(1) For the tech word with a "k", see
grok.
(2) For the AI chatbot, see
xAI.
(3) (Groq, Inc., Mountain View, CA) An AI company founded in 2016 by Google Engineers and Jonathan Ross, one of the designers of Google's Tensor Processing Unit. Originally called a Tensor Streaming Processor (TSP) Groq's Language Processing Unit (LPU) is a chip designed for linear algebra calculations, the primary processing required for inference execution. The LPU was designed to not only be faster than a GPU but to be ten times more energy efficient. See
Tensor Processing Unit.
GroqCloud
In 2024, Groq changed its focus to a cloud computing service using its LPUs to power several language models, including GPT and Llama, as well as Canopy text-to-speech and Whisper speech recognition models.
NVIDIA Licenses Groq Chips
In late 2025, NVIDIA entered into a non-exclusive licensing arrangment with Groq for its LPUs. Paying $20 billion for the license, this was NVIDIA's largest acquisition to date. Groq remains an independent company that continues to offer GroqCloud services to customers.