Pulse News

**How Runway Hopes to Outrun OpenAI, Google in the AI Video Race**

The valuable startup wants to let the giants jostle for the mass market while it focuses on winning over industry professionals and a landmark deal with Lionsgate.

There are two basic ways for startup founders to respond when much larger competitors encroach on their realm. They can soldier on with heads down and voices low, in silent determination. Or they can do as the founders of Runway did when the likes of Google and OpenAI entered the AI video field Runway helped pioneer: They can respond with swagger, hoping to leave observers wondering who the true underdog really is in this fight.

Indeed, Runway CEO Cristóbal Valenzuela didn't sit around and quietly grumble to himself while OpenAI took nearly a year to go from announcing Sora—a rival to Runway's video-generation technology that can conjure up photorealistic clips from a simple text prompt—to releasing its first version of the program in December. Instead, he publicly mocked his much bigger competitor. "I am announcing that there will be an announcement about an upcoming announcement where we will announce an announcement," he wrote on X, a thinly veiled critique of OpenAI's propensity for teasing future products that aren't quite ready for public consumption.

When I saw him a week after Sora's release at Runway's New York headquarters, Valenzuela gamely kept up his bluster. Indeed, he and his two co-founders, Anastasis Germanidis and Alejandro Matamala Ortiz, seemed more than a little fed up that OpenAI had immediately been crowned the victor in the race to develop AI video programs—with Google's Veo 2 seen as the only credible competitor. The founders could only shake their heads in disbelief: Runway has millions of monthly users and a head start of several years in AI video. The ink is barely dry on a flashy new partnership with Lionsgate, a major Hollywood studio, and Runway plans to begin funding filmmakers directly, further entrenching itself in entertainment. And it has hundreds of millions of dollars from investors such as Nvidia and General Atlantic.

With text-based chatbots more or less all capable of delivering the same thing at this point, video is quickly shaping up as the next lucrative front in AI. The Runway trio views their startup as the industry leader, with OpenAI and Google as the erstwhile competition.

"We see these other companies as creating concept cars while we're building a real car," said Valenzuela, 35, dressed in the standard uniform of every self-respecting urban artist: black T-shirt, black jeans, hipster sneakers. "A concept car can be a beautiful design-son that shows what you think you can achieve. But producing and selling 10 million cars is a whole different challenge. Right now, everyone's obsessed with concept cars—"

"Made by the same companies making concept bicycles and concept everything," said Matamala Ortiz, 40, who was dressed more or less like his partner.

"But our job is to make cars that are useful to people," Valenzuela said.

Is there a more improbable startup in AI than Runway? The prototypical venture in AI is founded by deep-learning specialists from places like Stanford University or the Massachusetts Institute of Technology. Runway, in contrast, emerged in 2018 from a program inside the Tisch School of the Arts at New York University, a place better known for shaping filmmakers than tech entrepreneurs. The three founders were wannabe artists who fell into the field, born of the desire to experiment with a new medium.

Yet Runway, which last year closed on a $450 million round valuing it at around $4 billion, boasts an impressive roster of marquee investors, including Google, Nvidia, and General Atlantic. Its business is growing quickly, too: The company hopes to hit $300 million in annualized revenue this year after having reached $80 million in annualized revenue in December 2024. That could help demonstrate to the wider world that there might even be money to be made building AI video models.

No less improbable is how Runway has proven its moneymaking potential: by convincing many in Hollywood that AI is actually good for the movie and television businesses. In 2022, Valenzuela began making regular treks to Hollywood to win over a skeptical—and sometimes outright hostile—industry. By 2024, he was spending roughly one week of each month in California, bouncing between meetings in Burbank, Culver City, Bel Air, Santa Monica, and the Hollywood Hills.

"I'll start at 9 a.m. and try to fit in at least six meetings a day," Valenzuela said. "Producers, actors, writers, directors, the guilds, the big studios, the production houses—pretty much everyone."

