Home 9 AI 9 Luma AI Raises $900M to Train Models on HUMAIN’s Project Halo

Luma AI Raises $900M to Train Models on HUMAIN’s Project Halo

by | Nov 26, 2025

New funding supports training large-scale multimodal models on a 2-gigawatt compute cluster to improve systems that interpret video, audio and language for real-world tasks
(L–R) Amit Jain, CEO of Luma AI, and Tareq Amin, CEO of HUMAIN, at the U.S.–Saudi Investment Forum in Washington, D.C., where Luma AI announced its $900 million Series C and partnership on Project Halo – a 2-gigawatt AI supercluster in Saudi Arabia that will power next-generation World Models and accelerate the path toward Multimodal AGI.

PALO ALTO, CA and WASHINGTON, DC – Luma AI has announced that it has raised $900M in Series C funding, led by HUMAIN, with participation from AMD Ventures, and investors Andreessen Horowitz, Amplify Partners, and Matrix Partners.

Luma AI is raising the stakes to build multimodal general intelligence, and its new series C will accelerate that effort by tapping HUMAIN’s project Halo supercluster. Luma plans to train its next-generation models on Project Halo, a 2-gigawatt AI supercluster in Saudi Arabia. The partnership connects Luma’s work on AI that can generate, understand and operate in the physical world with HUMAIN’s expanding infrastructure.

Luma AI and HUMAIN used the U.S.–Saudi Investment forum in Washington, D.C., to outline a roadmap to push multimodal intelligence forward. The announcement, made during the visit of Crown Prince Mohammed bin Salman, details how new funding and compute access will support Luma’s work on large-scale “World Models,” AI systems trained on video, audio and language. The companies aim to build products inside HUMAIN Create that can simulate real-world conditions for robotics, entertainment, advertising, gaming and personalized education.

“HUMAIN is the perfect partner for this next stage in Luma AI’s explosive trajectory,” said Amit Jain, CEO and co-founder of Luma AI. “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. We are also partnering deeply across customized models, go-to-market, and deployment of frontier capabilities to build a full end-to-end value-chain for multimodal AI.”

“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,” said Tareq Amin, CEO of HUMAIN. “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. Our investment in Luma AI, combined with HUMAIN’s 2GW supercluster, positions us to train, deploy, and scale multimodal intelligence at a frontier level. This partnership sets a new benchmark for how capital, compute, and capability come together.”

The 2-gigawatt supercluster will increase the compute available for multimodal AI work. Luma is building technology that can train on peta-scale multimodal data, a volume far larger than what current text-based large language models (LLMs) process. Luma says this capacity is needed for AI systems that interpret video, audio and language for real-world tasks. The cluster will also host inference systems that run these models and deliver real-time results to global users. “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,” added Jain.

Ray3 has demonstrated how Luma AI can turn initial research into commercial products, deployed in studios, advertising firms, including integration within Adobe products. With this funding, Luma AI plans to expand from entertainment and advertising into simulation, design and robotics.

Luma AI models were also the first launched within HUMAIN Create, a regional program aimed at developing AI models and products for Arabic-speaking users. These models are designed to handle cultural context, visual details and language variation so creators, businesses and governments can use AI that reflects local needs and identity.

Source: Luma AI

About Luma AI

Luma AI is a U.S.-based software company founded in 2021 that builds multimodal AI systems for generating and interpreting visual content. The company develops tools that create images, videos and 3D visuals for use in media, design, entertainment and other creative fields. Its flagship platform, Dream Machine, allows users to produce video and image content, while Ray3, introduced in 2025, generates videos and animations that follow real-world physical behavior. Luma also offers Photon and a developer API for teams that want to integrate its models into existing workflows. Its technology is used by entertainment studios, advertising firms and technology companies. Luma provides access through subscriptions and API plans. The company is headquartered in Palo Alto, CA, and is backed by investors including HUMAIN, Andreessen Horowitz, Amazon, AMD Ventures, NVIDIA, Amplify Partners, Matrix Partners and several individual investors.

About HUMAIN

HUMAIN is a Saudi Arabian AI company that develops AI infrastructure, cloud services, data centers and advanced AI models. The company provides next-generation data centers, high-performance cloud platforms and Arabic-language large language models for public- and private-sector organizations. It also builds multimodal AI tools that support customers in technology, government, media and other industries. HUMAIN delivers full-stack AI capabilities and designs end-to-end systems for clients that require regional or localized AI resources. The company was founded in 2025 and is headquartered in Riyadh, Saudi Arabia. It is backed by the Public Investment Fund (PIF) and supports both domestic and international users.