Groq 3 LPX inference rack
Entered full production August 24, 2026. Nebius is the first cloud to deploy it, with racks expected online later this year.
The numbers
What it is
An inference-focused rack built on the LPU technology NVIDIA licensed from Groq in a non-exclusive deal, paired with Vera Rubin NVL72 for low-latency token generation for agents. Groq remains independent and keeps its cloud business.
Why it matters to the stock. It is NVIDIA's answer to low-latency inference competition and is a test of how the Groq licensing deal turns into revenue.
News
NVIDIA Groq 3 LPX enters full production; Nebius first AI cloud to deploy
NVIDIA said its Groq 3 LPX inference accelerator is in full production as an extension of the Vera Rubin platform, aimed at faster token generation for agents.
GTC 2026: seven Vera Rubin chips in full production; Groq 3 LPX rack unveiled
At GTC 2026 NVIDIA said seven Vera Rubin chips are in full production and announced five rack types, including the Groq 3 LPX inference rack.