The general sentiment in Hollywood has been that AI meant flicking a switch—and presto, you have a movie. That would mean eliminating scores of human jobs. Instead, Valenzuela has labored to persuade Hollywood to view Runway as a tool for enhancing their creative work—and to convince the Town that Runway would better suit their needs than the tools released by OpenAI and Google, which Valenzuela thinks were designed for casual users. The argument isn't dissimilar to the one Adobe made a generation earlier to win over professional photographers to Photoshop and Lightroom.

In Valenzuela's whirlwind tours, his strategy has been to demonstrate the many practical ways Runway's tools could help professionals accomplish more ambitious creative work with the same budget—enabling production teams to create films with special effects that wouldn't be financially feasible without AI assistance, for example, rather than replacing the teams themselves.

Using Runway's tools, for example, a filmmaker can shift a camera angle even after a production has wrapped. They can swap in a new background, saving themselves the trouble and expense of creating one specifically for a shoot. They can change an outfit that isn't working for a character or retroactively give a shot the feel of a handheld camera. They can use a set of tools to impose expressions and head movements onto a cartoon figure or a famous actor's face.

One Hollywood-based Runway enthusiast gave the example of a scene from a movie script calling for 1,000 medieval soldiers on horseback lined up on a snowy hillside. That would have been prohibitively expensive to create using traditional special effects—but it's now possible to do it cost efficiently with Runway's AI.

Valenzuela's efforts paid off in a high-profile way when in September Lionsgate, the studio behind film franchises such as John Wick, The Twilight Saga, and The Hunger Games, announced a first-of-its-kind deal that allowed Runway to train a new model on the studio's 20,000-plus titles. The studios, production shops, and others experimenting with Runway generally are hesitant to speak publicly about their use of AI out of the fear of being seen as siding with the Cylons against actors, screenwriters, and others worried about being replaced.

"No one wants a bull's-eye on their back," explained one public relations representative when turning down an interview request on behalf of his Hollywood client.

The Lionsgate deal marked a watershed moment: a major studio stepping out of the shadows to openly embrace AI in the making of films. "If used correctly, Runway is going to enable films and television shows to get made that might not have ever been made before," said Lionsgate Vice Chair Michael Burns. Burns, who first met Valenzuela over lunch in November 2023, spoke of the "millions and millions of dollars" the studio would save by using AI. "I look at this as one of the great tools for the industry, and it's just the tip of the iceberg right now," Burns said.

And it's not just Hollywood. Nearly every big company these days employs creative teams to make videos for internal and external use, or they hire firms like Monks, a global marketing and production company that started using Runway a couple of years ago. Using Runway's tools, said Debora den Iseger, Monks' senior vice president of content and innovation, a project that in the past might have required 50 to 60 people can now be completed with just 15 to 20. Production timelines have shrunk from an average of 16 weeks to six. "We save between 50 and 80 percent on videos we create for our clients," she said.

Production companies and filmmakers could use Sora or another model to accomplish much of what they can do with Runway. OpenAI, of course, is a better funded, better known company and has in its employ an army of researchers able to match Runway's offerings feature for feature. But Runway's great hope is that its relationships with creatives give it a leg up in an increasingly competitive landscape that includes not just OpenAI and Google but Meta Platforms, Adobe, Tencent, ByteDance, and a raft of high-profile startups founded by AI researchers and funded by top-tier venture firms.

"When I've had meetings with other AI research labs, it's been striking how little they know about the industry they're trying to work with," Valenzuela said. While its competitors create models that impress their fellow engineers, Runway focuses on developing pro tools for film production and post-production processes that have quickly become the industry standard. "They're like brain surgeons, not general practitioners," Burns said of Runway. "They've really mastered this one specialty, which is why they're far ahead of the competition."

