Source: https://www.nature.com/articles/s41928-025-01477-0

Analysis of a breakthrough analogue computing chip developed by Peking University researchers, which uses Resistive Random-Access Memory (RRAM) to perform computations. This specialized chip is claimed to offer potential orders-of-magnitude improvements in throughput and energy efficiency over digital processors like the Nvidia H100 GPU for solving complex matrix equations, the core task of AI and HPC.

The innovation lies in its compute-in-memory (CIM) architecture and a hybrid iterative algorithm that solves the historical problem of analogue imprecision, overcoming the von Neumann bottleneck.

The report concludes that while the chip poses an asymmetric threat to digital dominance, its success hinges on overcoming significant hurdles, particularly the creation of a robust software ecosystem and scaling manufacturing, making this a pivotal development in the US-China technological competition.