Runway

Runway is a generative AI company founded in 2018 by three artists and technologists (Cristóbal Valenzuela, Alejandro Matamala, and Anastasis Germanidis) who met at NYU's Tisch School of the Arts ITP program. The through-line across its stories is a single conviction: that AI is a new tool for human creativity, not a replacement for it. Runway built its models, its product, and even its festival around filmmakers and designers, and in doing so helped invent the category of generative video and staked a defensible position inside it.

Category Creation: inventing generative video and crowning the artists who define it

The problem. When Runway began building generative models, "AI video" was not a market: there was no accepted way to make moving images from text or a prompt, no body of work, and no community of practitioners who called themselves AI filmmakers. A tool without a recognized use, and without users who understand why they need it, has no category to sell into.

The approach. Runway did two things at once: it shipped the models that made the category real, and it built the institution that legitimized the people using them. On the product side it released Gen-1 in February 2023, aiming to "do for video what Stable Diffusion did for images," followed by Gen-2, announced March 20, 2023, one of the first accessible commercial text-to-video systems. On the cultural side it launched the AI Film Festival in 2023 to celebrate and elevate a new class of AI-native directors.

How it solved it. The festival gave the emerging category a canon and a stage. Its inaugural 2023 edition drew about 300 submissions, from which jurors named ten finalists; within a few years submissions grew into the thousands with sold-out screenings at Lincoln Center, and by 2026 Runway widened it beyond film into design, fashion, and gaming. By manufacturing both the tool and the community of people who define its use, Runway made itself the reference point for a category it helped bring into existence.

Differentiation: a creative tool, not a research demo

The problem. Runway competes for generative video against far larger, better-funded labs (Google, OpenAI, and others) that can out-spend it on raw model scale. A pure capability race against those balance sheets is one a startup loses. Runway needed a basis of difference that did not reduce to "who has the biggest model."

The approach. Runway positioned itself as a tool built by and for creative professionals rather than a lab publishing benchmarks, embedding its models inside an actual production workflow (editing, rotoscoping, green-screen and generative features) aimed at filmmakers and designers. This artist-first framing was baked in from the founding team's Tisch background and carried through into partnerships with working productions.

How it solved it. The proof was real professional adoption: Runway's tools were used in the making of Everything Everywhere All at Once, which won the 2023 Academy Award for Best Film Editing, and in editing The Late Show with Stephen Colbert. Being the AI video company that credentialed filmmakers actually use, rather than the one with the flashiest demo, gave Runway a defensible identity that scale alone could not copy, and it kept the company competitive even as giants entered the space.

PMF: from a menu of ML tools to the thing people came for

The problem. Runway did not start as a video-generation company. Its early product was a broad platform of machine-learning "magic tools" for creators, a menu of capabilities spanning image and video editing. Breadth like that risks a product that many people try and few people need, with no single feature strong enough to build a business on.

The approach. Runway followed the pull. As its research pushed from image synthesis into motion, it concentrated the product around generative video, the capability where creator demand was clearly strongest, and shipped it in the forms professionals could actually reach: a browser app and an iOS release for Gen-2 by June 2023.

How it solved it. The signal that Runway had found fit was that its video models moved from novelty into real creative pipelines, showing up in an Academy Award–winning film and in television production rather than staying a demo. That validated pull carried commercially: Runway raised $315 million in February 2026 at a $5.3 billion valuation, up from roughly $3 billion the prior year, the kind of trajectory that follows a product people genuinely need rather than merely sample.