Runway started by supporting filmmakers. Now they are trying to beat Google with AI.

AI video generation startup Runway doesn’t have a typical Silicon Valley pedigree. There are no Stanford University founders, no former Google founders, no nine-figure seed rounds that bought time by ignoring revenue. The three founders (two from Chile and one from Greece) met at New York University’s Tisch School of the Arts and founded the company in New York.

Depending on who you ask, Runway could be one of the most influential AI companies today. Not because of what we’ve built so far, but because of what we’re going to build next.

For the past few years, the AI ​​industry has largely operated on the premise that intelligence resides in language. Large-scale language models like OpenAI’s ChatGPT and Anthropic’s Claude reflect that bet.

Runway, like all of its competitors, creates something different. Its founders believe that the next form of AI intelligence will not be built from text, but from videos and world models that learn not just how humans describe the world, but how the world works. That distinction sounds academic. That’s not what it means.

Runway co-founder and co-CEO Anastasis Germanidis said that training models directly on global observational data is the next frontier in AI. The first companies to get there won’t be the ones that perfect the language, he argues.

“We are fundamentally bound by our own understanding of reality,” Germanidis tells TechCrunch from Runway’s homey, sunny headquarters near Union Square.

“Language models are trained across the internet, on message boards and social media, and in textbooks to extract existing human knowledge,” Germanidis continued. “But to go beyond that, we need to leverage less biased data.”

Founded in 2018, Runway has built a reputation for video generation models, including the latest Gen-4.5, and AI tools that can transform text prompts into editable, cinematic content.

Today, Runway’s technology powers the production workflows of filmmakers and advertising agencies, and the company has signed deals with major media players like Lionsgate and AMC Networks. Its tools have also been used in films such as “Everything Everywhere All At Once.”

Runway is currently valued at $5.3 billion, and one of its founders says it will generate $40 million in annual recurring revenue in the second quarter of 2026.

If Runway’s bet that video generation is the path to becoming a global model pays off, the results will spread from Hollywood to drug discovery. Otherwise, Runway risks being surpassed by a much better-funded competitor: the top Google.

take a leap

Within the past six months, the startup has put its plans into action and expanded beyond video generation, launching its first global model in December and plans to launch another this year. (A world model is an AI system that simulates the environment enough to predict how it will behave.)

Runway is not alone in its pursuit of turning physics-aware video models into world models for short-term use cases such as interactive entertainment, gaming, and robot training. Startups Luma and World Labs are on a similar trajectory, and Google is also moving its Genie world model in the same direction.

Everyone wants some version of the same thing: AI that solves humanity’s toughest problems. While this is far from Runway’s original product, it is the result of both the emerging capabilities of technology and its founders’ inclination to follow where that technology leads.

Germanidis sees the world model as a scientific infrastructure. The more sensory data and observations that train a single model, the closer we get to a working digital twin of the universe, and the faster we can run experiments than any lab. He points out that much of the scientific process is just waiting for results. If we can compress this waiting time, we can also compress the progress itself.

“If we can develop scientists who are better than human scientists, we can accelerate progress in the way we understand the universe and solve problems,” Germanidis said.

moon shot

Germanidis fell in love with programming at the age of 11 in Athens and came to the United States at 18 to study neuroscience and film. He returned to computer science and worked at several technology companies in Silicon Valley until he decided he had enough of that culture. Co-CEO Cristóbal Valenzuela was born and raised in Santiago and studied economics as an undergraduate before working in film and then software. Chief innovation officer Alejandro Matamara Ortiz, also from Santiago, studied advertising and ran a design company.

The three met in 2016 while attending New York University’s ITP (Interactive Communications Program), a graduate program that Valenzuela describes as “an art school for engineers.”

Matamara-Ortiz said the co-founders all aspired to be filmmakers at some point in their lives. So Runway started with a simple mission. Can everyone become a filmmaker using AI?

