Runway started by helping filmmakers. Now it wants to beat Google at AI.

Runway, a startup focused on AI video generation, does not fit the usual Silicon Valley mold. It wasn’t founded by Stanford alumni, ex-Google employees, or supported by a massive initial funding round that allowed them to overlook revenue. Instead, its three founders—two from Chile and one from Greece—came together at NYU’s Tisch School of the Arts and established the company in New York.

Some consider Runway to be among the most influential AI companies today, not for its past achievements but for its future ambitions.

For years, the AI industry has primarily viewed intelligence as being rooted in language, a belief reflected in large language models like OpenAI’s ChatGPT and Anthropic’s Claude.

Runway, however, is taking a different approach. Its founders are convinced that the future of AI intelligence will be derived from video and world models that learn through observing the world, rather than relying solely on human descriptions. This may seem theoretical, but its consequences are significant.

According to Anastasis Germanidis, Runway co-Founder and co-CEO, training AI models on real-world data represents the next leap in AI development. He argues that the companies pioneering this approach won’t be those who excel in language alone.

“We’re essentially constrained by our perception of reality,” Germanidis shared with JS from Runway’s cozy, sunlit office near Union Square.

“Language models distill existing human knowledge from the internet, social media, and textbooks,” he continued. “To advance beyond that, we need to utilize less biased data.”

Since its founding in 2018, Runway has gained recognition for its video-generation models, including the latest Gen-4.5, and AI tools that transform text prompts into editable cinematic content.

Today, Runway’s technology enhances production workflows for filmmakers and advertising agencies. The company has partnerships with prominent media companies like Lionsgate and AMC Networks, and its tools have been utilized in films such as “Everything Everywhere All At Once.”

As of now, Runway is valued at $5.3 billion and, according to one of its founders, generated $40 million in annual recurring revenue in the second quarter of 2026.

If Runway’s strategy of focusing on video generation as a path to world models succeeds, its impact will span industries from Hollywood to pharmaceutical research. However, if it fails, Runway risks being surpassed by competitors with greater financial resources, particularly Google.

Taking the leap

In the past six months, Runway has executed its plan to venture beyond video generation. It launched its first world model in December and plans to introduce another by the end of this year. World models are AI systems capable of simulating environments to predict their behavior.

Runway is not alone in this endeavor. Competitors like Luma and World Labs are on a similar path, and Google is steering its Genie world model in the same direction.

The goal is universal: to develop AI that addresses humanity’s most complex challenges. This ambition marks a significant shift from Runway’s initial focus, driven by technological advancements and founders eager to explore new possibilities.

Germanidis views world models as essential scientific infrastructure. The more sensory data and observations a model is trained on, the closer it comes to creating a digital twin of the universe, enabling experiments far faster than any physical lab. He notes that much of the scientific process involves waiting for results. By reducing this waiting time, progress could be accelerated.

“If we can create a scientist superior to human scientists, we can speed up our understanding of the universe and solve problems more effectively,” Germanidis stated.

The moonshot

Germanidis discovered his passion for programming at age 11 in Athens and moved to the U.S. at 18 to study neuroscience and film. He later switched to computer science, working at various Silicon Valley tech companies before deciding to leave the culture behind. Co-CEO Cristóbal Valenzuela, from Santiago, studied economics before transitioning to film and software. Alejandro Matamala-Ortiz, also from Santiago, studied advertising and ran a design firm.

The trio met in 2016 at NYU’s ITP (Interactive Communications Program), which Valenzuela describes as an “art school for engineers.”

According to Matamala-Ortiz, the co-founders once dreamed of becoming filmmakers. Thus, Runway began with a straightforward mission: to make filmmaking accessible to everyone through AI.

Their first video generation model launched in February 2023, which was not particularly impressive compared to their current offerings. This prompted a shift in their mission: could they help everyone become a great filmmaker?

This vision required expanding their team to its current size of 155 employees across offices in New York, London, San Francisco, Seattle, Tel Aviv, and Tokyo. “Throughout this journey, we realized these models could grasp how the world operates and, with scaling, serve numerous other purposes,” Matamala-Ortiz explained.

Runway’s technology now extends to fields such as robotics, drug discovery, and climate modeling—areas that have long challenged researchers. Last year, they launched a robotics unit, which Germanidis claims has already achieved real-world testing and deployments.

Germanidis, along with others, sees the future in training a unified model across various modalities—text, video, voice, and other sensors—believing the compounding effect is vital.

His ambitious aim for Runway involves biological world models and anti-aging research, given sufficient time and resources.

Whether Runway can leverage its video expertise to excel in world models remains uncertain, especially with competition from well-funded and respected rivals. Google, former Meta scientist Yann LeCun, AI pioneer Fei-Fei Li, and numerous startups are all striving toward the same goal.

Kian Katanforoosh, CEO of Workera and a Stanford lecturer, noted that no one has yet demonstrated the transition from video intelligence to generalized reasoning through world models, but it’s not impossible. To make their vision a reality, Katanforoosh suggests, Runway must continue gathering resources, particularly computing power.

Runway has agreements with CoreWeave and Nvidia but hasn’t confirmed if it has dedicated cluster access, crucial for training frontier models.

“How can you build a foundational model without a cluster?” Katanforoosh questioned. “I don’t think anyone can.”

Runway has raised $860 million so far, including $315 million in February from partners like AMD Ventures and Nvidia. This is similar to what competitors Luma AI and World Labs have raised, according to PitchBook.

However, Runway faces formidable opponents like OpenAI, which has raised approximately $175 billion according to CEO Sam Altman, and Google, with its parent company Alphabet valued at $4.86 trillion. Google’s Veo model directly competes with Runway’s video generation, and its Genie world model targets the same long-term objectives Runway pursues.

Katanforoosh pointed to OpenAI’s video platform Sora, which shut down in March due to high compute costs and low revenue. His argument: resources alone don’t guarantee survival, nor do they assure Runway’s future.

Katanforoosh remains optimistic about Runway. He cited AI audio startup ElevenLabs, which outperformed OpenAI and Google on their benchmarks despite limited resources and prestige. Runway could adopt a similar strategy, he suggests.

Runway’s founders acknowledge this comparison. Valenzuela asserts the startup’s lack of Bay Area conformity gives them an advantage. Their diverse perspectives and the necessity to be resourceful, without Silicon Valley’s financial buffer, drove them to generate revenue early.

Michelle Kwon, Runway’s chief operating officer, stated that the company is not in a hurry to seek more funding, even as their computing needs grow.

“Their backgrounds have led them to be early, often correct, and foster a rapidly evolving culture,” Michael Dempsey, an early investor and managing partner at Compound, told JS.

For Valenzuela, this culture is rooted in his worldview. In his limited spare time, as a co-CEO and new father, he reads books, including works by Chilean poet Nicanor Parra, whom he regards as less formal and academic than Pablo Neruda, believing poetry should belong to the people, not adhere to rules.

“Rules are just invented,” Valenzuela stated. “That’s the driving force behind Runway. They say Silicon Valley is where startups are. Why? Those are just made-up rules. Erase them all and start anew.”