Runway as AI Infrastructure Provider

These partnerships turn Runway from a software vendor into an AI infrastructure provider for rights holders. Getty and Lionsgate are not just buying seats in a video editor, they are supplying licensed libraries that let Runway train models other companies cannot legally replicate. That creates a second business alongside subscriptions and APIs, where Runway can sell custom model development and ongoing licenses to large media, advertising, and entertainment customers that need safe training data and outputs tied to their own style and catalog.

• The Getty deal was built around an enterprise ready model trained on Getty's fully licensed creative library. In practice, that gives brands and agencies a way to generate campaign visuals and video from approved source material, without the legal ambiguity of scraping the open web.
  2 runwayml 4 axios
• The Lionsgate deal is more specific and higher value. Runway got access to a proprietary catalog of roughly 20,000 film and TV titles, and Lionsgate can use the resulting model in preproduction and postproduction workflows like storyboards, draft scenes, backgrounds, and effects. That is much closer to selling production infrastructure than creator software.
  1 lionsgate 3 sacra
• This is a real competitive wedge versus tools like Pika and OpusClip, which are strong product experiences but do not pair proprietary models with exclusive licensed media libraries. Runway's full stack matters here, because it can train the model, deploy it in production, and wrap it in editing workflows the customer already uses.
  3 sacra 5 sacra 6 sacra

The next step is a market where every large studio, broadcaster, stock media owner, and agency holding company wants its own house model. If that plays out, Runway can compound from selling faster editing into operating the licensed model layer for media companies, with higher contract values, stickier data relationships, and a stronger moat around enterprise video AI.