How Three Art Students Built the AI Tool Hollywood Couldn't Live Without

There's a version of this story where Cristóbal Valenzuela takes the job Adobe offered him in 2018.

He's a researcher at NYU, no formal AI background, studying how machine learning could change creative work. He builds something that catches Adobe's attention, they offer him a position on their AI team. It's a good offer. Safe. Credible. Adobe is the tool every creative professional already uses.

He says no.

Instead, he and two classmates incorporate Runway in December 2018 and spend the next two years getting rejected by investors who don't believe generative AI will matter.

"There's always doubt when you get 50 rejections," Valenzuela said later.

He kept going. And the company those 50 investors passed on is now worth $5.3 billion, used by every major film studio in the world, and pivoting toward something that makes video generation look like a warmup act.

This is how it happened.

📌 No sponsor this week - Projects of AI just relaunched. If your product belongs in front of 2,400+ AI founders and builders with a 69% open rate, grab a slot here →

The boy from Santiago who borrowed films from the library

Cristóbal Valenzuela was born in 1989 in Santiago, Chile. His route into technology was not the typical one.

He wasn't a programmer who discovered art. He was an artist who discovered programming.

As a teenager in Santiago, he made home videos with his friends, nothing serious, just the instinct to tell stories with a camera. He paid his way through college building websites and shooting promo videos for local businesses. In his free time, he borrowed films from the library and watched everything he could find.

He studied economics and business management at Adolfo Ibáñez University in Chile, and later added a master's in design. Not computer science. Not engineering. Design, economics, storytelling.

Then, in Chile, he came across a research project called Deep Dream, one of the earliest neural networks capable of generating images from patterns. The images it produced were strange and beautiful and nothing like what computers were supposed to be able to do.

That encounter changed everything.

Valenzuela was so captivated by what was happening in computer vision that he decided to move to New York to study it full-time at NYU, leaving a country and a career trajectory behind to chase something he found genuinely mind-bending.

He arrived at NYU's Tisch School of the Arts in 2016 with no deep learning experience, an economics degree, and a designer's eye. Most AI researchers would consider that a liability. Valenzuela considered it the point.

The NYU lab where everything started

At NYU, Valenzuela threw himself into research, specifically, how machine learning models could transform creative and artistic fields. Not AI for efficiency or automation. AI for imagination.

He researched applications of ML models for image and video segmentation for creative domains, conducting experiments such as integrating AI with Photoshop to streamline image editing tasks, gaining traction among creatives.

It was in this research context that he met the other two people who would build Runway with him:

• Anastasis Germanidis - a researcher from Greece who would become Runway's CTO. Technical depth, systems thinking, deep learning expertise.

• Alejandro Matamala - a designer and creative technologist from Chile who would become Chief Design Officer. The one who understood what creative professionals actually needed.

Three people. One from Chile via design and economics. One from Greece via computer science. One from Chile via creative technology.

None of them were the typical AI founder profile. All of them were obsessed with the same question: what happens when creative tools are rebuilt from scratch around machine intelligence?

A pivotal moment came when Adobe offered Valenzuela a position to continue his AI research work within their team in 2018.

He thought about it. Adobe is the gold standard for creative software. The offer was real validation. The safe choice was obvious.

He said no. He wanted to build the thing, not improve someone else's thing.

In December 2018, the three incorporated Runway.

Two years of nothing

Here's the part of the Runway story that gets skipped in the success narrative.

From December 2018 to roughly 2021, Runway existed but the world largely didn't care.

They were building AI tools for creative professionals before generative AI was a mainstream concept. Before DALL-E. Before Stable Diffusion. Before anyone outside of research labs talked about text-to-image or text-to-video as real things.

Investors weren't sold, not just on Runway's video AI startup, but on generative AI altogether.

Fifty rejections. The company kept building.

The product during this period was what they called "Magic Tools", a growing suite of AI-powered creative utilities. Remove background from video. Inpainting. Image-to-image style transfer. Colour grading. Tools that saved video editors hours of manual work per project. Runway grew its AI product pipeline to more than 30 AI Magic Tools by early 2023.

None of it was the breakthrough. All of it was building the muscle.

The moment everything changed

In 2022, two things happened that transformed Runway from a niche creative tool into something the entire industry was watching.

First: "Everything Everywhere All At Once."

The Academy Award Best Picture film "Everything Everywhere All At Once" used Runway's tools in its production, specifically Runway's green screen and visual processing technology that the filmmakers used in post-production. The film swept the 2023 Oscars, winning seven awards including Best Picture. Suddenly, the tool that 50 investors had rejected was part of an Oscar-winning production pipeline.