Unlike the typical Silicon Valley denizen who idolizes tech founders and venture capitalists, Valenzuela finds his heroes in a different world entirely. "Silicon Valley has this obsession with founders and investors," Valenzuela said. "I personally get really excited when I meet a legendary filmmaker." Growing up in Chile, Valenzuela dreamed of making movies. "But Chile does not have many really good film schools," he said. Instead, he studied economics and design, finding work with a local consulting firm, while on the side he pursued his passion for multimedia art installations that he exhibited in local galleries.

In 2016, Valenzuela met his co-founders at NYU's Interactive Telecommunications Program, which has been described as an art school for engineers and an engineering school for artists. Matamala Ortiz, who is also from Chile, had founded his own design studio at 19. "I was looking to focus on coding and creating cutting-edge tools for designers," he said. The Athens-born Germanidis had a computer science background and had devoted much of his energy to making computer-aided art. "There was starting to be this movement around neural networks," Germanidis said of his decision to get his master's degree, "and I wanted to be part of it."

Runway started as Valenzuela's master's thesis—an exploration of the various AI models available for artists and nearly ended there. Adobe offered him a job with a salary that "worked out to more money than my family had made in the previous 10 years combined," Valenzuela said. He had never thought of starting his own company and knew nothing about raising money for a startup—let alone the names of any VCs. Valenzuela shocked even himself when he turned Adobe down to continue working on what he had built. He and Matamala Ortiz lived on $2,000 a month working part-time at NYU, while Germanidis took a job as a back-end engineer at Zocdoc. Germanidis quit six months later to join his former classmates. Germanidis, who is one year younger than Valenzuela, took on the chief technology officer title, and Matamala Ortiz became chief design officer.

Fortunately for Valenzuela and the others, they didn't need to figure out venture capital to get going. The money found them. Sunil Dhaliwal, founder of Amplify Partners, met the trio in late 2018, when the founders were in Montreal for that year's Neural Information Processing Systems conference, the biggest event in AI. "They were a refreshing contrast to all those people who are like, 'We're the smartest computer scientists in the world; if we just align the ones and zeroes the right way, everyone will bend to our will,'" Dhaliwal said. "They weren't the most AI guys around, but they were the most passionate and clear-thinking creatives we came across. We basically just begged them to get involved." So Amplify wrote a $500,000 check for Runway's seed round, one of three investors in that funding.

Finding new funders, however, would prove to be a greater challenge. Many VCs are in the pattern-matching business, and Runway fit no one's template of a tech startup. "The VCs would be like, 'You went to some art school—you'll never know how to build a business,'" Valenzuela said. Those who showed initial interest seemed even less convinced of Runway's potential once they did their due diligence. "They're like, 'No one wants to invest in artists because artists have no money, which means no market,'" he said. It didn't help Runway's cause that the technology could produce little beyond fuzzy, sometimes distorted images.

Runway racked up more than 30 rejections over two years before announcing, in December 2020, an $8.5 million Series A round that relied heavily on the original three investors. "It was a leap of faith for all of us involved," said James Hong, founder of Hot or Not, who was among those taking part in Runway's A round. "You had to tell yourself, 'This is still crap, but you should have seen what it looked like last week.'"

Other AI startups famously spent lavishly to woo top AI researchers. Runway's first hire, in contrast, was Brannon Dorsey, who had earned his undergraduate degree at the School of the Art Institute of Chicago three years earlier. The company's second hire was also an artist who had embraced technology. Runway released its first product in mid-2019, just six months after its founding. ("As soon as we had something usable, we wanted to get it in people's hands," said Dorsey, offering a not-so-subtle dig at OpenAI.)

The first release of Runway was a repository of open-source models for artists and other creatives made by other people. Quickly, though, they realized the models' limits. "We know the level of control that's required to produce high-quality video," Matamala Ortiz said. "And we knew we couldn't build the applications we envisioned without control over the full stack supporting them."

Runway's AI technology was used to complete the Oscar-winning film "Everything Everywhere All At Once."

