Humain Leads $900M Series C for Luma AI’s Multimodal World Models

Humain secures Luma AI’s $900M Series C, backing their multimodal AI to train world models via Saudi Arabia’s 2 GW Project Halo supercluster.

Humain, the Saudi Arabia–based full-stack AI company backed by the Public Investment Fund (PIF), has led a $900 million Series C financing round for Luma AI, a U.S. company building advanced multimodal generative intelligence. 

Also Read: AWS, Humain Forge $5B+ AI Partnership to Launch “AI Zone”

The deal was formally announced in the U.S.–Saudi Investment Forum, coinciding with Crown Prince Mohammed bin Salman’s visit to Washington. 

Luma AI, founded in 2021 by Alberto Taiuti and Amit Jain, is developing what it calls “World Models”: AI systems capable of learning from video, audio, images, and language to simulate and understand real-world environments. 

With the new funding, Luma plans to accelerate the training of these large-scale models and expand into industries like robotics, simulation, entertainment, design, and personalized education. 

As part of the agreement, Luma AI will be a key customer for Humain’s Project Halo, a planned 2-gigawatt AI supercluster in Saudi Arabia. 

Humain CEO Tareq Amin said the partnership “positions us to train, deploy, and scale multimodal intelligence at a frontier level,” while Luma’s Amit Jain added that access to massive computers is “critical” for building their next-generation world models. 

Beyond compute, the two firms will work together on Humain Create, an initiative aiming to develop culturally aligned AI models trained on Arabic and regional data. This is designed to help enterprises, governments, and creators use AI that reflects their own language, culture, and identity. 

Strategic Implications

• The deal underscores Saudi Arabia’s ambition to become a global AI leader, not just via funding but also by building its own computer infrastructure.
• By training models locally, the partnership aims to contribute to AI sovereignty, reducing dependence on foreign infrastructure.

There’s a clear bet on multimodal AGI—moving beyond language models to systems that can operate in the physical world.