After digital computing has dominated the computer field for over half a century, Chinese scientists made major breakthroughs in new architectures. In October 2025, a team led by Sun Zhong from the Institute for Artificial Intelligence at Peking University, in collaboration with a research team of the School of Integrated Circuits, successfully developed a high-precision, scalable analog matrix computing chip based on resistive random-access memory (RRAM), according to Zhonghongwang, a news website managed by the National Development and Reform Commission (NDRC).

This breakthrough technology has, for the first time, increased the precision of analog computing to 24 bits fixed-point precision, with computational throughput and energy efficiency far exceeding existing top-tier GPUs, by a factor of hundreds to thousands.

Related NewsBOCI Raises ASMPT (00522.HK) TP to $98, Keeps Rating at Buy
Current mainstream CPUs and GPUs on the market are digital chips, adopting the Von Neumann architecture, which separates computing and storage functions, and achieving information computation and transmission through the compilation, calculation and decoding of binary code stream.

One of the advantages of analog computing based on RRAM is eliminating the process of converting data into binary digit stream and avoiding procedural data storage, thereby integrating the data calculation process with data storage, so as to achieve computing power release.