supplier

Groq

Maker of the LPU, a specialized chip for running (not training) models fast. In December 2025 Nvidia licensed its technology and hired away its founder, its president and other staff, paying $13.0B at closing and $4B more within a year by Nvidia's own filing (the press had reported $20B).

2 deals (2 T1) · 1 flow on the money map ·

Deals 2

sort: · · ·
  1. Dec 24, 2025 updated Oct 8
  2. Feb 2025 updated Nov 21

Money in and out 1 ribbon

Who pays Groq (left), and whom it pays (right). Width is a rough size in $B, per year for multi-year deals.

← swipe to see the whole chart →

sourced deal · 1 of 1

As a table · 1 ribbon
Notes in the record on Groq

LPU inference chips; KSA $1.5B commitment; core tech licensed to Nvidia 12/2025 — $13.0B at closing + $4B within a year (Nvidia FY2026 10-K; ~$20B press-reported — see nv-groq-license), founder + team departed; independent under new CEO. Jun 22 2026: raised $650M of new growth capital for its AI inference cloud, led by Disruptive and Infinitum (Groq's own release; no valuation given)