Seoul, March 16 -- Nvidia may unveil a new artificial intelligence inference chip architecture built around on-chip static random access memory, or SRAM, at the Nvidia GTC 2026 conference in San Jose, California, on Monday, raising questions about whether the design could reshape the structure of the AI memory market, Korea Herald reported.

The news report, quoting industry sources, noted that the proposed SRAM-based architecture is expected to take a different approach from the GPU designs currently used in AI data centres.

Today's GPUs process massive datasets by attaching multiple stacks of high-bandwidth memory, or HBM, next to the processor, enabling large volumes of data to be handled at extremely high speeds.

The SRAM-centered d...