Luma AI Raises $900 Million From Saudi-Backed HUMAIN
The startup secures major backing and gains access to one of the world’s largest AI compute platforms to train next-gen multimodal models
Luma AI has secured $900 million in new financing in a round led by HUMAIN, the Saudi sovereign wealth fund-backed AI firm. The startup, known for Dream Machine and the Ray model, says the capital will accelerate its work on large-scale World Models. The latest infusion brings Luma’s total funding to more than $1 billion; the company’s current valuation hasn’t been disclosed, though it’s reported to be around $4 billion.
In addition, Luma announces it will be one of the first customers of Project Halo, a two-gigawatt AI supercluster HUMAIN is building in Saudi Arabia. When completed, it stands to be one of the largest dedicated AI compute deployments ever built.
All of this comes as Saudi Crown Prince Mohammad bin Salman Al-Saud begins an official state visit to the United States. It also unfolds alongside the U.S.-Saudi Investment Forum, where reports indicate that Luma will not be the only company expected to receive financial backing from HUMAIN.
All of this coincides with the official U.S. visit by Saudi Crown Prince Mohammad bin Salman Al-Saud, and also on the sidelines of the U.S.-Saudi Investment Forum, where it’s reported that Luma won’t be the only beneficiary of HUMAIN funds
A Milestone Series C Funding Round
“HUMAIN is the perfect partner for this next stage in Luma’s explosive trajectory,” Luma chief executive and co-founder Amit Jain says in a statement. “To create AI that can help humanity in the physical world and expand our understanding of the universe, we need to build systems that can learn from a quadrillion tokens of information—roughly the collective digital memory of humanity—contained in video, image, audio, and language. HUMAIN is deploying frontier compute infrastructure at impressive speed, and this is critical to achieving Luma AI’s mission.”
But HUMAIN isn’t the only investor in this round. Luma also received “significant participation” from AMD and Amazon, along with existing investors Andreessen Horowitz, Amplify Partners, and Matrix Partners.
Luma’s mission is to build multimodal general intelligence, which it defines as AI capable of generating, understanding, and operating in the physical world. This isn’t the same as artificial general intelligence that OpenAI and Meta have been talking about for months. Luma’s pursuit is to develop models that can perceive, reason about, and act in real-world environments. In other words, it wants to build AI for robots, sensors, and mixed-reality applications.
To succeed, it requires not only a healthy war chest but also incredible computing power. Like many other AI companies and startups, Luma is going to the one place that can help make this happen: the Middle East.
This isn’t the first meeting between Luma and HUMAIN either. The two have partnered to develop new ways of interactive storytelling, gaming, and entertainment.
“This investment underscores an important point in HUMAIN’s philosophy: we are not only funding the next wave of AI, we’re building the full value chain that makes it possible,” HUMAIN chief executive Tareq Amin remarks. “Luma AI is an exceptional U.S.-based global frontier startup pushing the boundaries of multimodal world models. Their technical ambition, research velocity, and proven ability to turn foundational breakthroughs into real products make them uniquely aligned with HUMAIN’s vision.”
Tapping Into a Supercluster
To generate the compute required to realize multimodal intelligence, Luma AI is teaming with HUMAIN to develop one of the largest compute infrastructure buildouts. By building out a two-gigawatt AI supercluster, both companies seek to “train, deploy, and scale multimodal intelligence at a frontier level.”
When completed, it will support training on peta-scale multimodal datasets, which are up to 10,000 times larger than those handled by today’s LLMs. Luma plans to use the cluster not only for training, but also for real-time, global inference, enabling its AI to operate on real-world tasks at scale. As Jain points out, “it is only through this level of systems and algorithmic integration that it is possible to build models that can understand and simulate the universe.”
There is no timeline for when Project Halo is expected to be completed.
The scale of compute they're planing to deploy is staggering. Training on peta-scale multimodal datasets could really change what's posible with world models. The partnership with HUMAIN seems stratigic, especially given the global race for AI infrastructure. Curious how this will impact their competitiveness against other multimodal systems.