Runway AI, Inc., also known as Runway and RunwayML, is an American company headquartered in New York City that specializes in generative artificial intelligence research and technologies. The company develops commercial text-to-video and video generative AI models, along with tools for generating images and other multimedia content. Runway's models have been used in films such as Everything Everywhere All at Once and in editing television shows like The Late Show with Stephen Colbert.
Founded in 2018 by Cristóbal Valenzuela, Alejandro Matamala, and Anastasis Germanidis, Runway focuses on generative AI for video, media, and art. It builds proprietary foundational model technology used in filmmaking, post-production, advertising, editing, and visual effects. In 2024, the company released Gen-3 Alpha, a video generation model, and continued to develop enterprise-oriented products.
Founding and funding
Runway was founded in 2018 by the Chileans Cristóbal Valenzuela and Alejandro Matamala and the Greek Anastasis Germanidis, who met at New York University Tisch School of the Arts ITP program. The company initially raised US$2 million in seed funding in 2018 to develop a platform for deploying Machine learning models within multimedia applications.
Runway raised US$8.5 million in a Series A round in December 2020, US$35 million in a Series B round in December 2021, and US$50 million in a Series C round on December 21, 2022. In June 2023, the company closed a US$141 million Series C extension at a US$1.5 billion valuation, with participation from Google, Nvidia, and Salesforce, to build foundational multimodal AI models for content generation. On April 3, 2025, Runway raised US$308 million in a round led by General Atlantic, valuing the company at over US$3 billion.
Technology and products
Runway's core technology centers on generative models for video and image synthesis. Its proprietary models are designed for tasks such as text-to-video generation, video editing, and visual effects. The company's tools are used in filmmaking, post-production, advertising, and editing. In 2024, Runway released Gen-3 Alpha, a video generation model aimed at producing more coherent and controllable output.
The underlying systems are built with Deep learning methods. Like many modern generative models, Runway's video and image systems use neural networks trained on large collections of visual data. These models often rely on transformer architectures, which process relationships across spatial and temporal dimensions using Multi-Head Attention mechanisms. In text-to-video systems, Cross-Attention layers can be used to align text prompts with visual features. Some video generation pipelines also adopt an Encoder-Decoder Architecture structure to map between text, image, and video representations.
Training such models involves a range of established techniques. Researchers typically select Loss Functions suited to the target output, use optimizers such as Adam, and apply learning rate schedules to stabilize convergence. Regularization methods such as Dropout and Batch Normalization are common in deep networks. Gradient clipping is often used to prevent exploding gradients during training, while Data Augmentation helps models generalize from limited examples. For deployment, techniques such as Model Pruning can reduce model size.
At inference time, generative models can be controlled through sampling strategies. For text and image generation, Temperature Scaling adjusts the randomness of outputs, while Top-P (Nucleus) Sampling limits the candidate set during sampling. These controls are also relevant to video generation models that produce frames from text prompts.
Enterprise tools and education
Runway has developed enterprise-oriented products for media organizations, including custom model training and workflow tools. The company's approach has been to offer models that can be adapted to a studio's specific needs rather than only providing a general-purpose generator. Runway also maintains an education program called Runway for Educators, which provides resources for instructors and students working with generative media. Its consumer-facing iOS app targets individual users who want to create short videos and images on mobile devices. The combination of enterprise, education, and consumer offerings distinguishes Runway from research-only AI organizations.
Stable Diffusion
In August 2022, Runway co-released an improved version of its latent diffusion model called Stable Diffusion. Stable Diffusion is an open-source deep learning text-to-image model developed with the CompVis Group at LMU Munich and supported by compute from Stability AI. The model uses a U-Net backbone with residual connections to denoise latent representations, and it conditions on text through cross-attention layers. Stable Diffusion became widely used by artists and developers and helped popularize open-source generative image tools.
Partnerships
In September 2024, Runway announced a partnership with Lionsgate Entertainment to create a custom AI video generation model trained on the studio's proprietary catalog of over 20,000 film and television titles. The model is designed exclusively for use by Lionsgate-affiliated filmmakers and cannot be accessed by other Runway users. The partnership was part of Runway's broader push into enterprise tools for media companies.
In June 2025, AMC Networks partnered with Runway for AI content creation. AMC planned to use Runway's technology to generate marketing images and help pre-visualize shows before production begins. These collaborations reflect Runway's focus on serving professional filmmaking and television production workflows.
Film festivals
Since 2023, Runway has hosted an annual AI Film Festival (AIFF) for films created using artificial intelligence tools. In its initial year, the festival received 300 submissions and held screenings in small New York City theaters. By 2025, the festival had received over 6,000 submissions, reflecting the rapid growth of interest in AI-assisted filmmaking. Runway has also partnered with the Tribeca Film Festival since 2024 to showcase AI-generated films. The festival program is part of Runway's effort to support artists working with generative media and to build a community around AI filmmaking.
Recognition
In June 2023, Time magazine selected Runway as one of the 100 Most Influential Companies in the world. The recognition came shortly after the company's Series C extension and reflected the growing visibility of generative AI in creative industries. Runway's inclusion highlighted the role of video generation tools in the broader generative AI boom.