Runway AI launches $10M fund

Runway is no longer content to sell AI tools. It now wants to shape the companies that are built with them. On March 31, Runway announced a new $10 million fund for early-stage startups working across AI, media, and what it calls world simulation. The company says it will usually write checks of up to $500,000 for pre-seed and seed-stage teams, with room to expand the fund later. It also said this is not a brand-new experiment. Runway has already been investing quietly for several years. (runwayml.com) That quiet part matters, because it shows what this launch really is. Runway is formalizing a strategy it has already been testing. The company says its past bets include Cartesia, LanceDB, and Tamarind Bio. Those names are revealing. Cartesia works on real-time audio generation. LanceDB builds database infrastructure for AI systems. Tamarind Bio uses AI for protein design. This is a wider map than AI filmmaking. Runway is signaling that it sees its future not just in making prettier video clips, but in backing the infrastructure and applications that might grow around multimodal models. (runwayml.com) That helps explain why the fund arrived alongside a second announcement. The same day, Runway launched Builders, a startup program that gives Seed through Series C companies free API credits and early access to some of Runway’s newer products. The program is already live with a first cohort that includes Cartesia, MSCHF, Oasys Health, Spara, Subject, and Supersonik. On its public application page, Runway says startups can receive up to 500,000 API credits, Tier 5 usage limits, and access to Runway Characters. Money gets founders in the door. Credits and product access give them a reason to stay. (runwayml.com) This is the part of the story that matters. Runway is trying to become a platform company. Its pitch is that generative video and conversational AI are converging, and that the next wave of products will be built on top of that stack. TechCrunch reported that Runway’s founders describe this as a push toward “video intelligence,” meaning interactive and real-time systems rather than one-off media generation. That sounds abstract until you look at what Runway has been shipping. In December 2025, it introduced GWM-1, a “general world model” designed to simulate reality in real time. Runway says the model family includes versions for explorable environments, conversational avatars, and robotics. (techcrunch.com) Seen in that light, the fund is less about venture investing than product strategy. Runway cannot possibly build every application for avatars, simulated worlds, customer support, entertainment, training, and robotics by itself. Startups can explore those edges faster. If they build on Runway’s APIs, they also help prove that Runway’s models are useful beyond filmmaking and advertising, the markets that first made the company famous. The fund’s three thesis areas make that ambition plain: frontier AI research, application-layer products, and new forms of media and content. (runwayml.com) The timing is not accidental. In February, Runway announced a $315 million Series E led by General Atlantic, with participation from Nvidia, Adobe Ventures, AMD Ventures, Fidelity, and others, saying the money would help it train the next generation of world models. TechCrunch reported that the company has now raised close to $860 million in total and carries a post-money valuation of about $5.3 billion. A company with that much capital does not launch a $10 million fund because it needs financial returns from tiny startup checks. It does it to steer an ecosystem. (runwayml.com) Runway has tried versions of this before in narrower form. In September 2024, it launched the Hundred Film Fund to support AI-augmented film projects, with grants ranging from $5,000 to more than $1 million and up to $2 million in Runway credits across projects. That program was aimed at artists. The new fund is aimed at founders. Same instinct, bigger target. The company that once sold creative software is now trying to finance the layer above it, one $500,000 check at a time. (runwayml.com)