Runway needed to be both a research lab and product company if it was going to succeed. The founders raised another $35 million at the end of 2021, which they used to move into an office in New York's Tribeca neighborhood—a hub for film and media—and add research scientists to its staff. One standout hire, Patrick Esser, co-authored a landmark paper, "High-Resolution Image Synthesis With Latent Diffusion Models," published in April 2022. That research introduced latent diffusion models, a more efficient method for generating high-quality images. This approach became the foundation for Stable Diffusion, an open-source AI image generator, released later that summer by Stability AI, which democratized access to high-quality AI image creation tools.

"That was the inflection point," Valenzuela said. That, he said, is what told the wider world that AI video was "more than a toy." At the end of 2022, Runway secured another $50 million in a Series C that valued the company at $500 million—followed by an additional $141 million Series C extension that included Google, Nvidia, and Salesforce as investors. The release of ChatGPT and the explosion of interest in generative AI elevated Runway's profile as a pioneer of text-to-video technology.

The company's tools were starting to make a mark in the real world. Visual effects artist Evan Halleck discovered Runway while he was working on "Everything Everywhere All at Once," a film that practically swept the 2023 Oscars with seven awards. The team used Runway to make googly-eyed rocks walk along a canyon's edge in a pivotal scene in the movie. "I was cutting out the characters, placing them cleanly on a plate shot in minutes versus what takes half a day," Halleck told Variety in 2023.

The comedy team at "The Late Show With Stephen Colbert" started using Runway to create video gags. "I've done a shot in five minutes, which normally would have taken me five hours," said Ryan Mauskopf, a graphic artist with "The Late Show." "You're not going through it frame by frame, you're just tapping a couple of things and it does it."

By mid-2023—long before any large competitors had entered the field—Runway had developed 35 "magic tools" catering to professionals working in all aspects of video creation, including animation, visual effects, and post-production work. "It took five years, but we proved that video is worth spending time on," Valenzuela said. "Then everyone suddenly realized it was worth building."

Today, Runway operates out of spacious offices in New York's Chelsea neighborhood that feel like they're part art gallery and part design studio. One corner houses a production studio staffed by a dozen creatives who produce content aimed at demonstrating Runway's capabilities. "Half our team has an arts background," Valenzuela said of a staff that has grown to around 100 employees. By staying in close contact with creatives, he said, "we'll stay ahead of the competition."

At most AI companies, research scientists and product people occupy different spheres. Inside Runway, they sit at adjoining desks. "At one table, you've got a Ph.D. who's been contributing to fundamental research in the space working really closely with someone who has been editing and post-producing films or content for the previous 20 years," Valenzuela said.

Building on this fusion of technology and creativity, Runway inaugurated an annual AI Film Festival that last year drew 3,000 entries. The 2025 festival this spring will showcase the works of 10 finalists at screenings in both New York and Los Angeles. In September, the company announced that it was devoting $5 million to The Hundred Film Fund, which it established to encourage those who want to push the boundaries of the medium. One film fund recipient, Jeremy Higgins, has been using Runway's tools for the past couple of years to make short animated clips both as art and for clients.

Landing Lionsgate has opened new horizons for Runway. Valenzuela stressed the significance of working closely with a studio that's intent on using its tools to widen the palette of its filmmaking—not incidentally saving a few bucks along the way. But the strategic alliance also served as a neon sign for a new kind of deal that moves Runway beyond a service that charges $95 per user per month. "After the Lionsgate announcement, it's like half of Hollywood has been reaching out to get the same deal or try to learn more about it," Valenzuela said.

In December, Oscar-winning writer and filmmaker Roger Avary gave the company another high-profile endorsement while appearing on "The Joe Rogan Experience" alongside Quentin Tarantino. Avary recounted Valenzuela's promise: "Literally, whatever you think you can't do, ask us because we probably will be able to do it in a couple of days."

So ingr