Intel / video

"An endless demand for compute" | Jonathan Ross, founder of Groq

Source: youtube.com
Author: Julia TurcTop Viber
Date: Published Jun 11, 2026

Context

The cheapest way to run an AI model depends partly on how quickly its answer must arrive. In this interview, the founder of -chip company Groq acknowledges that GPUs can offer better economics when processing large batches of requests. The founder presents Groq's processors as attractive toward the faster-response end of that tradeoff.

The explanation centers on spreading a model across chips with fast on-chip memory and scheduling computation in advance. Those choices aim to reduce waiting during generation. This is the supplier's account, rather than a matched hardware , but it identifies the comparison that matters: cost at the response speed an application needs, rather than one universal winner on price or speed.

Terms in this piece · Glossary

inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.

Key quotes

“As long as there are unsolved problems in civilization, we will have a need for more compute.”
Jonathan Ross

“Why that's leading to an absolute explosion in the usage is AI using AI using AI.”
Jonathan Ross

“My response was how hard could it be? Turns out it was very hard but we didn't know so jumped in.”
Jonathan Ross

“If you can come up with the right question, AI can go answer it for you.”
Jonathan Ross

“The more compute we provide for the world, the more compute people are going to want.”
Jonathan Ross

Read the source www.youtube.com