Fractile
Fractile is building chips to run large language models two orders of magnitude faster. Existing hardware is good for training LLMs, but very poorly suited to subsequent inference of the trained…
Fractile
Fractile is building chips to run large language models two orders of magnitude faster. Existing hardware is good for training LLMs, but very poorly suited to subsequent inference of the trained…
TL;DR
What they do: Build processors and full-stack systems that fuse memory and compute to accelerate LLM inference
Performance claim: Up to 25x faster inference and ~1/10th the cost (as stated by the company)
Founded: 2022
Notable funding: $220M Series B (May 2026 reported)
Headcount: 83 employees (reported)
Company Overview
Problem Domain
AI inference hardware — reducing latency and cost of running large language models.
Founded
2022
Industry
Computer Hardware Manufacturing
Funding Track Record
$15M (reported)
Reported seed round following exit from stealth; participating investors reported include Kindred Capital, NATO Innovation Fund, and Oxford Sciences Enterprises
£5M
ARIA Scaling Compute programme grant (reported)
$220M
Reported lead investors named
Investor Signal
“Factorial Funds, Accel, Founders Fund, Kindred Capital, NATO Innovation Fund, Oxford Sciences Enterprises, Innovia Capital, Cocoa, and several angel investors (reported)”
Founders
What we do
Related Companies
| Company |
|---|
 |
 |
 |
 |
 |