Luma AI

Luma AI (operating as Luma Labs, Inc., and increasingly branded simply as Luma) is a generative AI company that builds text-to-video, 3D, and image models, best known for its Dream Machine platform and its Ray series of video models. Its own company pages list its headquarters as Redwood City, California, although press coverage through mid-2026 described it as Palo Alto-based.^[41][42][48] Founded in 2021 by Amit Jain, Alex Yu, and Alberto Taiuti, the company first gained attention for neural radiance field (NeRF) based 3D capture from smartphone video before expanding into video generation. In November 2025 Luma AI raised a $900 million Series C led by Saudi Arabia's HUMAIN at a valuation of more than $4 billion, the largest single round in its history, and partnered with HUMAIN on Project Halo, a planned 2-gigawatt AI supercluster in Saudi Arabia.^[2][40][47]

Luma AI's flagship Ray series powers text-to-video, image-to-video, and video-to-video generation through the consumer-facing Dream Machine application (rebranded during 2026 as the Luma App) and a developer API. Its September 2025 model, Ray3, was marketed as "the world's first reasoning video model and the first to generate high fidelity 16 bit HDR."^[5] When Ray3 launched, Luma said Dream Machine powered creation for more than 30 million users.^[5] As of October 2026, no funding round after the November 2025 Series C had been announced; Luma's careers page listed more than 250 team members and total funding of more than $1.3 billion.^[41]

During 2026, the company expanded beyond video generation into agentic creative software with Luma Agents, unified image understanding and generation with its Uni-1 model family, and open robotics research through the Open Physical AI Lab, while signing advertising-industry partners including Serviceplan Group and Publicis Groupe.^{[28][29][33][37]} In the second half of the year it built out the agent workspace with reusable Skills, layer-based image editing, storyboard-first video, and one-click "mini-apps," and added third-party video models such as Seedance 2.5, MiniMax H3, and Kling 4.0.^{[50][53][54][55][56][61][62]} Two of the three co-founders, Yu and Taiuti, had left the company by October 2026, and Chief Scientist Jiaming Song was reported to have departed in May 2026.^{[43][44][45][46]}

History

Who founded Luma AI, and when?

Luma AI was co-founded in 2021 by Amit Jain, Alex Yu, and Alberto Taiuti. Jain and Taiuti both previously worked at Apple. Jain was a systems and machine learning engineer there; he led development of the Passthrough feature for the Apple Vision Pro and worked on integrating the first LiDAR sensors into the iPhone.^[15][68] Taiuti worked at Apple on AR/VR, and his other prior roles included Senior Autonomy Software Engineer at Skydio and a position at Codeplay Software; he studied at the University of Abertay Dundee in Scotland.^[44] Alex Yu was a researcher at UC Berkeley, where he worked with Professor Angjoo Kanazawa and co-authored neural rendering papers including pixelNeRF (CVPR 2021), PlenOctrees (ICCV 2021), and Plenoxels (CVPR 2022).^[43] When TechCrunch profiled the company in January 2024, it described Yu as having been an AI researcher at UC Berkeley and Jain as an Apple employee building out the Vision Pro's multimedia experiences at the time they founded Luma.^[3]

The company's initial product focused on NeRF-based 3D capture, allowing users to create photorealistic 3D scenes from smartphone video. The technology used deep learning networks to process 5D coordinates (location and view direction), outputting volume density and RGB radiance for each point in the scene. Users could walk around a subject while recording video with a phone, and the AI would analyze camera movement and scene geometry to reconstruct a full 3D representation.

Startup funding databases record a seed round of roughly $4 million to $4.3 million in October 2021; Luma has not published the round's terms.^[74] By November 2024 the company named Andreessen Horowitz, Amplify Partners, Matrix Partners, General Catalyst, and South Park Commons among its investors.^[67] Luma's smartphone capture app, launched in 2021, had passed two million users by January 2024, according to TechCrunch.^[3]

Series A and 3D Capture Growth (2023)

On March 20, 2023, Luma AI announced a $20 million Series A round led by Amplify Partners, with participation from NVentures (NVIDIA's venture arm) and General Catalyst. Further terms, including valuation, were not disclosed.^[69] In April 2023 the company named Kanazawa, a UC Berkeley professor associated with the nerfstudio framework, as an advisor; Radiance Fields headlined the role as Chief Scientific Advisor.^[71]

During this period, the company released its iOS app for 3D capture, which allowed anyone with an iPhone 11 or newer to create lifelike 3D models without specialized equipment such as LiDAR scanners. The app supported exporting to industry-standard formats including USDZ, glTF, and OBJ, making captures compatible with tools like Blender, Unity, and Unreal Engine.

In April 2023, Luma AI released a plugin for Unreal Engine, allowing users to import and render photorealistic NeRF captures directly inside the game engine. The plugin included tools for scene cleanup, cropping, culling, and relighting via analytical or sun lights.

In June 2023, Luma AI hired Matthew Tancik, a co-author of the original NeRF paper and co-founder of the nerfstudio framework.^[1][70] He had previously collaborated with co-founder Alex Yu on papers such as PlenOctrees and Plenoxels.^[70]

In November 2023, Luma AI launched Genie (unrelated to Google DeepMind's world model of the same name) as a research preview, a text-to-3D generation tool first offered on Luma's Discord server.^[1][3] Genie attracted tens of thousands of users on its first day, and Jain told TechCrunch it had reached 100,000 users in four weeks.^[1][3]

Series B, Genie, and Key Hires (January 2024)

On January 9, 2024, Luma AI raised $43 million in a Series B round led by Andreessen Horowitz (a16z), with other new and existing backers participating.^[1][3] TechCrunch, citing a source familiar with the matter, reported that the round valued Luma at between $200 million and $300 million, and that the company's total funding stood at more than $70 million. Luma also said it would train new models on a cluster of about 3,000 NVIDIA A100 GPUs.^[3] Alongside the funding announcement, Luma released Genie 1.0 for general availability.^[1]

a16z's investment thesis cited Luma's position in generative 3D, writing that "generative 3D models are on the cusp of their own breakthrough into the mainstream" following inflection points in 2D image generation in 2022 and computer vision in 2012.^[1] The firm highlighted the strength of Luma's technical team, which by this point included Jiaming Song as Chief Scientist.^[1] Song, who holds a Ph.D. from Stanford University under advisor Stefano Ermon,^[46] created the DDIM (Denoising Diffusion Implicit Models) algorithm, which the MIT Technology Review profile says accelerates diffusion models by up to 50 times while keeping sample quality.^[14] DDIM was subsequently adopted in production systems including DALL-E 2, Imagen, and Stable Diffusion.^[14]

When did Dream Machine launch?

