dr Dominic Smith, summarising public announcements by Aston University and UKRI.