Publication date: 22nd July 2026
The past decade has witnessed a surge of interest in indoor photovoltaics (IPVs), and the opportunities they offer in sustainably powering autonomous Internet of Things (IoT) electronics [1,2]. Simultaneously, we are witnessing the exponential, pervasive rise of aritifical intelligence, which will severely test the limits of underpinning infrastructure, including electricity and water supply. An alternative paradigm is available, where rather than computing within centralised servers, computations are performed locally across a network of billions of small, autonomous nodes, much like the IoT. This is known as edge computing, or distributed intelligence, and can save on the substantial energy costs of communication, but relies on dependable and resilient local energy supplies.
In this talk, I will discuss the opportunities afforded by emerging IPV materials that now offer indoor power conversion efficiencies >35%, substantially exceeding commercial standard IPV based on hydrogenated amorphous silicon (7-16% efficiency). There are three sections. The first discusses energy requirements of edge computing, and the tiers of computatinal tasks accessible depending on the terabyte operations per second (TOPS) that can be carried out, and opportunities to improve the performance efficiency of computing units through neuromorphic devices. The second section discusses the emerging IPV technologies that are now able to provide the energy required for edge computing, the main advances that enabled these improvements, and current materials challenges that need to be addressed to further increase efficiencies towards their radiative limits. I will also discuss emerging nontoxic and stable IPV materials that also have potential for powering edge computing. Finally, I will close the loop and discuss promising energy storage solutions that can meet the high charge/discharge rates required for edge computing, and which can work in synergy with IPV to provide a reliable energy source.
I acknowledge support from UKRI through a Frontier Grant (no. EP/X029900/1), awarded through the 2021 ERC Starting Grant scheme, as well as the UK Innovation and Knowledge Centre in Neuromorphic Computing Hardward, link: https://neuroware-ikc.com/