On June 12, 2024, Luma AI launched Dream Machine, a text-to-video generation platform powered by its first Ray model. The tool allowed anyone to type a descriptive prompt and generate a short high-definition video clip in minutes. Dream Machine gained one million users within four days of launch.^[14] By November 2024 Luma said it had 25 million registered users.^[67]

The launch attracted wide attention on social media. Users created video versions of images generated with Midjourney, as well as moving recreations of famous artworks such as "Girl with a Pearl Earring" and popular internet memes including Doge and the distracted boyfriend meme. Los Angeles director Ellenor Argyropoulos posted a Pixar-style animation of a girl in ancient Egypt that, according to Futurism, drew 7.5 million views on X within a few days.^[66] Not all of the attention was positive: a Dream Machine trailer called "Monster Camp," posted by Luma, drew criticism for featuring a recognizable Mike Wazowski from Pixar's Monsters, Inc. (see Reception and criticism below).^[65]

Photon Image Model and Dream Machine Service (November 2024)

On November 25, 2024, Luma AI introduced Photon, a text-to-image foundation model.^[9] The Photon family included two variants: Photon (full quality) and Photon Flash (optimized for speed). Luma claimed the models were 800 percent faster and cheaper than comparable image generation services.^[9] Luma's Photon page priced Photon at 1.5 cents per 2-megapixel 1080p image and Photon Flash at 0.4 cents per image.^[73]

Photon introduced several capabilities new to image generation at the time, including a large context window for visual generative models and consistent character generation from a single reference image (in beta). Alongside Photon, Luma relaunched Dream Machine as a free and subscription service with paid tiers at $9.99, $29.99, and $99.99 per month (described as hobbyist, explorer, and professional tiers) plus an enterprise tier.^[9][67]

Series C and Ray2 (December 2024 - January 2025)

On December 6, 2024, Luma AI raised $90 million in an initial Series C round with participation from new investors Amazon and AMD, alongside existing backers a16z, Amplify Partners, and Matrix Partners.^[16][17] This brought total funding to approximately $157 million.^[16] Other new participants in the round included Factorial Funds, LDV Capital, and Hanwha.^[17]

On January 15, 2025, Luma AI released Ray2, its second-generation video model.^[4] Ray2 was built on a new multi-modal architecture with 10x the compute power of Ray1.^[4] The model was trained directly on video sequences, teaching it to understand motion as continuous flow rather than discrete snapshots.

Ray2 supported video generation at 540p, 720p, and 1080p resolution with 4K upscaling, producing clips of 5 to 10 seconds that could be extended to approximately 30 seconds. The model introduced structured prompting, seamless looping, camera movement controls (crane, tracking, dolly), keyframe control, style references, and color grading. Ray2 was available exclusively to paid Dream Machine subscribers, via the iOS app, and through the developer API.

Luma AI's Ray2 model was also made available through Amazon Bedrock, making AWS the first cloud provider to offer Luma's video models to enterprise customers.^[10] Under the partnership, AWS became the preferred compute partner for Luma AI, and Luma began training its foundation models on Amazon SageMaker HyperPod infrastructure.^[18]

Ray2 Flash, Modify Video, and Dream Lab LA (March-August 2025)

In March 2025, Luma released Ray2 Flash, a variant of Ray2 that the company described as 3x faster and 3x cheaper, bringing Ray2's text-to-video, image-to-video, audio, and control capabilities to all paid subscriber tiers.^[19] The same month, Luma's research team published Inductive Moment Matching (IMM), a paper proposing a new class of generative models that sample in one or a few steps (see Research below).^[20]

On May 15, 2025, Luma announced a strategic partnership with HUMAIN, the Saudi AI company backed by the kingdom's Public Investment Fund, spanning compute, go-to-market, and investment, and began deploying Ray2 on HUMAIN infrastructure.^[21] In June 2025, the company released Modify Video, a video-to-video tool that restyles or retextures real footage (changing wardrobe, props, environments, or overall aesthetic) while preserving the original motion and camera dynamics, without requiring motion capture or green screens.^[22] Modify with Instructions followed on August 12, 2025, adding natural-language video editing with object removal and swapping, virtual sets, character refinements, and restyling, with adjustable Adhere, Flex, and Reimagine strength levels.^[24]

In July 2025, Luma opened Dream Lab LA, a Los Angeles studio intended as a creative incubator and training ground where filmmakers and studios learn to integrate AI into production workflows. Veteran creative executive Verena Puhm was named head of the studio and Jon Finger was hired as creative workflow executive.^[23]

How big was the $900 million Series C, and what is Project Halo?

On November 19, 2025, at the U.S.-Saudi Investment Forum during Crown Prince Mohammed bin Salman's visit to Washington, D.C., Luma AI announced a $900 million Series C investment.^[2] The round was led by HUMAIN, a company owned by Saudi Arabia's Public Investment Fund (PIF) that delivers global full-stack AI solutions, with participation from AMD Ventures and existing investors Andreessen Horowitz, Amplify Partners, and Matrix Partners.^[2] CNBC confirmed that the round valued Luma AI at upwards of $4 billion.^[40][47] Amit Jain, Luma's co-founder and CEO, said: "HUMAIN is the perfect partner for this next stage in Luma AI's explosive trajectory."^[2] HUMAIN CEO Tareq Amin framed the deal as a value-chain bet, saying: "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."^[2]

As part of the deal, Luma AI became a customer of HUMAIN's Project Halo, a planned 2-gigawatt AI supercluster in Saudi Arabia that would be one of the world's largest compute infrastructure buildouts.^[2] Luma AI stated it was developing technology "capable of efficiently training on peta-scale multimodal data," handling "1,000-10,000 times more information than the current frontier Large Language Models (LLMs)."^[2] The partnership also included HUMAIN Create, an initiative to build sovereign AI models trained on Arabic and regional data, with Luma working on what it described as the world's first Arabic video model.^[2]

Ray3 and Adobe Partnership (September 2025)

On September 18, 2025, Luma AI released Ray3, which it described as "the world's first reasoning video model and the first to generate high fidelity 16 bit HDR."^[5] Ray3 was more than twice the size of Ray2 and introduced several technical firsts. It generated physically accurate videos with improved fidelity, instruction following, and temporal coherence. Ray3 was the first video generation model capable of producing High Dynamic Range (HDR) video in professional ACES2065-1 EXR format across 10-, 12-, and 16-bit pipelines, the color encoding standard used in film and television finishing.^[5] Luma said Ray3 "can actually think through what a user is asking, plan complex scenes, and judge whether its own output makes sense."^[5]

Ray3 included reasoning-driven generation, improved video-to-video workflows with character reference and keyframe support, a Draft Mode for rapid iteration, and an HDR pipeline.^[5] Adobe integrated Ray3 into Adobe Firefly and Firefly Boards as Luma's first launch partner for the model; for its first two weeks Ray3 was available only in Firefly and in Luma's Dream Machine.^[8] Users on paid Firefly or Creative Cloud Pro plans received unlimited Ray3 generations through October 1, 2025.^[8] Content generated with Ray3 in Firefly could be synced to Creative Cloud accounts for editing in Premiere Pro and other Adobe applications.

