← All companies

Lamb Labs

Active

Lightning fast chips with hardcoded AI models

S26·Summer 2026·B2B·London, England, United Kingdom

About

Lamb Labs is building MPUs (Model Processing Units), custom chips that hardcode LLM weights directly into the chip. Unlike GPUs, which move weights between memory and compute, MPUs keep the model on-chip, targeting up to 20,000+ tok/s and a 63× higher intelligence per watt.

Change history · 3 recorded

  1. August 23, 2026
    • Description rewritten07:00 PM
  2. August 21, 2026
    • Pitch updated07:00 PM
  3. August 2, 2026
    • Tags updated07:00 PM