Mixed Creative and Research Teams Win
Cristóbal Valenzuela, CEO of Runway, on the state of generative AI in video
In AI video, product advantage comes from turning raw model capability into tools creatives actually want to use, which is why mixed teams matter more than pure research depth. Runway built around artists, researchers, and product builders because video quality is judged by the human eye, and the work is not just training models, but shaping motion, pacing, collaboration, and editing workflows into software that feels fast and useful in production.
-
Runway’s own workflow shows why. The company trains models, builds video ingestion and rendering systems, and packages them into browser based tools for rotoscoping, inpainting, subtitles, and generation. That stack needs technical model work and taste about what an editor needs on screen, frame by frame.
-
The user base is not just ML engineers. Runway sells to filmmakers, marketers, post production teams, and small businesses making ads, reels, demos, and VFX shots. A team building for all of those users needs people who understand storytelling, aesthetics, and messy creative workflows, not just benchmark performance.
-
This is also how Runway differs from both horizontal model labs and narrow wrappers. OpenAI and Google can bundle video into larger AI suites, while tools like Pika and OpusClip focus on specific jobs. Runway is trying to own the full filmmaking workflow, which rewards teams that combine research talent with product and creative judgment.
The next wave of AI video companies will look less like pure labs and more like new age creative studios with deep engineering. As video models improve, the winners will be the teams that can spot new creative behaviors early, turn them into repeatable tools, and ship them into real production workflows before larger rivals do.