AI is getting ever-better at generating video, and this potential is being reflected in the valuations of pure-play AI video generation companies.
Runway, one of the pioneers in AI-powered video generation tools, announced today it has secured $315 million in Series E funding at a $5.3 billion valuation, marking a significant milestone for the generative AI sector as venture capital continues flowing into frontier model development.
General Atlantic led the round, with participation from a roster of strategic and financial investors including NVIDIA, Adobe Ventures, AllianceBernstein, AMD Ventures, Fidelity Management & Research Company, Mirae Asset, Emphatic Capital, Felicis, and Premji Invest.
In a blog post, co-founder and CEO Cristóbal Valenzuela framed the capital raise around what he called “world models”—advanced AI systems capable of simulating realistic video content. “World models are the most transformative technology of our time,” Valenzuela wrote. “This capital lets us pre-train the next generation of world models and bring them to new products and industries.”
From Art School Project to Generative Video Leader
Founded in 2018 by Valenzuela, Alejandro Matamala, and Anastasis Germanidis—three graduates of NYU’s Tisch School of the Arts—Runway began as a machine learning toolkit for artists before pivoting to video generation and editing. The New York City-based startup has since established itself as a leading platform for AI-driven media production, competing in an increasingly crowded field that includes Google’s Veo3, OpenAI’s Sora, and a host of Chinese names.
Runway’s web-based platform features over 30 “Magic Tools” enabling text-to-video generation, image-to-video conversion, object removal, slow-motion effects, and sophisticated editing capabilities like motion brush and camera controls. The company has released several flagship models, including Gen-3 Alpha for high-fidelity generation up to 10 seconds, Gen-4 with improved character and environment consistency, and its latest Gen-4.5, which the company says excels in benchmarks for HD video generation from text prompts.
Hollywood to Social Media: Broad Market Adoption
The startup has carved out a notable presence in professional filmmaking and advertising. Runway’s tools contributed to visual effects in the Oscar-winning film Everything Everywhere All at Once and have been adopted by The Late Show with Stephen Colbert for production workflows. The platform’s ability to prototype scenes and create effects without expensive equipment or lengthy production cycles has resonated with creative professionals seeking to accelerate turnaround times.
Beyond high-end production, Runway serves the exploding demand for short-form video content on platforms like Instagram Reels and YouTube Shorts. The company offers a credits-based pricing model spanning individual creators to enterprise clients, with an iOS app extending its reach to mobile users. Real-time collaboration features and integrations with tools like Adobe Firefly have helped embed Runway into existing creative workflows.
Strategic Backing From Tech Giants
The participation of NVIDIA and Adobe Ventures signals strong strategic interest from companies positioned across the AI value chain. NVIDIA, whose GPUs power most generative AI training and inference, has been an active investor in AI application companies. Adobe’s involvement is particularly notable given the company’s parallel development of its own Firefly video generation tools—suggesting potential partnership or integration opportunities despite competitive dynamics.
The inclusion of AMD Ventures also highlights intensifying competition in AI chip markets, as AMD seeks to expand its footprint in the generative AI ecosystem beyond NVIDIA’s dominance.
The World Models Vision
Valenzuela’s emphasis on “world models” reflects an ambitious technical direction. While the term has various interpretations in AI research, in the context of video generation it typically refers to systems that learn comprehensive representations of how the physical world operates—understanding physics, object permanence, spatial relationships, and temporal consistency. Such capabilities would represent a significant leap beyond current text-to-video models, which often struggle with maintaining consistency across longer sequences or accurately modeling complex physical interactions. Several companies are working on such models, including Google, which has recently made Genie 3 available to the public, and pioneering AI researcher Fei-Fei Li’s new startup named World Labs.