Cristóbal Valenzuela
Last reviewed
Sources
14 citations
Review status
Source-backed
Revision
v1 · 1,844 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Sources
14 citations
Review status
Source-backed
Revision
v1 · 1,844 words
Add missing citations, update stale details, or suggest a clearer explanation.
Cristobal Valenzuela (born 1989) is a Chilean technologist, artist, and entrepreneur who co-founded Runway, the New York based generative AI company, in 2018 and has served as its chief executive officer since its founding. Trained in economics and in media art rather than in computer science, Valenzuela built Runway around the idea of putting machine learning tools directly into the hands of artists and filmmakers. Under his leadership the company became a co-creator of the original Stable Diffusion image model in 2022 and a pioneer of commercial text-to-video generation through its Gen series of models. He was named to the inaugural TIME 100 Most Influential People in AI list in September 2023.[1][2][3]
Valenzuela was born in 1989 and raised in Chile. He studied at Adolfo Ibanez University in Santiago, where he earned a bachelor's degree in economics and business management between 2008 and 2012 and later a master's degree in the arts and design field. He worked as a teaching and research assistant at the university's School of Design and taught there as an adjunct, developing the interest in the intersection of computation and creative practice that would define his career.[1]
In 2016 he moved to New York City to attend the Interactive Telecommunications Program (ITP) at New York University's Tisch School of the Arts, a graduate program known for blending art, design, and emerging technology. He completed his media arts degree there in 2018 and stayed on as a researcher, working alongside the educator Daniel Shiffman. During this period Valenzuela contributed to open-source creative-coding projects, including the ml5.js library that makes machine learning accessible inside the browser, and his early experiments were supported by Google and the Processing Foundation. His artwork was exhibited in venues across Latin America and the United States, including the Santiago Museum of Contemporary Art.[1]
Runway grew directly out of Valenzuela's graduate work at ITP. His thesis explored how to make machine learning models usable by people without engineering backgrounds, packaging them behind a simple interface so artists and designers could run them without writing code. In 2018 he turned that idea into a company, co-founding Runway in New York City with two collaborators he met at NYU: Alejandro Matamala, a fellow Chilean and designer, and Anastasis Germanidis, a Greek-born technologist who became the company's chief technology officer. Valenzuela has been chief executive since the start.[1][2]
The founders framed the company's mission as building "new tools for human imagination," and the early product reflected that. Runway's first platform let creators access a catalog of machine learning models for tasks such as image segmentation, style transfer, and motion tracking through a desktop application, and it became popular with visual artists, video editors, and educators. Runway raised a 2 million dollar seed round in 2018 to develop this platform for deploying machine learning models, followed by an 8.5 million dollar Series A in December 2020 and a 35 million dollar Series B in December 2021.[2] As the underlying research in generative models advanced, Valenzuela steered the company from a toolkit for running other people's models toward training and shipping its own foundation models for media.[3]
Runway was one of the original creators of Stable Diffusion, the open-source text-to-image model whose 2022 release was a turning point for generative AI. The model was built on a research method called latent diffusion, introduced in the paper "High-Resolution Image Synthesis with Latent Diffusion Models" and presented at the CVPR 2022 conference. That paper was co-authored by the CompVis (Computer Vision and Learning) group at Ludwig Maximilian University of Munich, including Robin Rombach, Andreas Blattmann, Dominik Lorenz, and Bjorn Ommer, together with Patrick Esser, a researcher at Runway. Latent diffusion ran the generative process in a compressed latent space rather than directly on pixels, which cut the computational cost enough to make high-quality image generation practical on consumer hardware.[4][5]
The original Stable Diffusion model was released in August 2022, with Stability AI providing funding and compute alongside the CompVis researchers and Runway. In October 2022 Runway independently published the weights for an improved checkpoint, Stable Diffusion version 1.5, on the Hugging Face platform. The release briefly became contentious: Stability AI, which had been holding back its own 1.5 release over stated legal concerns, filed a takedown request alleging an intellectual property leak. Runway noted that Esser was a co-author of the underlying latent-diffusion research and a Runway employee with legitimate rights to publish derived weights, and Stability withdrew the request. Version 1.5 went on to become the canonical, most widely used Stable Diffusion checkpoint in the open-source community. For Valenzuela's company, the episode established Runway as a serious research lab rather than only an applications vendor.[5][6]
