AUTO-UPDATED

Alibaba Cloud plans to use fewer Western chips, to boost its already huge AI margins

Waymo is developing custom 5nm machine learning accelerators to enhance its autonomous vehicle capabilities, aiming to maintain a competitive technological advantage over Tesla in the self-driving market.

Key Points

  • Waymo’s new custom-designed robocar chips are engineered to deliver over 1,000 TOPS with ultra-low latency.
  • The 5nm ML accelerators are specifically optimized to process complex sensor data for autonomous navigation.
  • This hardware development represents a strategic move to secure a performance lead against Tesla’s proprietary FSD hardware.
  • The initiative highlights the growing trend of automotive companies moving toward vertical integration of specialized silicon for AI workloads.

Why it Matters

This shift toward custom silicon underscores the critical importance of specialized hardware in the race to achieve reliable, scalable autonomous driving. By controlling the underlying chip architecture, Waymo can optimize performance and power efficiency in ways that off-the-shelf components cannot match.
Theregister.com Published by Simon Sharwood
Read original