Luma AI Secures $900M Investment Led by Saudi’s Humain

Video generation startup Luma AI has secured a staggering $900 million in a funding round spearheaded by Humain, an artificial intelligence entity backed by Saudi Arabia’s Public Investment Fund (PIF). This investment underscores the growing significance of generative AI and Saudi Arabia’s ambition to become a major player in the global AI landscape.

The round, which also saw participation from Advanced Micro Devices’ venture capital arm, alongside existing investors Andreessen Horowitz, Amplify Partners, and Matrix Partners, was unveiled at the U.S.-Saudi Investment Forum on Wednesday. This move elevates Luma AI’s valuation to over $4 billion, according to sources familiar with the matter.

Luma AI is pioneering the development of multimodal “world models,” sophisticated AI systems capable of learning from a confluence of data types, including text, video, audio, and images, to construct realistic simulations. CEO Amit Jain, in an interview, highlighted that these models surpass traditional large language models (LLMs) by extending their capabilities beyond text-based analysis, thus facilitating more effective interaction within the physical world.

“With this funding, we plan to scale our and accelerate our efforts in training and then deploying these world models today,” Jain stated, signaling an aggressive push towards deploying their advanced AI capabilities.

The company’s flagship product, Ray3, released in September, is a reasoning video model adept at interpreting prompts to generate videos, images, and audio. Luma AI claims that Ray3 currently delivers performance metrics that surpass OpenAI’s Sora 2 and are comparable to Google’s Veo 3, based on internal benchmarks. This positions Luma AI as a formidable competitor in the rapidly evolving video generation space.

Humain, launched earlier this year, aims to establish Saudi Arabia as a global AI hub by delivering a comprehensive suite of AI capabilities. The company is helmed by Tareq Amin, a seasoned industry executive who previously held leadership positions at Aramco Digital and Rakuten Mobile. His experience in telecommunications and digital transformation brings significant expertise to Humain’s mission.

Beyond the investment, Luma AI and Humain are collaborating to construct Project Halo, a massive 2-gigawatt AI supercluster located in Saudi Arabia. According to Jain, this infrastructure will represent one of the largest deployments of graphic processing units (GPUs) worldwide, enabling the training of extremely large and complex AI models. Industry analysts suggest that this supercluster will likely utilize cutting-edge GPU architectures from Nvidia or AMD, providing the computational horsepower needed to push the boundaries of AI research and development. The strategic importance of such infrastructure cannot be overstated, as the availability of powerful compute resources is a key bottleneck in AI development today.

This investment mirrors a broader trend of tech companies investing heavily in supercomputing infrastructure. Meta recently announced its plans for a 1-gigawatt supercluster, Prometheus, while Microsoft has deployed a supercomputing cluster using Nvidia’s GB300 NVL72 platform. These investments reflect the growing demand for computational resources required to train increasingly complex AI models.

“Our investment in Luma AI, combined with HUMAIN’s 2GW supercluster, positions us to train, deploy, and scale multimodal intelligence at a frontier level,” Amin stated. “This partnership sets a new benchmark for how capital, compute, and capability come together.” The sheer scale of the proposed supercluster indicates a significant bet on the future of AI and its potential to transform various industries.

The partnership also includes the Humain Create initiative, aimed at developing sovereign AI models trained on Arabic and regional data. This initiative underscores the importance of cultural representation in AI-generated content. Jain emphasized that Luma AI’s models and capabilities will be deployed to businesses across the Middle East, highlighting a commitment to regional economic development and technological advancement.

Jain also addressed the issue of data bias in AI models, noting that current models often underrepresent regions outside of the U.S. and Asia due to the reliance on internet-scraped data. “It’s really important that we bring these cultures, their identities, their representation — visual and behavioral and everything — to our model,” Jain said, emphasizing the need for inclusive and culturally sensitive AI development. This is a crucial consideration as AI becomes more pervasive in various aspects of life.

The rapid advancements in AI-generated content have also raised concerns regarding copyright infringement. Luma AI’s Dream Machine platform faced accusations of IP copying earlier this year. Jain addressed these concerns, stating that the company has implemented robust safeguards, including AI-powered detection systems, to prevent unauthorized use of copyrighted material. “Even if you really try to trick it, we are constantly improving it,” he claims. The ongoing battle against copyright infringement will likely continue to shape the development and deployment of AI-generated content tools.