After Stable Diffusion, Valenzuela concentrated Runway on generative video, an area where the company became an early commercial leader. Its Gen family of models advanced quickly from short, stylized clips to longer, more controllable, and more physically consistent footage. The progression is summarized below.[2][7][8]
| Model | Released | Capability |
|---|---|---|
| Gen-1 | February 2023 | Video-to-video: restyles an existing clip using a text prompt or reference image |
| Gen-2 | March 2023 (beta), June 2023 (general) | Adds text-to-video and image-to-video generation from scratch |
| Gen-3 Alpha | June 2024 | Higher fidelity, longer clips, and finer control, trained on new infrastructure |
| Gen-4 | March 2025 | Consistent characters, objects, and scenes across shots using reference images |
| Gen-4.5 | December 2025 | Improved realism with native synchronized audio; basis for the GWM-1 world model |
Gen-1 and Gen-2 were among the first generally available text-to-video systems, and Gen-3 Alpha, released in June 2024, sharply improved motion and visual quality. Gen-4, announced on March 31, 2025, addressed one of the format's hardest problems by letting creators keep a character or setting consistent from one generated shot to the next, which made the tool far more useful for narrative filmmaking. Valenzuela has consistently positioned these systems not as replacements for filmmakers but as new instruments for them, arguing that cheaper generation could let studios make many films for the cost of a single large production.[7][8][9]
To connect its models to professional filmmaking, Runway launched Runway Studios, its production and entertainment arm, and began running an annual AI Film Festival (AIFF). The festival started in 2023 with roughly 300 submissions and screenings in small New York theaters, partnered with the Tribeca Festival beginning in 2024, and grew to more than 6,000 submissions by 2025. In 2024 Runway and Runway Studios introduced the Hundred Film Fund, a program offering grants from 5,000 dollars up to 1 million dollars, plus Runway credits, to filmmakers using AI, with the company taking no ownership of the resulting work. The fund was initially set at 5 million dollars with room to grow to 10 million.[2][10]
The company also signed direct deals with the entertainment industry. In September 2024 Runway announced a first-of-its-kind partnership with the studio Lionsgate, training a custom video model on the studio's proprietary catalog of film and television titles for use by Lionsgate-affiliated filmmakers. In June 2025 Runway agreed to a partnership with AMC Networks covering marketing imagery and pre-visualization. These deals made Valenzuela a prominent voice in the contentious debate over AI's role in Hollywood.[2][9]
Runway became one of the most highly valued companies in generative media. After its early venture rounds, it raised 50 million dollars in a Series C on December 21, 2022, at a 500 million dollar valuation, then extended that round in June 2023 with an additional 141 million dollars at a 1.5 billion dollar valuation, led by Google, Nvidia, and Salesforce. On April 3, 2025, the company closed a 308 million dollar Series D led by General Atlantic, with participation from Fidelity, Baillie Gifford, Nvidia, and SoftBank, valuing Runway at more than 3 billion dollars. On February 10, 2026, Runway raised a 315 million dollar Series E, again led by General Atlantic and joined by Nvidia, Adobe Ventures, AMD Ventures, Fidelity, and Felicis Ventures, at a 5.3 billion dollar valuation. By 2026 the company had raised on the order of 860 million dollars in total.[2][3][11][12]
As of 2026 Valenzuela remains co-founder and chief executive of Runway, and he has reframed the company's long-term ambition around general world models: AI systems that simulate environments, actions, and consequences rather than merely generating video clips. He has described AI-generated video as a "prequel" to this larger goal, with applications spanning entertainment, gaming, robotics, and synthetic training data. In December 2025 Runway revealed GWM-1, its first general world model family, built on top of the Gen-4.5 video model. Unlike the company's earlier systems, GWM-1 is autoregressive, predicting the world frame by frame and allowing a user to intervene in real time, and it shipped in three variants: GWM Worlds for explorable simulated environments, GWM Avatars for conversational characters, and GWM Robotics for robot policy training.[13][14]
Valenzuela's path, from an economics student and media artist in Chile to the chief executive of a multibillion-dollar AI lab, is frequently cited as an example of how the generative AI wave drew leaders from outside traditional computer science. He continues to describe himself foremost as an artist, and Runway under his direction has remained focused on tools for creative expression even as its underlying research has expanded toward simulating the physical world.[1][3]