Develops superconducting computer chips for AI inference that replace conventional semiconductor GPUs, aiming to reduce power consumption by 10–100x near term and approach the physical limit of computing efficiency
Develops superconducting computer chips for AI inference that replace conventional semiconductor GPUs, aiming to reduce power consumption by 10–100x near term and approach the physical limit of computing efficiency. Its chips perform classical digital logic, are designed to work with existing software, and are described as fabricable using standard lithography.