In October 2025, Luma published a Ray3 evaluation report stating that third-party evaluations placed the model ahead of competing video generators on most measured dimensions, including motion artifacts (where Ray3 exhibited artifacts less often and less severely than any other tested model) and temporal consistency.^[25]

London Office and Global Expansion (December 2025)

On December 2, 2025, Luma AI opened its first international office in London, United Kingdom. The company appointed Jason Day, a former executive at Monks and WPP, as Head of EMEA to lead international expansion across creative, advertising, gaming, and entertainment sectors.^[7] Luma said it expected to create 200 roles in London in 2026 across research, engineering, partnerships, and strategic development; CNBC reported the target as roughly 200 hires by early 2027, which it said would make up about 40% of the company's workforce.^[7][13] The announcement quoted Caroline Ingeborn, Luma's chief operating officer.^[7] The company chose London in part because of proximity to institutions like DeepMind and the strength of the UK's AI research community.^[13]

Ray3 Modify (December 2025)

In December 2025, Luma AI released Ray3 Modify, a model designed for hybrid workflows combining real-world footage with AI generation.^[11] Ray3 Modify introduced four capabilities: Keyframe Control, Character Reference, enhanced Modify Video, and Start-and-End Frame control.^[11] The model allowed creative teams to provide a start frame and end frame, then generate transitional footage between them. It could retain an actor's original motion, timing, eye line, and emotional delivery while transforming the surrounding scene, costumes, or environment.^[11] The release came on December 18, 2025; TechCrunch noted that Ray3 Modify built on the Modify Video capabilities Luma had first introduced in June 2025.^[26]

Ray3.14 and Riyadh Office (January-February 2026)

On January 26, 2026, Luma AI launched Ray3.14, an update that delivered native 1080p generation, 4x faster generation speeds, and per-second pricing that was 3x cheaper than Ray3.^[6] The model applied the reasoning engine from Ray3 more powerfully to animation and professional video workflows, producing higher detail adherence and the strongest temporal stability Luma had achieved. Modify Video support was extended to 18-second clips.^[6]

On February 11, 2026, Luma AI announced plans to open a Riyadh, Saudi Arabia office to accelerate the HUMAIN Create initiative.^[12] The company also named Publicis Groupe Middle East as its preferred AI creative partner across the MENA region.^[12] Through the partnership, Publicis would integrate Luma AI's generative video and multimodal AI technologies into creative production workflows for brands across the Middle East and North Africa.

Luma Agents, Uni-1, and the Cannes Lions Push (February-April 2026)

In early 2026, Luma mounted a sustained push into the advertising industry. On February 2, 2026, the company announced The Luma Dream Brief, a global competition offering $1 million to any team that wins a Gold Lion at the June 2026 Cannes Lions festival with work created using Luma AI; submissions ran through March 22, 2026.^[32] On February 19, 2026, Serviceplan Group, Europe's largest independent agency network with more than 6,500 employees across 43 locations, agreed to deploy Luma's creative AI across its global operations, with Luma serving as its AI technology partner.^[33]

On March 5, 2026, Luma launched Luma Agents, AI agents that handle creative workflows end to end across text, image, video, and audio while maintaining context across assets, collaborators, and iterations. The agents orchestrate Luma's own models alongside third-party systems including Google's Veo 3 and Nano Banana Pro, ByteDance's Seedream, and ElevenLabs voice models. Early customers included Publicis Groupe, Serviceplan, Adidas, Mazda, and HUMAIN, and TechCrunch reported that one customer converted a $15 million annual advertising campaign into localized ads for multiple countries in 40 hours for under $20,000.^[28]

Later in March 2026 (with availability announced on March 23), Luma released Uni-1, its first Unified Intelligence model and its first model to combine image understanding and image generation in a single architecture.^[29] Uni-1 is a decoder-only autoregressive transformer in which text and image tokens share a single sequence, allowing the model to perform structured reasoning before and during image synthesis rather than denoising images from noise as diffusion models do.^[29] According to Luma, Uni-1 ranked first on human-preference Elo ratings for overall quality, style and editing, and reference-based generation, led the RISEBench reasoning benchmark ahead of Google's Nano Banana 2 and OpenAI's GPT Image 1.5, and cost roughly 10 to 30 percent less than comparable models at high resolution.^[29][30]

The advertising campaign culminated in April 2026: Luma submitted 21 finalist entries from the Dream Brief competition to Cannes Lions after the initiative drew roughly 400 submissions in under eight weeks^[34]; Johannesburg agency Boundless used Luma Agents to deliver Mazda's first AI-produced commercial, a heritage film supporting the MX-5's South African relaunch that went from concept to approved campaign in under two weeks^[35]; and Luma joined Jon Erwin's Wonder Project to launch Innovative Dreams, a production services company, R&D lab, and post-production and visual effects firm backed by Amazon Web Services that combines performance capture, virtual production, and generative AI in what the companies call Realtime Hybrid Filmmaking. Innovative Dreams' first project, The Old Stories: Moses, a three-part companion special to House of David starring Ben Kingsley and O-T Fagbenle, was produced with Amazon MGM Studios for Wonder's subscription service on Prime Video.^[36]

Uni-1.1 API, Open Physical AI Lab, and Ray3.2 (May-June 2026)

On May 5, 2026, Luma released the Uni-1.1 API, exposing Uni-1's image generation and natural-language editing through a REST interface that supports up to nine reference images per request, with launch partners including Envato, Comfy, Runware, Flora, Krea, Magnific, Fal, and LovArt.^[31]

On June 1, 2026, the company announced the Open Physical AI Lab, an open science initiative aimed at solving generalization in physical AI. The lab invites outside robotics teams and researchers to test whether Luma's world models trained on video, image, and 3D data can help machines generalize beyond narrow tasks, and CEO Amit Jain framed the open approach as a way to prevent physical AI infrastructure from being controlled by one or a few companies.^[37] Jain told Semafor that Luma was not building robot hardware; instead, outside robotics labs and engineers could "build your own systems on top of" its technology, and he said critical robotics software and infrastructure should not be "controlled by one or two or three companies."^[48] Semafor described the lab as an expansion beyond the company's video models that would use its video training data to help robots operate in unpredictable real-world environments.^[48]

On June 9, 2026, Luma introduced Ray3.2, an update to the Ray3 family developed with input from entertainment, advertising, and gaming creatives. Ray3.2 added frame-level control with up to 16 keyframes per clip, native 1080p generation at up to 20 seconds, HDR generation with 16-bit EXR export, performance tracking for up to 8 faces simultaneously, and integrated reframing tools, and shipped alongside a developer API.^[38] Luma said it was the first time the full control surface of a Ray model had been exposed through an API.^[38]

Leadership changes (2026)

By October 2026 two of Luma's three co-founders had moved on. Alberto Taiuti co-founded Reactor, a developer platform for real-time generative video and world models, which emerged from stealth on May 28, 2026 with $59 million led by Lightspeed Venture Partners; Amplify Partners, which had led Luma's Series A, also invested.^[45][69] Reactor's announcement described Taiuti as having served as Luma's CTO and built the infrastructure behind its 3D and video platforms.^[45] (See Reactor.) Alex Yu's personal website, as of October 2026, describes him as the former co-founder and CTO of Luma, where he worked on the Ray (Dream Machine) video model and the Genie text-to-3D model, and says he is now an AI researcher and founder in San Francisco working on image generation.^[43] Neither departure was the subject of a Luma announcement that could be found.

