Humain turns to Groq's inference chips for Saudi Arabian AI push

Appears to already be operating Groq chips.


Despite being announced just this week, Saudi Arabia's new AI business is offering access to Groq inference chips.


The Sovereign Wealth Fund-backed Humain this week announced multi-billion dollar partnerships with Nvidia, AMD, and Amazon Web Services to build out compute across the country and elsewhere.


It has yet to clarify its relationship with Groq, a smaller chip company developing dedicated inference hardware as well as its own GroqCloud.


But on the new Humain website, the company claims that it supports some 1.5 million developers currently in partnership with Groq (on Groq's own cloud website, it only notes 1m+ developers overall).


On the site, Humain states that it offers access to Groq's LPU chips to allow developers to run open models like Llama, Mixtral, and Whisper "with blazing-fast inference."


Groq first deployed a cluster in Saudi Arabia in December 2024, and this February announced a $1.5 billion deal to expand in the country with a Dammam data center. At the time, it demonstrated its hardware alongside Aramco Digital CEO Tareq Amin.


Amin is now CEO of Humain.

DCD has contacted Groq for clarification on whether those investments and existing clusters have shifted to Humain ownership.

Read Also
New CPC Solution Tackles Growing Liquid Cooling Needs for AI
Waste heat from Météo France supercomputers to be used in Toulouse district heating system
Stack secures AU$1.3bn green financing in Australia to fund Melbourne campus

Research