- When the Runway founders first came up with the idea for an AI art startup, they said they were crazy.
- Generative AI startups are now fan favorites of both VCs and the public.
- The Runway CEO spoke to Insider about how the company’s technology will make filmmaking more accessible.
When Cristobal Valenzuela started building his generative AI company Runway in 2018, investors, filmmakers and advertisers told the co-founder he was crazy.
“They were like, ‘Why do you want to build an AI tool for video and filmmaking?'” he told Insider. “People didn’t understand the value.”
But five years later, the startup, which offers AI-powered image and video editing tools that can generate complete videos from text prompts, has captured the interest of both investors and the general public. According to Forbes, Runway has already secured nearly $100 million of his funding from VCs including Amplify Partners, Lux Capital, Coatue and Felicis Ventures, and the latest round puts him at a $500 million valuation. , which also serves as an award-winning critical production.blockbusters like All, Anywhere, At Once.
In conversations Insider had with AI-focused VCs, Runway came up again and again as a startup unafraid to push the boundaries of what was possible.
Sara Guo, founder of Conviction, who is not a Runway investor, said, “So many companies fail because of lack of scope ambition.” We figured the way to get people to use it was to build an entire video editor ourselves.”
The intersection of art and technology
Valenzuela is not your typical tech founder. During his undergraduate years he studied economics and business administration, received a master’s degree in design, and teaches at the Adolfo Ibáñez University in Santiago, Chile.
In 2016, Valenzuela left her native Chile to cultivate a renewed interest in artificial intelligence and machine learning by attending NYU’s Interactive Telecommunications Program, a graduate program at the intersection of technology and the arts.
There, he met future co-founders Anastasis Germanidis and Alejandro Matamara Ortiz, and the three of them explored the use of neural networks outside of their usual applications such as self-driving cars, especially in tools for artists and filmmakers. I started looking for ways to use it.
Valenzuela’s diverse background, which spans business, technology and art, is integral to building his product and company, and this multidisciplinary culture is what he’s trying to bring to his runway, he told Insider. The 44-person strong small team is made up of a mix of researchers, scientists, engineers and artists, he said.
“Next to people inventing new computer graphics techniques, there are people working on video effects,” he said.
It can be hard to find a hybrid of a scientist and an artist, but this structure encourages collaboration between teams, such as video and creative teams that quickly validate insights for researchers, and Runway helps end users We guarantee that you will stay close to, Valenzuela explained.
full stack approach
As a “full-stack applied AI research company,” as Valenzuela explains, Runway is made up of three components:
First, the startup builds its own basic models, such as Gen-1 and Gen-2, that can create videos from user-provided text, images, or video clips. The roots of Runway model training really lie in image generation. The startup initially worked with researchers from the University of Munich to build the first version of Stable Diffusion, after which Stability AI trained the model on additional data.
Runway’s infrastructure team then works to deploy these models to customers in a reliable and secure manner. This is a key requirement when creating films of Hollywood scale, he said, Valenzuela.
The final component is Runway’s application layer, which consists of tools anyone from individual creators to large film studios can use to edit images and videos.
Runway’s full-stack strategy has appealed to many Silicon Valley venture capitalists for its flexibility, control and defensibility, the VC told Insider.
However, the startup didn’t originally take this approach to cater to investors, but out of sheer necessity. It wasn’t.” That meant the startup’s founding team needed complete control over the product so they could experiment easily, he explained, Valenzuela.
Due to the time required to build this entire technology stack (4+ years for Runway), no direct competitors of Runway’s scale have emerged, but enterprise use like Synthesia AI-generated video startups working on cases are also gaining attention. Investor.
Currently, the startup uses a consumption-based pricing model, setting different tiers, including free ones, based on the length of the video or number of images users are trying to generate.
art for everyone
Many customers, like New Balance, now use Runway’s tools when brainstorming, storyboarding, and prototyping to get a basic idea of what different approaches might look like. Once they’ve committed to their final plans, filmmakers and creators will be able to combine the startup’s tools to accelerate the filming and production process, including automatically removing and replacing video backgrounds, he added.
Valenzuela said many of these features were developed in collaboration with customers who often describe problems and obstacles they encountered while making videos to Runway.
His past experience in art and design also made it easier for the Runway co-founder to empathize with the hurdles of filmmaking, he said.
In the future, Valenzuela is most excited to build “creative augmentation tools” that work in any modality, from text to audio to video. In many ways it reflects the way humans consume content today. That means watching clips, reading subtitles, and digesting audio at the same time, he said.
And while new technologies can also create divisions between the haves and have-nots, Valenzuela says Runway’s generative AI is helping to improve the quality of video, especially in a world where high-end production films are made by fewer and fewer people. I think it just makes production more accessible and easier. video effects expert.
“Artists of every generation have a responsibility to use the best possible technique to create art,” he said. “Art is a point of view, it doesn’t have to be technical or sophisticated, it just has to convey something meaningful.”