On May 31, 2026, the Chinese technology outlet QbitAI reported that Chief Scientist Jiaming Song, the first author of DDIM, had announced on LinkedIn that he had left Luma that month after nearly three years, without naming a new employer.^[46] Amit Jain remained chief executive,^[48] and Caroline Ingeborn continued to speak for the company as chief operating officer.^[7][57]

Skills, Layers, Scenes, and new third-party models (June-August 2026)

After the Agents launch, Luma shipped a run of features for the agent workspace. On June 3, 2026, it joined Human After All, an AI creator studio launched in Paris by Webedia-Elephant, as a creative partner alongside Google and ElevenLabs.^[51] Luma Skills, released June 16, let users save a multi-step workflow inside Luma Agents (an input, a set of tools, and plain-language instructions) and rerun it on new assets, invoke it by slash command, or share it with a link or as a bundled package.^[50]

On July 29, Luma introduced Layers, a Uni-1 feature that generates images as independently editable layers and decomposes uploaded images into layers, so a product, a background, or a line of copy can be changed without regenerating the rest of the frame.^[53] On August 11 it launched Luma Scenes, which turns a brief or reference images into an editable storyboard of keyframes generated together by Uni-1; users approve each scene before rendering the sequence with Ray3.2 or ByteDance's Seedance 2.^[56]

Luma also widened the set of outside models available inside the workspace. On August 6 it added ByteDance's Seedance 2.5, which Luma said could generate 30-second clips in a single pass, and MiniMax H3, which generates video and matched audio together at native 2K and 24 frames per second in clips of 5 to 15 seconds, initially for self-serve accounts only.^[54][55] On August 13 Luma and the emotion-analytics company Dumbstruck announced a "Creative Intelligence" partnership in which Dumbstruck measures audience responses to ads and Luma Agents carry out the resulting edits; Wayfair was named as a brand exploring the approach.^[57]

AMD inference and the Arabic interface (July-September 2026)

In a July 22, 2026 engineering post, Luma said that thousands of AMD Instinct MI325X GPUs, supplied with compute from partners including TensorWave, served inference for every major Luma model, including Ray3.2, Uni-1, and Luma Agents, while its training platform ran on a separate cluster. Luma reported that roughly 90% of its code ran on AMD's PyTorch backend without changes, that moving the video-to-video pipeline to production took two to three engineers about two weeks, and that human evaluators found no quality regression.^[52]

On September 2, 2026, Luma said that Luma Agents, which already accepted Arabic prompts and produced Arabic outputs, would roll out a fully Arabic-language interface later in the year, its first non-English interface language. The company tied the update to its HUMAIN partnership and to what it called its regional headquarters in Riyadh, and quoted Jason Day, Head of EMEA.^[58]

Variants, Claude Motion, Kling 4.0, and Mini-Apps (October 2026)

On October 1, 2026, Luma launched Variants, which takes one approved static ad and produces versions for five standard placements (from 9:16 Story to 16:9 widescreen) and for multiple languages while keeping the logo, headline, and call to action intact.^[59] On October 8, Luma announced that animations from Anthropic's Claude Motion, then in beta on Claude Team and Enterprise plans, could be opened in Luma through an MCP connector and then restyled, reframed, or expanded into other formats with its Ray and Uni models.^[60] Luma added Kling 4.0, a third-party video model, on October 23, and on October 27 described eight one-click "mini-apps" in Luma Agents: Face Swap, Camera Angles, Light Transfer, Relight Studio, Virtual Try-On, Background Swap, Ad Variants, and Product Shots.^[61][62]

Products and Technology

Dream Machine

Dream Machine is Luma AI's consumer-facing web platform and iOS application for AI video generation. Launched in June 2024, it allows users to generate video from text prompts, images, or existing video clips. The platform supports text-to-video, image-to-video, video-to-video, and video extension workflows. During 2026 Luma reorganized the consumer product around Luma Agents as the Luma App at app.lumalabs.ai; its official company-information page (last updated July 13, 2026) asks that the name "Dream Machine" no longer be used, calling it "an older deprecated model replaced by Ray," and names Ray3.2 and UNI-1.1 as the current video and image models.^[42] See Luma Dream Machine for the product's history.

Luma said Dream Machine had 25 million registered users in November 2024 and more than 30 million users by September 2025; its careers page listed "30M+" users worldwide as of October 2026.^[5][41][67] Luma does not publish revenue. Sacra, a private-markets research firm, estimated that Luma reached $8 million in annualized revenue as of December 2024, shortly after Dream Machine subscriptions launched, and still showed that figure as of October 2026.^[16]

Pricing

Luma's consumer pricing has changed several times. At the November 2024 relaunch, paid tiers cost $9.99, $29.99, and $99.99 per month.^[67] As of October 2026, Luma listed the following plans for the Luma App, with yearly billing saving up to 20%:^[63]

The same pricing page listed third-party models available inside the app alongside Ray3.2, Ray3.14, and Uni-1, including Google's Veo 3.1 and Nano Banana models, OpenAI's GPT Image models, ByteDance's Seedance and Seedream models, Kling 3.0 and 4.0, MiniMax H3, and Flux 3.^[63]

Ray Model Family

The Ray models are Luma AI's core video generation models, powering the Dream Machine platform (now the Luma App) and API.

Ray2 was trained directly on video sequences rather than still images, which Luma credits with its ability to produce natural, continuous motion.^[4] Alex Yu has described the original Dream Machine model as built on a transformer architecture trained directly on videos, generating 120 frames from an image or text prompt in under 120 seconds.^[43]

Genie (Text-to-3D)

Genie is Luma AI's text-to-3D generation tool. It converts natural language prompts into 3D models in under 10 seconds, with a separate refine step for higher quality.^[43] The system follows a two-stage process: first generating four low-resolution previews in approximately 10 seconds, then allowing users to refine their chosen variant into a high-quality model with PBR (Physically Based Rendering) textures and optimized mesh topology.

Genie generates quad meshes and materials at configurable polygon counts, exporting in standard formats compatible with professional 3D pipelines. It is accessible via web, iOS, and Discord. The tool targets developers, 3D artists, game designers, and VR/AR creators who need rapid 3D asset prototyping.

