Apple, Xiaomi and Nvidia chase terabyte-class bandwidth for AI chips

DIGITIMESSeptember 3, 2026en

For decades, semiconductor performance gains largely came from shrinking process nodes, from 22nm and 14nm to 7nm and 3nm. AI is changing the constraint: processors are getting faster, but moving enough data to feed them is increasingly difficult.

This is a short summary published by AI Global Wire. The full article is owned and hosted by DIGITIMES — open it there to read it in full.

• Verktyg