Lamb Labs
Lamb Labs is a YC Summer 2026 (S26) company. Based in London, England, United Kingdom.
About Lamb Labs
Summary
Lamb Labs builds MPUs (Model Processing Units): custom chips that hardcode an entire LLM's weights directly into silicon, eliminating the memory-bandwidth bottleneck that forces GPUs to shuttle weights between memory and compute. The first MPU carries Larry, a one-bit quantized 27B-parameter model compressed to 3.8 GB, targeting throughput above 20,000 tokens per second. The product replaces GPU-based inference stacks for workloads where speed and energy efficiency are the primary constraints.
Go-to-market
Lamb Labs targets B2B customers who need fast, power-efficient AI inference at the hardware level, progressing from a live software demo of Larry on GPU today through FPGA prototypes toward custom silicon.
From the Y Combinator listing
Lightning fast chips with hardcoded AI models
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.
Facts
- B2B / B2B
- London, England, United Kingdom
- Active
- Early
Founders (2)
- Thomas Lanning
Score every batch against your own thesis
GPAgent's sourcing agent reads every company in a batch, scores each one against your fund's thesis in a private workspace, and lines up the founders worth meeting before Demo Day. You decide who to contact. Your scores, notes, and pipeline never appear on these pages.
How Accelerator intelligence works