Genie was first released as a research preview in November 2023 and reached general availability with version 1.0 in January 2024.^[1]

Photon (Image Generation)

Photon is Luma AI's text-to-image foundation model, released on November 25, 2024.^[9] The model family includes Photon (full quality at 1.5 cents per 2MP 1080p image) and Photon Flash (speed-optimized at 0.4 cents per image).^[73] Luma claimed Photon was 800% faster and cheaper than comparable models at launch.^[9] By July 2026 Luma said Photon no longer existed as a separate product, with Uni-1.1 as its current image model.^[42]

Photon supports a large context window for visual generative models and can generate consistent characters from a single input reference image. The model is available through the Dream Machine web interface and via API.

Uni-1 and Luma Agents (Unified Intelligence)

Uni-1, released in March 2026, is the first model in Luma's Unified Intelligence family and marked an architectural departure for the company's image generation: instead of stitching a generator to a separate reasoning system at inference time, Uni-1 is a single decoder-only autoregressive transformer that processes interleaved text and image tokens, so instruction understanding, reasoning, and pixel generation run in one forward pass.^[29] The model supports text-to-image generation, natural-language image editing, and reference-based generation with multiple reference inputs. At 2K resolution, Luma priced Uni-1 API generation at approximately $0.09 per text-to-image output; VentureBeat compared this with about $0.101 for Google's Nano Banana 2 and $0.134 for Nano Banana Pro at the same resolution, though Nano Banana 2 remained cheaper at lower resolutions.^[29][30] The Uni-1.1 API, released May 5, 2026, added Build and Scale pricing tiers, output in standard aspect ratios, and generation times of roughly 31 seconds per image.^[31]

Luma Agents, launched March 5, 2026, sit above the model layer: they plan and execute multi-step creative briefs across Dream Machine, Uni-1, the Ray models, and selected third-party models, and can critique and revise their own outputs.^[28] Through 2026 Luma added reusable Skills, Layers, Scenes, Variants, and one-click mini-apps to the agent workspace (see History above), and its company-information page lists Publicis Groupe, Serviceplan, Mazda, Dentsu, and HUMAIN among the teams using Luma Agents.^{[42][50][53][56][59][62]}

3D Capture (NeRF and Gaussian Splatting)

Luma AI's original product was a 3D capture tool built on neural radiance field technology. The iOS app allows users to capture 3D scenes by walking around a subject while recording video on an iPhone (model 11 or newer). The AI processes the video to reconstruct 3D geometry, lighting, and material properties.

Key capabilities of the 3D capture system include:

• Capture of fine details, reflections, and view-dependent lighting effects
• Support for both NeRF and Gaussian Splatting representations
• Export to USDZ, glTF, OBJ, and native NeRF/Gaussian Splat formats
• Editing tools for adjusting lighting, removing backgrounds, and modifying materials
• Compatibility with Blender, Unity, Unreal Engine, and other 3D engines
• No LiDAR or specialized hardware required

The Unreal Engine plugin (released April 2023) allows photorealistic NeRF and Gaussian Splatting captures to be imported, cropped, relit, and combined with traditional Unreal Engine VFX and real-time graphics.

API and Developer Platform

Luma AI provides a developer API for integrating video and image generation into third-party applications, billed separately from consumer subscriptions. As of October 2026 the API offered Ray3.2 for video and Uni-1.1 for images in two tiers: Build (pay per output, no commitment, rate limits and no latency SLA) and Scale (dedicated capacity with guaranteed throughput and latency).^[64] Luma's published Build prices for Ray3.2 were:^[64]

Those prices are for standard dynamic range output; HDR output costs twice as much and HDR with EXR export three times as much.^[64] The Ray3.2 API exposes multi-keyframe control with up to 16 keyframes, video-to-video at up to 20 seconds, and Reframe.^[42][64] The Ray2 model is also available through Amazon Bedrock, allowing AWS customers to access Luma's models through a managed API.^[10]

Leadership and Team

As of October 2026, Luma's careers page listed more than 250 team members, with open roles in Redwood City, London, Munich, Riyadh, New York, and Los Angeles (the Dream Lab studio).^[41] In December 2025, CNBC reported that the roughly 200 planned London hires would make up about 40% of the company's workforce.^[13]

Funding

The publicly reported rounds below add up to roughly $1.06 billion. Luma's careers page, as of October 2026, stated total funding of more than $1.3 billion without itemizing the difference.^[41] No new funding round had been announced as of October 2026; Semafor and CNBC were still citing the November 2025 round and its roughly $4 billion valuation in mid-2026.^[48][49]

Before the HUMAIN-led round closed, Sacra wrote that Luma was seeking to raise at least $1.1 billion at a valuation of $3.2 billion, with Middle Eastern investors in discussions.^[16] Luma's own announcements in 2025 and 2026 list HUMAIN, Andreessen Horowitz, AWS, AMD Ventures, NVIDIA, Amplify Partners, and Matrix Partners among its backers.^[7][57]

Partnerships and Integrations

Amazon Web Services

Luma AI and AWS formed a strategic partnership in late 2024.^[18] AWS became the preferred compute partner for Luma AI, and Luma's Ray2 model was integrated into Amazon Bedrock.^[10] Luma trains its foundation models on Amazon SageMaker HyperPod and plans to bring models to AWS Trainium and Inferentia chips for lower-cost training and inference.^[18]

The partnership was unveiled at AWS re:Invent on December 4, 2024, where Luma previewed Ray2 ahead of its January 2025 release.^[18] In April 2026, AWS also backed Innovative Dreams, the hybrid AI-filmmaking production services company that Luma launched with Wonder Project.^[36]

Adobe

Luma announced in April 2025 that Ray2 was coming to Adobe Firefly and Firefly Boards.^[39] In September 2025 Adobe integrated Ray3 into Firefly and Firefly Boards as Luma's first launch partner for that model, making Adobe the first partner to offer Ray3 outside Luma's own platform.^[8] The integration allows users to generate video within Adobe's creative tools and sync content to Premiere Pro for editing.

HUMAIN and Saudi Arabia

Through the $900 million Series C investment in November 2025, Luma AI became a customer of HUMAIN's Project Halo, a 2-gigawatt AI supercluster in Saudi Arabia.^[2] The partnership includes development of Arabic-native foundation models under the HUMAIN Create initiative and a Riyadh office that Luma announced on February 11, 2026.^[12] By September 2026 Luma referred to its "regional headquarters in Riyadh," and its careers page listed open roles there.^[41][58] The partnership also produced the planned Arabic-language interface for Luma Agents announced in September 2026.^[58]

The relationship predates the investment: Luma and HUMAIN first announced a strategic partnership on May 15, 2025, covering compute capacity, go-to-market collaboration, and investment, with Luma deploying Ray2 on HUMAIN infrastructure to serve creative customers.^[21]

Publicis Groupe

In February 2026, Publicis Groupe Middle East became Luma AI's preferred AI creative partner across the MENA region, integrating Luma's generative video technologies into advertising and brand production workflows.^[12] Publicis Groupe was also among the first enterprise customers of Luma Agents in March 2026.^[28]

AMD and TensorWave

AMD participated in Luma's December 2024 round, and AMD Ventures joined the November 2025 Series C.^[2][16] By July 2026 Luma said it ran production inference for all of its major models on thousands of AMD Instinct MI325X GPUs, with compute from partners including TensorWave.^[52]

Other 2026 partnerships

• Anthropic: Since October 8, 2026, animations made with Anthropic's Claude Motion can be opened in Luma through an MCP connector and turned into finished video in other formats.^[60]
• Dumbstruck: An August 2026 "Creative Intelligence" partnership pairs Dumbstruck's emotion analytics with edits carried out by Luma Agents.^[57]
• Webedia-Elephant: Luma is a creative partner in Human After All, an AI creator studio launched in Paris in June 2026.^[51]
• Wonder Project: Co-founded Innovative Dreams, a hybrid production services company backed by AWS, in April 2026.^[36]

Competitive Landscape

How does Luma AI compare to Runway, Sora, and Veo?

