Runway started by helping filmmakers. Now it wants to beat Google at AI.

Runway, an AI video generation startup, is taking a unique approach to artificial intelligence development by prioritizing video and "world models" over language-centric AI. While many major AI labs focus on large language models, Runway believes the next frontier of AI intelligence will emerge from systems that learn how the world works through observational data, rather than solely relying on human descriptions. This distinction, though seemingly academic, carries significant implications.

Founded in 2018, Runway built its reputation on powerful video-generation models, including its cutting-edge Gen-4.5. These tools allow users to transform text prompts into editable, cinematic content, revolutionizing production workflows for filmmakers and advertising agencies. Runway's technology has been instrumental in major media projects and has secured partnerships with prominent industry players like Lionsgate and AMC Networks.

Runway's co-founder and co-CEO, Anastasis Germanidis, asserts that training AI models directly on real-world observational data is crucial for advancing AI beyond its current limitations. He argues that language models, while powerful, are inherently bounded by existing human knowledge and biases. To achieve true breakthroughs, Germanidis believes, AI must engage with unbiased data directly from the environment.

The company's innovative strategy extends to "world models," AI systems that can simulate environments and predict outcomes. Runway launched its first world model in December and plans another soon, aiming to create a digital twin of the universe for accelerated scientific experimentation. This ambitious goal stems from the founders' belief that such models can accelerate scientific progress by compressing the time required for research and development.

Runway's unique vision and rapid advancements have positioned it as a significant player in the AI landscape, challenging established norms and attracting a valuation of $5.3 billion. If its bet on video generation and world models pays off, the impact could resonate across diverse sectors, from entertainment to drug discovery, potentially outperforming even tech giants like Google in the race for advanced AI.