Luma AI
Category: Foundation Models / LLMs
A multimodal AI foundation lab developing 'World Models' that understand and simulate the physical world through high-fidelity 3D and video generation. Luma AI was founded in 2021. The company is led by Amit Jain. Based in Redwood City, United States. Team size: 101-500. Total funding raised: $1.3B+. Latest round: Series C. Key investors include HUMAIN (Saudi PIF), Andreessen Horowitz (a16z), NVIDIA, Amazon, AMD Ventures, Amplify Partners, Matrix Partners.
AMW Analysis
Luma AI is a Palo Alto-based foundation model lab founded in 2021, building "World Models" aimed at simulating the physical world through 3D and video generation, with a stated value proposition of physics-consistent visual generation and reasoning. The company has raised $1.07B in total funding, most recently a $900 million Series C backed by investors including HUMAIN (Saudi PIF), Andreessen Horowitz, NVIDIA, Amazon, AMD Ventures, Amplify Partners, and Matrix Partners, tied to a partnership with HUMAIN on Project Halo, a 2 gigawatt AI supercluster in Saudi Arabia. Since that raise, product news has centered on consolidating fragmented generative workflows into a single system: the Unified Intelligence architecture and Uni-1 model, paired with Luma Agents, are positioned to maintain creative context across text, image, video, and audio generation in one pipeline rather than separate tools. This has extended into applied deployments, including production-services work with Wonder Project through the Innovative Dreams venture, which is set to debut a series on Amazon Prime Video using Luma Agents for real-time hybrid filmmaking. Ray3 usage is also priced on a per-task basis. Together, the news flow shows Luma moving from foundation-model development toward productized creative tooling and media-industry partnerships following its large capital raise.
AMW analysis, generated from 6 tracked news signals.
- Founded
- 2021
- Headquarters
- Redwood City, United States
- Team size
- 101-500
- Total funding
- $1.3B+
Value proposition
Provides physics-consistent, high-fidelity visual generation and reasoning models that bridge the gap between generative AI and physical reality simulation.
Products and solutions
Luma Agents / Dream Machine (agentic creative platform for video, image, audio, text), Ray3 / Ray3.2 (reasoning / frame-directable cinematic video models & API), Uni-1 / Uni-1.1 (unified reasoning image generation & editing API), Layers (Uni-1-powered layered image editing), Luma Skills (repeatable agent workflows), Luma API (enterprise multimodal model integration), Genie (generative 3D foundation model), Luma App (NeRF / Gaussian Splatting 3D capture), Open Physical AI Lab (open science physical AI research), Project Halo (2GW AI supercluster with HUMAIN)
Unique value
Pioneered the transition from static 3D capture to 'World Models'—AI systems that don't just generate pixels but understand physical laws and spatial reasoning.
Target customer
Creative professionals, VFX and film studios, game developers, e-commerce enterprises, and government infrastructure partners.
Industries served
Entertainment & Media, Gaming & Metaverse, E-commerce & Advertising, Robotics & Simulation, Government & Infrastructure
Technology advantage
Combines proprietary neural rendering (NeRF/Gaussian Splatting) with massive-scale compute via the 2GW 'Project Halo' supercluster and specialized regional models like the Arabic World Model.
How they differentiate
Focuses on 'World Models' that prioritize physical reasoning and spatial consistency over simple pixel generation, utilizing proprietary neural rendering and 3D foundation models (Genie).
Main competitors
OpenAI (Sora), Google (Veo), Runway, Kling, Pika Labs, Black Forest Labs
Key partnerships
HUMAIN (Saudi PIF-backed; Project Halo 2GW supercluster; HUMAIN Create / Arabic-native models), Publicis Groupe Middle East (preferred generative AI creative partner, MENA), Serviceplan Group (global creative AI deployment), Wonder Project / Innovative Dreams (AWS-backed hybrid AI+filmmaking production JV), AMD & TensorWave (production inference partnership), Amazon / AWS (Series C investor; Innovative Dreams support), NVIDIA (compute / historical investor), Adobe (Ray2 / Firefly integration announced 2025)
Notable customers
Serviceplan Group, Publicis Groupe, Dentsu, Boundless (Mazda commercial), Wonder Project / Innovative Dreams, Mazda, Aperture, Uncharted, HUMAIN Create users, VFX/film studios and enterprise creative teams
Major milestones
Launched Dream Machine high-fidelity video generation, Launched Ray3 reasoning video model (Sep 2025); Ray3.2 with frame-level control (Jun 2026), $900M Series C led by HUMAIN at ~$4B valuation (Nov 2025); Project Halo 2GW supercluster, Opened London office (Dec 2025) and Riyadh office; Publicis MENA preferred AI partner (Feb 2026), Serviceplan Group global creative AI partnership (Feb 2026), Wonder Project + Luma launch Innovative Dreams AI production company with AWS (Apr 2026), Launched Uni-1/Uni-1.1 image models & API; Layers; Luma Agents & Skills (2025–2026), Announced Open Physical AI Lab (Jun 2026), Named to CNBC Disruptor 50 2026
Growth metrics
Official site: 250+ team members; 30M+ users worldwide; Total funding $1.3B+; offices Redwood City (HQ per careers), London, Riyadh; Project Halo 2GW supercluster partnership with HUMAIN.
Market positioning
Frontier AI research lab transitioning from 3D capture tools to large-scale physical world simulation and reasoning video models.
Geographic focus
North America and the Middle East (specifically Saudi Arabia via Project Halo).
Patents and IP
Holds multiple patents in neural rendering, AI-driven data security, and decentralized storage (led by CTO Alex Yu).
About Amit Jain
Amit Jain is the co-founder and CEO of Luma AI. Prior to founding Luma, he worked at Apple as a Systems and Machine Learning Engineer on the Vision Pro team (passthrough and related CV/camera work) and earlier LIDAR integration for iPhone. He also has experience at Product Hunt and Circle Medical. He studied Mathematics and Physics at university (not a Stanford PhD).
Latest news about Luma AI
- Luma and Wonder Project launched Innovative Dreams, a production-services venture that will debut the AI‑enhanced series “The Old Stories: Moses” on Amazon Prime Video. Their Luma Agents enable real‑t
- Luma AI just launched Unified Intelligence architecture with Uni-1 model and Luma Agents, tackling the industry's fragmented AI workflow problem where separate models lose creative context. Their earl
- Luma AI just launched its Unified Intelligence architecture with Uni-1, a single multimodal model trained across audio, video, image, language, and spatial reasoning. Unlike fragmented AI toolchains t
- Luma just launched Luma Agents powered by Unified Intelligence, a single multimodal reasoning system that coordinates text, image, video, and audio generation end-to-end—replacing the fragmented multi
- Peeyush Ranjan and Mukesh Bansal launched Luma, an AI-first edtech startup across the US and India teaching students how to think. With a 10-member tech team, the platform targets Math and Physics to
More Foundation Models / LLMs companies
Official website: https://lumalabs.ai