Luma AI operates in the AI video generation market alongside several other companies.

The comparison above reflects the products as commonly positioned rather than independent benchmark results. The field shifted during 2026: Semafor reported that OpenAI had shut down its Sora video generator earlier in the year while keeping the technology for internal robotics work.^[48] Luma itself now offers several competitors' video models, including Veo 3.1, Kling, and Seedance, inside its own app, positioning its agent workspace rather than any single model as the product.^[54][61][63] When announcing its Series C, Jain told CNBC that Ray3 benchmarked higher than OpenAI's Sora 2 and around the same level as Google's Veo 3.^[47]

In image generation, Luma's Uni-1 competes directly with Google's Nano Banana family and OpenAI's GPT Image models, with Luma positioning reasoning quality and lower per-image cost as its differentiators.^[29][30]

Reception and criticism

Dream Machine's June 2024 launch drew copyright criticism. "Monster Camp," an AI-generated trailer among a series of Dream Machine videos Luma posted, featured a recognizable Mike Wazowski from Pixar's Monsters, Inc.; Jain told The Verge that the video "was created from a user uploaded image that contains the visuals" and that the system "didn't generate the characters" on its own, and said Luma used "industry standard programmatic methods" to moderate content. He did not answer The Verge's questions about what content Luma's models were trained on.^[65] In November 2025 Jain told CNBC that Luma had built "very robust systems" using models it trained to detect such content.^[47]

CNBC ranked Luma 35th on its 2026 Disruptor 50 list, citing its Ray3 model, the Innovative Dreams venture with AWS, and the HUMAIN-led round.^[49] In Debarghya Das's "Ultimate Neolab List," a 2026 compilation of new AI labs, Luma appears with a "Borderline" classification, explained as "Video product revenue, but moving into world models and physical AI."^[72]

What is Luma AI used for?

Luma AI's tools are used across several industries:

• Film and television: Pre-visualization, storyboarding, concept art animation, VFX prototyping, and HDR content generation
• Advertising and marketing: Rapid creation of video ads, campaign variations for different regions, product demonstrations
• Gaming: 3D asset generation with Genie, environment prototyping, cinematic trailer creation
• E-commerce: Product photography in 3D, interactive 3D product views on websites
• Architecture and real estate: 3D scanning of physical spaces, virtual tours
• Education: 3D capture for teaching materials, visual explanations of concepts
• Social media: Short-form video content creation, meme generation, creative experimentation

Technical Architecture

Luma AI's video models are built on a multimodal transformer architecture trained directly on video, and Adobe described Ray3 as built on "a new multimodal reasoning system."^[4][8][43] In a July 2026 engineering post, Luma explained that its video and world models are served not as single networks but as pipelines of submodels, including variational autoencoders, ControlNets, LoRAs, and a denoiser backbone, each with its own compute and memory profile; it said large per-GPU memory let it keep every submodel of a pipeline resident at once. Its multimodal Uni-1 model, the first such model Luma had served to customers, required KV caching and a disaggregated serving setup.^[52]

Key technical characteristics of the Ray model family include:

• Direct video training: Ray models are trained on video sequences rather than individual frames, giving them an understanding of motion as continuous flow
• Multi-modal inputs: The architecture processes text instructions alongside visual inputs (images, video frames) within a unified model
• Reasoning engine (Ray3+): Ray3 introduced a reasoning layer that enables the model to generate physically accurate interactions, logical event sequences, and coherent cause-and-effect chains ^[5]
• HDR pipeline: Ray3 was the first video generation model to support native HDR output in ACES2065-1 EXR format at up to 16-bit precision ^[5]
• Resolution scaling: Models generate at base resolutions (540p, 720p, 1080p) with optional upscaling to 4K (8.3 million pixels per frame)

For 3D capture, Luma AI uses both NeRF (neural radiance field) representations and Gaussian Splatting. NeRF represents scenes as continuous volumetric functions, while Gaussian Splatting uses explicit 3D Gaussian primitives for faster rendering. Both methods can be exported for use in game engines and 3D software.

Luma frames its long-term agenda as building multimodal world models rather than larger language models. Jain has described the goal as a "universal imagination engine" trained on multimodal data approaching 1.2 quintillion tokens across tens of petabytes, arguing that video, audio, and image data carry far more information about the physical world than text.^[15] The company says its training systems are designed for peta-scale multimodal data, 1,000 to 10,000 times more information than frontier LLMs process.^[2] With Uni-1, Luma also departed from diffusion-based image generation in favor of a decoder-only autoregressive transformer that reasons over interleaved text and image tokens before rendering pixels.^[29]

Research

Luma's research team, led until May 2026 by Chief Scientist Jiaming Song, publishes work on efficient generative modeling.^[46] In March 2025, Linqi Zhou, Stefano Ermon, and Song released Inductive Moment Matching (IMM), which proposes a class of generative models trained in a single stage, without pre-trained initialization, that sample in one or a few steps. Unlike consistency models, IMM guarantees distribution-level convergence; the method reached 1.99 FID on ImageNet-256x256 using 8 inference steps and a state-of-the-art two-step FID of 1.98 on CIFAR-10.^[20]

In November 2025, Zhou, Mathias Parger, Ayaan Haque, and Song released Terminal Velocity Matching (TVM), a generalization of flow matching that models transitions between any two diffusion timesteps and regularizes behavior at the terminal time, providing an upper bound on the 2-Wasserstein distance between model and data distributions. TVM achieved 3.29 FID on ImageNet-256x256 with a single network evaluation and 1.99 FID with four evaluations, matching diffusion-quality output at a fraction of the inference cost; the paper was published as a conference paper at ICLR 2026 and the ImageNet code was open-sourced.^[27]

The Open Physical AI Lab, announced June 1, 2026, extends this research agenda from generative media toward robotics, applying Luma's world models and evaluation frameworks to the generalization problem in physical AI in collaboration with outside researchers, academia, and hardware partners.^[37]

See also

• Luma Dream Machine
• HUMAIN
• Reactor
• AI video generation
• Runway (company)
• List of neolabs
• Neolab