Matamara-Ortiz said that after releasing its first video generation model in February 2023 (which is surprisingly unimpressive compared to what Runway is putting out now), its mission evolved to: “Can anyone be a great filmmaker?”

We needed to grow the team into what it is today. The company has 155 employees and offices in New York, London, San Francisco, Seattle, Tel Aviv, and most recently, Tokyo. “But through this process, we learned that these models can help us understand how the world works and, if scaled, can be useful for many other different things,” he added.

Robotics, drug discovery, climate modeling, and more are the types of problems that have puzzled researchers for decades. Last year, Runway launched a robotics division, which Germanidis said is already undergoing real-world testing and deployment.

Germanidis, like others, sees the field moving toward training a single model with different modalities, such as text, video, audio, and other sensors, and believes compounding effects are important.

Given enough time and resources, a moonshot goal for his own Runway technology is biological world modeling and anti-aging research.

The jury is still out on whether Runway can bring its video advantage to the world’s models, and its competitors aren’t waiting either. Runway was one of the first companies to tackle AI video generation, but World Models is in a different breed with well-funded and well-regarded competitors. Google, former meta-chief scientist Yan Lekun, AI’s “godmother” Fei-Fei Li, and the growing startup field are all chasing the same goal.

Kian Catanforouch, CEO of AI skills benchmarking company Workera and lecturer at Stanford University, pointed out that while no one has yet proven the jump between video intelligence and generalized inference via world models, that doesn’t mean it’s impossible. He said if Runway wants to make its global model bet a reality, it needs to keep gathering resources, including top-of-the-line computers.

Runway has contracts with CoreWeave and Nvidia, but hasn’t confirmed whether they have dedicated cluster access (the type of guaranteed large-scale computing needed to train Frontier models).

“How can we build a foundational model without clusters?” asked Catanforouche. “I don’t think anyone can do that.”

Runway has raised $860 million to date, including a $315 million round in February from strategic partners including AMD Ventures and Nvidia. That’s about the same level as its closest competitors, Luma AI and World Labs, which have raised $900 million and $1.29 billion, respectively, according to PitchBook.

But Runway will also face off against incumbents like OpenAI, which has raised about $175 billion for CEO Sam Altman, and tech giant Google, whose parent company Alphabet is valued at $4.86 trillion. Google is Runway’s biggest threat. The company’s Veo model competes directly with Runway’s video generation business, while its Genie World model targets the same long-term space that Runway is aiming for.

Katanforoosh nods to OpenAI’s shutdown of video platform Sora in March, saying it was spending about $1 million in computing costs per day and had just $2.1 million in revenue. His argument is that resources alone cannot guarantee survival. There are no guarantees about the runway either.

Katanforoosh has no intention of canceling Runway. He pointed to AI audio startup ElevenLabs, which outperformed OpenAI and Google in its own benchmarks despite lacking resources and track record. Runway could follow a similar strategy, he argues.

Runway’s founders aren’t immune to this comparison. Valenzuela said the startup’s lack of Bay Area “standardization” gives it an advantage. He argues that not only are they more diverse in their ideas, but without their connections to Silicon Valley, they might have been forced to be more aggressive, and they lacked the war chest that many of their peers had access to, meaning they didn’t need to generate revenue early.

And Michelle Kwon, Runway’s chief operating officer, said the company is in no rush to raise more funding, even though its computing demands are increasing as it scales.

“Their background has allowed them to build a culture that moves early, often gets right, and moves incredibly quickly,” Michael Dempsey, an early investor and managing partner at Compound, told TechCrunch.

For Valenzuela, that culture starts with how he sees the world in the first place. As a co-CEO and a new father, he spends all his free time reading. One of them describes the Chilean poet Nicanor Parra as the antithesis of Pablo Neruda. Informal and non-academic, he holds the view that poetry belongs to the people, not the rules.

“The rules are just rules that they invented,” Valenzuela said. “That’s the driving force behind how we do things on the runway. They say Silicon Valley is here and startups are there. Why? Those are just rules that were made. Erase everything and start over.”