British technology company Signaloid has demonstrated its C0-ASIC for physical AI. The application-specific integrated circuit (ASIC) is designed for robotics and industrial automation, with the company projecting a 1000-fold improvement in performance-per-watt compared to current hardware.

The chip utilises Signaloid’s “distribution-extended compute” technology, known as UxHw. Unlike traditional processors that handle single numbers, this hardware represents values as probability distributions. This allows the system to consider multiple scenarios simultaneously, which is a requirement for autonomous navigation and predictive maintenance.

The technology is intended to enable autonomous mobile robots (AMRs) to navigate factories more safely and allow programmable logic controllers (PLCs) to better estimate equipment wear. Signaloid claims that a single execution on its platform can replace the millions of repetitions usually required by classical iteration-based computing methods.

“The compute workloads that Signaloid’s UxHw is designed for are used across many aspects of computing, from physical AI and robotics to supply-chain modelling,” said Phillip Stanley-Marbell, founder and CEO of Signaloid.

Cloud and FPGA-based versions of the technology have already shown significant speedups in infrared sensor analysis. The new ASIC will complement existing hardware modules that integrate with Bosch Rexroth’s industrial controllers.