Dr Kilian Stenning, Rayd Technologies

Brain-inspired computing for energy-efficient AI 

The challenge

The use of AI is booming – from voice assistants to recommendation algorithms – but training and running it comes at a cost: massive energy use and data needs.  

Data centres, the vast buildings full of servers and storage that power the digital world, including AI, already consume around 2.5% of the UK’s electricity, and that could quadruple by 2030. Globally, these centres use about 1.5% of all electricity, and demand is only growing. 

The challenge before us is urgent: how can we keep expanding AI without having a profound impact on our planet? Dr Kilian Stenning, CEO and co-founder of Rayd Technologies and a Research Fellow at Imperial College London, has a bold answer.

The innovation

Dr Stenning is developing new computer hardware and algorithms that use light and brain-inspired learning dynamics to process data and images extremely efficiently – mimicking the way the human brain works to compute information with very limited data and energy. He predicts that, with this technology, computation could become between 1,000 and 10,000 times more efficient than current approaches. 

By making the hardware smaller and more efficient, and leveraging the brain-inspired learning dynamics, the technology could drastically improve AI's speed, energy and data efficiency, and lower data centre emissions, making it far more environmentally friendly. 

Over the 10 years of the fellowship, Dr Stenning and his team will develop this technology, with the intention that it could enable more sustainable AI across sectors such as manufacturing, healthcare and robotics in the long term.

The innovator

Dr Stenning originally trained in nuclear engineering before completing a more physics-based PhD, where he explored unconventional approaches to computation. Motivated by the question of how to make computing more efficient, he began investigating systems that would be capable of processing information in new ways, leading to his current research.

• Website
• LinkedIn

We’re really rethinking how computation happens at a fundamental level.