Luma AI has finalized a $900 million Series C funding round, led by HUMAIN, a Public Investment Fund company specializing in full-stack AI solutions.
The investment, which also saw participation from AMD Ventures and existing backers like Andreessen Horowitz, fuels the company’s ambition to develop multimodal Artificial General Intelligence capable of simulating and interacting with the physical world. This financing is directly coupled with a strategic partnership focused on infrastructure development in Saudi Arabia.
HUMAIN will construct Project Halo, a massive 2-gigawatt AI supercluster, positioning it among the world’s largest compute builds dedicated to AI training. Luma AI intends to utilize this dedicated compute capacity to train its next-generation World Models, which represent a progression beyond current Large Language Models. These foundational models learn across video, audio, and language data, aiming to capture a broader scope of human digital memory for realistic simulation.
The partnership framework establishes Luma AI as a primary customer of the new HUMAIN infrastructure, creating a vertically integrated pipeline from foundational research to scaled deployment.
This synergy is crucial for training models on the peta-scale multimodal datasets required for advanced real-world utility. The announcement occurred during the U.S.–Saudi Investment Forum in Washington, D.C., highlighting the geopolitical focus on securing advanced AI capabilities. This investment mirrors broader market activity, considering that OpenAI competitor Anthropic also recently closed a substantial funding agreement, indicating strong investor interest in frontier model development.
Beyond core simulation capabilities, the partnership includes HUMAIN Create, an initiative focusing on developing AI models specifically trained on Arabic and regional data. This effort aims to provide culturally aligned AI tools for enterprises and governments across the Middle East and North Africa (MENA) region.
Luma AI’s existing platform, notably its Ray3 video model, already sees commercial deployment within major studios and brands, including integration into Adobe’s product suite.
The fresh capital will expand this focus beyond entertainment and advertising into areas like robotics and advanced design simulation.
Consequently, the combination of substantial capital and guaranteed access to frontier compute resources positions Luma AI to accelerate its roadmap toward deploying AI systems that understand and operate effectively outside purely digital environments.
This infrastructure commitment is a prerequisite for achieving the scale necessary for true world modeling.
