Buy or subscribe
The researchers — who are based at Microsoft Research in Cambridge, Barclays in London and the University of Cambridge — combined analogue electronics with three-dimensional optics in a system that can perform an iterative fixed-point search without digital conversion. The analogue optical component executes the matrix–vector multiplication and the analogue electric component executes the nonlinear operations, subtraction and annealing. Each loop iteration takes 20 ns. The team show that the analogue computer can accelerate equilibrium models with up to 4,096 weights at 9-bit precision, classify images and perform nonlinear regression tasks.
Original reference: Nature 645, 354–361 (2025)
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$32.99 /30days
cancel any time
Learn more
Subscribe to this journal
Receive 12 digital issues and online access to articles
$119.00 per year
only $9.92 per issue
Learn more
Prices may be subject to local taxes which are calculated during checkout
Author information
Authors and Affiliations
Nature Electronics https://www.nature.com/natelectron/
Katharina Zeissler
Authors
- Katharina Zeissler
View author publications
Search author on:PubMedGoogle Scholar
Corresponding author
Correspondence to Katharina Zeissler.
Rights and permissions
About this article
Cite this article
Zeissler, K. An analogue computer for AI and optimization. Nat Electron (2025). https://doi.org/10.1038/s41928-025-01473-4
Published:
DOI: https://doi.org/10.1038/s41928-025-01473-4