References

1. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10"Investing in Luma AI." Andreessen Horowitz, January 9, 2024. a16z.com/...investing-in-luma-ai
2. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12"Luma AI Raises $900 Million Series C Led by HUMAIN And Partners on 2 Gigawatt AI Supercluster in Saudi Arabia." Luma AI Press Release / BusinessWire, November 19, 2025. businesswire.com/...en
3. ^1 ^2 ^3 ^4 ^5 ^6 ^7"Luma raises $43M to build AI that crafts 3D models." TechCrunch, January 9, 2024. techcrunch.com/...o-build-ai-that-crafts-3d-models
4. ^1 ^2 ^3 ^4"Ray2 Release: Next-Generation AI Video Model." Luma AI Changelog, January 15, 2025. lumalabs.ai/...introducing-ray2
5. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10"Luma AI launches Ray3." Luma AI Press Release, September 18, 2025. lumalabs.ai/...ray3
6. ^1 ^2"Luma AI Launches Ray3.14, Eliminating Quality-Speed-Cost Tradeoff in Generative Video." BusinessWire, January 26, 2026. businesswire.com/...en
7. ^1 ^2 ^3 ^4 ^5 ^6 ^7"Luma AI Opens for International Business With Former WPP Executive Jason Day Leading New London Office." Luma AI Press Release, December 2, 2025. lumalabs.ai/...jason-day-leading-new-london-office
8. ^1 ^2 ^3 ^4"Unlock New Creative Possibilities with Luma AI's Ray3 Video Model - Now in Adobe Firefly." Adobe Blog, September 18, 2025. blog.adobe.com/...y3-video-model-now-adobe-firefly
9. ^1 ^2 ^3 ^4 ^5"Luma AI updates Dream Machine with new Photon image generation model and more." SiliconANGLE, November 25, 2024. siliconangle.com/...efreshed-ui-subscription-tiers
10. ^1 ^2 ^3"Luma AI's Ray2 video model is now available in Amazon Bedrock." Amazon Web Services Blog, 2025. aws.amazon.com/...-now-available-in-amazon-bedrock
11. ^1 ^2 ^3"Introducing Ray3 Modify." Luma AI, December 18, 2025. lumalabs.ai/...ray3-modify
12. ^1 ^2 ^3 ^4"Luma AI to Open Riyadh Office to Accelerate HUMAIN Create." BusinessWire, February 11, 2026. businesswire.com/...en
13. ^1 ^2 ^3"Nvidia-backed startup Luma AI announces major London expansion." CNBC, December 2, 2025. cnbc.com/...ma-ai-announces-major-london-expansion
14. ^1 ^2 ^3 ^4"Jiaming Song." MIT Technology Review Innovators Under 35. innovatorsunder35.com/...jiaming-song
15. ^1 ^2 ^3"Building AI That Understands Our World Like We Do, with Amit Jain." Amplify Partners. amplifypartners.com/...stands-our-world-like-we-do
16. ^1 ^2 ^3 ^4 ^5 ^6"Luma AI revenue, funding & news." Sacra. sacra.com/...luma-ai
17. ^1 ^2 ^3"Luma AI, Inc. announced that it has received $90 million in funding from a group of investors." MarketScreener, December 6, 2024. marketscreener.com/...-group-of-investors-48550518
18. ^1 ^2 ^3 ^4"Luma AI Announces New Video Model Ray 2 Will Soon be Available to Consumers, Professionals, and Developers." Amazon Press Center, December 4, 2024. press.aboutamazon.com/...fessionals-and-developers
19. ^1 ^2"Introducing Ray2 Flash - 3x faster, 3x cheaper new model." Luma AI on X, March 2025. x.com/...1898056614684381218
20. ^1 ^2Zhou, Linqi; Ermon, Stefano; Song, Jiaming. "Inductive Moment Matching." arXiv:2503.07565, March 10, 2025. arxiv.org/...2503.07565
21. ^1 ^2"Luma is Partnering with HUMAIN AI to Accelerate the Arrival of Multimodal AGI." Luma AI, May 15, 2025. lumalabs.ai/...luma-joins-forces-with-humain
22. ^"Introducing Modify Video - AI Video-to-Video Editing." Luma AI Changelog, June 2025. lumalabs.ai/...introducing-modify-video
23. ^"Luma AI Launches 'Dream Lab' Studio In LA, Taps Veteran Execs Verena Puhm And Jon Finger For Key Posts." Deadline, July 10, 2025. deadline.com/...-verena-puhm-jon-finger-1236453424
24. ^"Introducing Modify with Instructions." Luma AI, August 12, 2025. lumalabs.ai/...modify-with-instructions
25. ^"Ray3 Evaluation Report - State-of-the-Art Performance for Pro Video Generation." Luma AI, October 14, 2025. lumalabs.ai/...ray3-eval
26. ^"Luma releases a new AI model that lets users generate a video from a start and end frame." TechCrunch, December 18, 2025. techcrunch.com/...video-from-a-start-and-end-frame
27. ^Zhou, Linqi; Parger, Mathias; Haque, Ayaan; Song, Jiaming. "Terminal Velocity Matching." arXiv:2511.19797, November 2025. arxiv.org/...2511.19797
28. ^1 ^2 ^3 ^4"Luma launches creative AI agents powered by its new 'Unified Intelligence' models." TechCrunch, March 5, 2026. techcrunch.com/...-new-unified-intelligence-models
29. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8"Luma AI's Uni-1 could be the first real challenger to Google's Nano Banana image dominance." The Decoder, March 2026. the-decoder.com/...les-nano-banana-image-dominance
30. ^1 ^2 ^3"Luma AI launches Uni-1, a model that outscores Google and OpenAI while costing up to 30 percent less." VentureBeat, March 2026. venturebeat.com/...tscores-google-and-openai-while
31. ^1 ^2"Introducing the Uni-1.1 API: Intelligence You Can Direct. Aesthetic You Can Ship." Luma AI, May 5, 2026. lumalabs.ai/...uni-1-1-api
32. ^"Luma AI Offers $1 Million for Cannes Lions Gold Winner." BusinessWire, February 2, 2026. businesswire.com/...n-for-Cannes-Lions-Gold-Winner
33. ^1 ^2"Serviceplan Group Deploys Creative AI Across Global Operations in Partnership With Luma AI." BusinessWire, February 19, 2026. businesswire.com/...ns-in-Partnership-With-Luma-AI
34. ^"Luma Submits 21 AI-Generated Finalists to Cannes Lions After Inspiring 400 Ads in Less Than 8 Weeks." Luma AI, April 9, 2026. lumalabs.ai/...luma-dreambrief-submission-finalists
