Anastasis Germanidis is a co-founder and co-CEO of Runway, an AI research company developing generative video and interactive world models. His work combines machine learning with tools for artists, filmmakers and simulated environments.[1][2]
Start with Runway’s WorldPrompt and the Engineering of Real-Time Worlds →Background
NYU Tisch identifies Germanidis as an Interactive Telecommunications Program alumnus from the class of 2018. He met his Runway co-founders through that program. In an early project with Cristóbal Valenzuela, Uncanny Road, a user painted a simple arrangement of roads, sky and buildings; a model converted the color map into a photo-like scene. The interface made a model’s image-generation ability available through a familiar drawing operation, anticipating the creative-tool focus of Runway.[4][5]
Career
Germanidis spent much of Runway’s history as chief technology officer. In his September 2026 Outliers interview, he says his role recently changed to co-CEO, while much of his day-to-day collaboration with co-founder Cristóbal Valenzuela remained similar. Runway’s current summit page names him co-CEO and presents him in its opening keynote and research discussion. The company’s program places world models, real-time video and robotics together as parts of its broader real-world intelligence work.[2][1]
Contributions
In his Runway interview, Germanidis describes world models as an extension of the company’s video-model research. Generating plausible motion requires some representation of how the visual world changes. An interactive system adds actions so the next frames can respond to a user or controller. The conversation considers uses beyond making a finished clip, including simulated environments and interfaces. These are research and product directions for Runway.[3]
Tools, writing & talks
Writing, research & talks
Start here
Conversation about interactive world models, real-time video and simulation.
Start with the WorldPrompt conversation for an explanation of what interaction adds to generated video. Basic familiarity with generative AI helps; Uncanny Road is a concrete earlier example.
Interview guest and Runway research leader.[3]Original source- Runway’s Bet Beyond Video: World Models, Robotics, and the Neural OS — Anastasis Germanidisyoutube.com
Runway's co-founder covers the path from Gen-1 to Gen-4.5, world models as physics simulators for robotics, neural-rendered interfaces, real-time generation, and a proposed test for world models.
Catalog workOriginal source - Uncanny Roadtisch.nyu.edu
An early drawing-to-image experiment that turns a simple scene map into photo-like textures.
Co-creator with Cristóbal Valenzuela.[5] - Runway AI Summitsummit.runwayml.com
Company talks and research discussions spanning world models, robotics and real-time video.
Opening-keynote speaker and research-discussion participant.[1]