A Review of AI Computing Hardware for Enabling 6G Wireless Networks: Neuromorphic Chips, NPUs, and TPUs

The introduction of 6G Networks and its deep integration of Artificial Intelligence (AI) necessitates powerful and efficient hardware processors to manage massive data flows, ensure network reliability, and support demanding use cases. To overcome these challenges, the industry is shifting away from conventional processing architectures (CPUs and GPUs) toward specialized hardware accelerators. This paper reviews three critical categories of processing chips: Neuromorphic chips, Neural Processing Units (NPUs), and Tensor Processing Units (TPUs). It also reviews other emerging chips that are still relatively unknown such as Quantum Processing Units (QPUs) and Photonic Integrated Circuits (PICs). Specifically, this paper will review different research tasks and display tables of hardware efficiency and wireless system metrics.

Preparing paper insights from the available abstract and paper details.

Authors

Yu-Dong Yao; Jason Lee

Published

Jul 16, 2026

Journal or venue

AI Engineering

DOI

10.53941/aieng.2026.100010

Paper type

journal article

Source

Crossref