35. ^"South African Agency Boundless Delivers Mazda's First AI-produced Commercial Using Luma Agents." BusinessWire, April 15, 2026. businesswire.com/...d-Commercial-Using-Luma-Agents
36. ^1 ^2 ^3"Wonder Project And AI Firm Luma Launch Innovative Dreams As Home For 'Hybrid Filmmaking'." Deadline, April 16, 2026. deadline.com/...reams-hybrid-filmmaking-1236861786
37. ^1 ^2 ^3"Luma Announces The Open Physical AI Lab: An Open Science Effort to Solve Generalization in Physical AI." Luma AI, June 1, 2026. lumalabs.ai/...luma-open-physical-ai-lab
38. ^1 ^2 ^3"Luma Introduces Ray3.2 Model & API: Complete Creative Control for Video Generation." Luma AI, June 9, 2026. lumalabs.ai/...introducing-ray-3-2
39. ^"Ray2 is coming to Adobe Firefly & Firefly Boards." Luma AI, April 24, 2025. lumalabs.ai/...ray2-is-coming-to-adobe-firefly
40. ^1 ^2"AI video shift powers Luma's $4 billion valuation." EMARKETER, November 19, 2025. emarketer.com/...owers-luma-s--4-billion-valuation
41. ^1 ^2 ^3 ^4 ^5 ^6"Join us" (careers page). Luma AI, accessed October 27, 2026. lumalabs.ai/join
42. ^1 ^2 ^3 ^4 ^5"Luma AI: LLM Info" (official company information page, last updated July 13, 2026). Luma AI. lumalabs.ai/llm-info
43. ^1 ^2 ^3 ^4 ^5 ^6 ^7Yu, Alex. "Alex Yu's Website." Accessed October 27, 2026. alexyu.net
44. ^1 ^2 ^3Taiuti, Alberto. "Alberto Taiuti's website." Accessed October 27, 2026. albertotaiuti.com
45. ^1 ^2 ^3 ^4"Reactor Emerges from Stealth with $59M to Build the Platform for Real-Time AI Worlds." Reactor press release via Cision/Newswire, May 28, 2026. newswire.ca/...m-for-real-time-ai-worlds-813500054
46. ^1 ^2 ^3 ^4 ^5"DDIM creator Jiaming Song announces departure" (in Chinese). QbitAI, May 31, 2026. qbitai.com/...427104
47. ^1 ^2 ^3 ^4 ^5"Luma AI raises $900 million in funding round led by Saudi AI firm Humain." CNBC, November 19, 2025. cnbc.com/...in-funding-led-by-saudi-ai-firm-humain
48. ^1 ^2 ^3 ^4 ^5 ^6"AI video startup Luma AI makes the jump into robotics with 'open' lab." Semafor, June 1, 2026. semafor.com/...he-jump-into-robotics-with-open-lab
49. ^1 ^2"35. Luma AI" (2026 CNBC Disruptor 50). CNBC, May 19, 2026. cnbc.com/...luma-ai-cnbc-disruptor-50-ranking
50. ^1 ^2 ^3"Introducing Luma Skills: Build a Creative Workflow Once, Run It Forever." Luma AI, June 16, 2026. lumalabs.ai/...luma-skills
51. ^1 ^2"Human, after all." Luma AI, June 3, 2026. lumalabs.ai/...human-after-all
52. ^1 ^2 ^3"Luma Runs Production Inference on AMD and Tensorwave." Luma AI, July 22, 2026. lumalabs.ai/...ion-inference-on-amd-and-tensorwave
53. ^1 ^2 ^3"Introducing Layers." Luma AI, July 29, 2026. lumalabs.ai/...introducing-layers
54. ^1 ^2 ^3"Seedance 2.5: Now Live in Luma." Luma AI, August 6, 2026. lumalabs.ai/...seedance-2.5-live-in-luma
55. ^1 ^2"MiniMax H3 Available Now." Luma AI, August 6, 2026. lumalabs.ai/...max-h3-now-available-in-luma-agents
56. ^1 ^2 ^3"Introducing Luma Scenes." Luma AI, August 11, 2026. lumalabs.ai/...introducing-luma-scenes
57. ^1 ^2 ^3 ^4 ^5"Luma and Dumbstruck Launch Creative Intelligence for Advertising." Luma AI, August 13, 2026. lumalabs.ai/...eative-intelligence-for-advertising
58. ^1 ^2 ^3"Luma AI to Launch Native Arabic-Language Interface for Saudi and Arabic-Speaking Creative Community." Luma AI, September 2, 2026. lumalabs.ai/...-arabic-speaking-creative-community
59. ^1 ^2"Live Now: Luma Variants." Luma AI, October 1, 2026. lumalabs.ai/...introducing-luma-variants
60. ^1 ^2"Start in Claude Motion. Finish it in Luma." Luma AI, October 8, 2026. lumalabs.ai/...Luma-Claude-Motion-Launch-Partnership
61. ^1 ^2 ^3"Kling 4.0 in Luma: Fast Action and Big Camera Moves." Luma AI, October 23, 2026. lumalabs.ai/...kling-4-in-luma-ai-video
62. ^1 ^2 ^3"Luma Mini-Apps: One-Click Tools Inside Luma Agents." Luma AI, October 27, 2026. lumalabs.ai/...luma-mini-apps
63. ^1 ^2 ^3"Plans & Pricing." Luma AI, accessed October 27, 2026. lumalabs.ai/pricing
64. ^1 ^2 ^3 ^4"Build with Luma APIs." Luma AI, accessed October 27, 2026. lumalabs.ai/api
65. ^1 ^2Pulliam-Moore, Charles. "An AI video tool just launched, and it's already copying Disney's IP." The Verge, June 18, 2024. theverge.com/...ai-monster-camp-monsters-inc-pixar
66. ^"This AI-Generated Pixar Style Animation Just Might Blow Your Mind." Futurism, June 15, 2024. futurism.com/...ai-generated-pixar-style-animation
67. ^1 ^2 ^3 ^4 ^5"People Can Show the World What They See With Launch of New Luma AI Dream Machine Service." Luma AI press release (Business Wire) via Silicon UK, November 25, 2024. silicon.co.uk/...new-luma-ai-dream-machine-service
68. ^1 ^2"Amit Jain" (speaker profile). TechCrunch, accessed October 27, 2026. techcrunch.com/...amit-jain
69. ^1 ^2 ^3"Luma AI Announces $20M Series A Round." Radiance Fields, March 20, 2023. radiancefields.com/...announces-20m-series-a-round
70. ^1 ^2 ^3"Luma AI Hires Matthew Tancik." Radiance Fields, June 6, 2023. radiancefields.com/luma-ai-hires-matthew-tancik
71. ^1 ^2"Nerfstudio Professor Angjoo Kanazawa joins Luma AI as Chief Scientific Advisor." Radiance Fields, April 17, 2023. radiancefields.com/...-as-chief-scientific-advisor
72. ^Das, Debarghya. "The Ultimate Neolab List." debarghyadas.com, accessed October 27, 2026. debarghyadas.com/neolabs
73. ^1 ^2"Luma Photon." Luma AI, accessed October 27, 2026. lumalabs.ai/photon
74. ^1 ^2"Luma AI - $4M Seed." Startup Intros, accessed October 27, 2026. startupintros.com/...2021-10-01-luma-ai-seed