HKU Team Bypasses Von Neumann Bottleneck with Ultra-low Power 2D Material AI Chip

YiMaTweets · x · 2026-08-11

A research team from the HKU Faculty of Engineering, led by Professor Can Li and Dr. Guoyun Gao, has developed a new computer chip that bypasses the traditional 'von Neumann bottleneck.'

To address the slow and energy-consuming data transfer between memory and processors caused by the AI boom and edge computing, the team utilized 2D materials to create a chip with minimalist architecture. It performs complex searches directly where data is stored. Achieving record-breaking energy consumption of under 0.1 femtojoules per operation, the chip significantly boosts processing speed while slashing power usage.

Original post →

More from Infra

Infra channel →