Runway

Market position

Runway is described as an established player in generative video, best known for physics-aware video generation models, with Gen-4.5 reported to beat Google and OpenAI video models on several benchmarks. It competes in the emerging world-models field against World Labs and Google DeepMind, both of which have released publicly available models. Wikipedia notes Runway's use in notable film and television productions, and Time selected it as one of the 100 Most Influential Companies in June 2023. [4][6]

Runway positions a single underlying "Real-World Intelligence" model stack across three product surfaces (creative, developer, robotics), aggregates both its own and third-party models in one consumer app, and emphasizes world models that simulate physics and dynamics rather than language-only systems. Its Lionsgate model is trained exclusively on that studio's catalog and restricted to Lionsgate-affiliated filmmakers. [0][1][3][4][6]

Technology

Runway develops foundation models it calls General World Models, intended to simulate environments and support planning and action. GWM-1 is its state-of-the-art general world model, used for photorealistic simulation, policy inference and synthetic data generation in robotics. Gen-4.5 is its latest video model, supporting text-to-video high-definition generation with native audio, long-form multi-shot generation, character consistency and advanced editing tools; TechCrunch reports it outperformed Google and OpenAI video offerings on several benchmarks. Other capabilities include image generation, image-to-video and video-to-video, consistent characters/objects/locations across shots, and an agent that plans, generates and assembles complete videos from natural-language instructions. The company earlier co-developed the Stable Diffusion latent diffusion text-to-image model. Compute capacity is expanded through a deal with CoreWeave. [0][1][3][4][6]

Go-to-market

Self-serve free trial and subscriptions across web and an iOS app, a developer API with self-service key issuance, and direct enterprise sales for large organizations and robotics customers. Runway builds industry credibility through partnerships with media companies (Lionsgate, AMC Networks, Adobe), community programs (Discord, Runway Academy, Runway for Educators) and its annual AI Film Festival plus a partnership with the Tribeca Film Festival. [0][1][2][3][4][6]

Individual creators and creative teams, filmmakers, designers, marketers and advertisers; media and entertainment companies and studios; developers building on generative media APIs; and, more recently, gaming and robotics organizations. [0][1][3][4]

Geography

Headquartered in Manhattan, New York City, serving customers worldwide. Runway announced London as its European headquarters and world model research hub, with a planned $100 million investment in the UK AI ecosystem and a new Head of Europe. One secondary source also lists offices in San Francisco, Seattle, London and Tel Aviv plus remote staff. [0][1][5][6]

History

After its 2018 founding, Runway raised a $2 million seed round in 2018 to build a platform for deploying machine learning models in multimedia applications, an $8.5 million Series A in December 2020, a $35 million Series B in December 2021, a $50 million Series C in December 2022, and a $141 million Series C extension in June 2023 at a $1.5 billion valuation from investors including Google, Nvidia and Salesforce. In August 2022 it co-released an improved version of its latent diffusion model, Stable Diffusion, with the CompVis Group at LMU Munich, supported by compute from Stability AI. It announced a Lionsgate partnership in September 2024 and an AMC Networks partnership in June 2025, raised $308 million led by General Atlantic in April 2025 at a valuation over $3 billion, released its first world model in December 2025, and raised a $315 million Series E in February 2026 at a $5.3 billion valuation. Valenzuela has said the company raised capital roughly every twelve months while maintaining twelve to twenty-four months of runway. [4][6][7]

Risks & controversies

Runway operates in a compute-intensive market with intensifying competition in world models from World Labs and Google DeepMind. A Forbes article referenced by Wikipedia questions credit-taking around Stable Diffusion's success. User reviews on the App Store include complaints that the Agent fails to follow editing instructions, cannot reliably reference uploaded assets or make incremental edits, and that failed generations consume paid credits. Public sources also conflict on basic facts: one aggregator profile attributes the company to a different founder and 2014 founding date, and headcount figures range from 86 to roughly 140. [3][4][5][6]