Word spread. Film editors talked to other film editors. The creative professional grapevine is tight and fast. Runway started appearing in conversations it had never been part of before.

Second: the generative AI moment.

Late 2022 brought DALL-E 2, Stable Diffusion, and ChatGPT in rapid succession. The entire world suddenly understood what generative AI could do. Runway had been building in this space for four years. When the mainstream arrived, Runway already had infrastructure, relationships, and a product that worked.

In June 2023, Runway announced a $141 million extension to its Series C at a $1.5 billion valuation, from investors including Google, NVIDIA, and Salesforce. The same investors who hadn't returned calls in 2019 were now competing to get in.

Gen-1, Gen-2, and the Hollywood moment

The product that put Runway on the global map wasn't the Magic Tools suite. It was the Gen models, Runway's text-to-video and image-to-video generation technology.

Gen-1 launched in early 2023. Gen-2 followed in June 2023. For the first time, a creative professional could type a description or provide a reference image and receive a short, coherent video clip in return. Not perfect. Not photorealistic. But directionally extraordinary showing what was possible in a way that made every filmmaker, every ad agency, every content studio sit up and pay attention.

Lionsgate signed a deal with Runway, with vice chairman Michael Burns telling the Wall Street Journal it would save the company "millions and millions of dollars" in production costs. "We do a lot of action movies, so we blow a lot of things up and that is one of the things Runway does," Burns said.

By 2026, Runway's customers included every major film studio, not just indie productions experimenting at the edges, but the studios responsible for the biggest theatrical releases in the world.

The art students from NYU had become Hollywood's infrastructure.

The Adobe moment, revisited

There's something worth sitting with in the Adobe decision.

When Valenzuela said no to Adobe in 2018, the company he was turning down was worth approximately $120 billion. Creative professionals used Adobe products for everything, Photoshop, Premiere, After Effects. Adobe was the creative software market.

By February 2026, Adobe Ventures was a participant in Runway's $315 million Series E funding round.

Adobe went from offering Valenzuela a job to investing in his company.

The thing he built by saying no became something Adobe needed to invest in. That's the complete arc in one sentence.

What Runway is building next, and why it's bigger than video

Runway is pivoting from filmmaking tools to world models, AI systems that learn from observational data rather than language.

This deserves a moment of context because it's not a small shift.

A world model doesn't generate video from text. It builds an internal representation of how physical reality works, cause and effect, physics, motion, spatial relationships and uses that representation to simulate environments. Runway released its first world model in December 2025 and views the technology as central to tackling major challenges across fields like medicine, climate, energy, and robotics.

The company that started with a designer from Chile asking "what if AI could help artists make better videos" is now building systems that could simulate drug interactions, climate scenarios, and physical environments.

That's not a pivot. That's an evolution. The creative intuition that made Runway good at video the understanding that the world has physics, that motion has logic, that scenes have coherence turns out to be exactly the right foundation for building systems that model reality.

In February 2026, Runway raised $315 million in a Series E at a $5.3 billion valuation, led by General Atlantic with participation from NVIDIA, Adobe Ventures, AMD Ventures, and Fidelity. Total raised: over $1 billion across all rounds.

By October 2025 Runway had reached approximately $300 million in annualized revenue, the largest Western AI video platform by revenue.

516 people. $5.3 billion. Every major film studio. And now: world models.

What I keep coming back to

Valenzuela spent his childhood borrowing films from a library in Santiago and making home videos with his friends. He studied economics and design, not computer science. He moved to New York specifically to chase something he found beautiful, the intersection of machine intelligence and creativity.

He wasn't building toward an exit. He was building toward a vision: what creative work looks like when AI is not a tool but a collaborator.

The 50 rejections didn't change that vision. Adobe's job offer didn't tempt it. The Oscar win validated it. The Hollywood deals confirmed it.

Valenzuela puts it this way: "We shape our tools and thereafter our tools shape us.”

That's not a founder talking about product-market fit. That's an artist talking about what he believes.

That's the Runway origin story. On Friday I'm going inside the product, the Gen model architecture, why Runway won Hollywood while Sora struggled, and what every creator and builder can steal from how they grew. See you then.

Hit reply - have you used Runway? What did you make? I read every reply.

- Projects of AI

P.S. - Adobe offered Cristóbal Valenzuela a job in 2018. He said no and built a $5.3 billion company. Adobe Ventures invested in his Series E in 2026. The best validation of a decision is when the person you turned down invests in what you built instead.

Enjoying this? Forward it to one person building in AI. That's the only growth strategy I have.

Want to reach 2400+ AI founders and builders? Sponsor Projects of AI