Machine learning and artificial intelligence are currently transforming society, but they demand an enormous amount of computing power. At the opposite end of the computer-power spectrum are extreme-edge devices, which are made using ultra-low-cost electronic components. Such devices execute basic computations locally before sending compressed data to the cloud and often need to be installed on flexible substrates — think of wearable sensors that process biometric data, brain implants for recording neural activity or applications in the ‘Internet of Things’. Writing in Nature, Yan et al.1 report the implementation of AI chips directly on flexible substrates, paving the way to ultra-low-cost AI systems for use in extreme-edge technology.
Nature 649, 1114-1115 (2026)
doi: https://doi.org/10.1038/d41586-026-00037-6
References Yan, A. et al. Nature 649, 1165–1171 (2026). Yan, A. et al. Adv. Func. Mater. 34, 2304409 (2024). Ozer, E. et al. Nature 634, 341–346 (2024). Çeliker, H., Dehaene, W. & Myny, K. Nature 629, 335–340 (2024). Verma, N. et al. IEEE Solid-State-Circuits Mag. 11, 43–55 (2019). Geng, D. et al. Nature Electron. 6, 963–972 (2023). Download references
Competing Interests The authors declare no competing interests.
Related Articles
Subjects