Real time is one of the things that is happening, that we are working hard on, releasing models that work in real time in partnership with Nvidia, so that one can iterate in real time with these models and it can be applied to the creative area as well as to new media, to video games, to interactive experiences and to robotics. And physical AI is something that again is coming at us, and we haven't seen as much investment in robotics as what's happening nowadays. Nowadays, every day a new company doing robotics is announced, and as we start seeing that these models, this brain layer of robotics, begins to work, what we're going to see is a verticalization of robotics in every possible industry it can be applied to. So that's what I think is going to happen from here on, not only in 2027 but from 2017 onwards.
The government of Chile, I believe, invests a couple of tens of millions of dollars a year in digital infrastructure, in cables. We have a connection cable with Asia, for example, to connect the country, right? And nowadays it's very obvious, it's like, of course we have to connect the country. How much money do we have to invest? Hopefully we invest more. And that's a need of every country in the world. I believe that over the next 12 years, many countries in the world are going to start investing very heavily in sovereignty, where you'll see quite large budgets and numbers from countries making investments they've never made before, like what we talked about earlier, that this is a fundamental strategic position for them.
Welcome to IALAF, a podcast from Diario Financiero where we approach artificial intelligence from different angles and how this technology is affecting our world, our companies and even the government from different fields. Today we have two luxury guests. Joining us from New York are Cristóbal Valenzuela and Alejandro Matamala, two of the three founders of Runway. Very welcome and thank you so much for being here.
Thank you very much for the invitation.
Yes, thanks for the invitation.
Well, for those who don't know what Runway is, I'll give a brief definition. They're an applied artificial intelligence research startup and multimedia technology company specializing in the generation and editing of videos, images and audiovisual content with multiple use cases across different industries. Well, perfect. [laughter]
Throughout their history, Runway, which started in 2018, has raised tens, hundreds of millions of dollars and today they're valued at more than 5 billion dollars. And it's worth saying that they're one of the main, I don't know whether to say, artificial intelligence laboratories or artificial intelligence companies in the world. They're based in New York. And anyway, thank you again for being here. I want to talk about several things, about the progress you've had these years at Runway. Robotics and physical AI I think are also a great angle to approach, but first I want to know the reasons for your visit here in Chile, an express visit. There, Cristóbal, you made some spoiler on X, you had meetings with the government, a photo at La Moneda, and you brought two Chileans back to their country. What are you going to be doing during these days?
Yes, do you want to tell a bit?
Well, we're growing, we're in a growth stage at the company, an internationalization stage. Recently we opened an office in London, a commercial office where we also cover Europe, the Middle East and Africa. And an office in Paris, which is more of a research center. Paris and London are two very important hubs in artificial intelligence research, and because Europe is our second market. And recently Cristóbal was in Tokyo because it's our third market, because we have several investors from there and because robotics is very strong there and it makes a lot of sense for us to have a presence.
I was recently in Brazil, in São Paulo and Rio, exploring Latin America. Brazil and Korea are the markets that come after the ones I told you about, and we're seeing where to position ourselves, where to establish ourselves in Latin America. But being there, we said, why don't we think about home? Why don't we explore what the possibility of doing something here would be? And we started having some conversations, and this week we're here because we have the idea of coming to establish Runway, a more physical presence with a commercial arm, but also with a research and development arm. And that's why we're meeting, we met with the government, with the Minister of Science, with the Presidency, to discuss how we can create an advanced research laboratory with UC Santiago and in line with what we've been doing.
That's great. And what was the reception like from the government, from the authorities?
Yes, there's a lot of interest. We met with several ministers, we're continuing conversations with CORFO and with other institutions, and there's interest above all because we're from here, because we're Chilean, and because we're working on the cutting edge. And we believe there's a very interesting window to do research and development here in Chile, and above all because there are very interesting specific industries like mining, like agriculture, where we can deploy this technology in the future of robotics.
Well, let's talk a bit about your development off-mic, Cristóbal. We talked about how we met, I interviewed you for the first time around September 2024, which is an eternity in the times of artificial intelligence. I remember that after the interview you went to another interview there on one of the panels, and you were showing in one of your presentations the progress it had had since the first model you released, I imagine it was Gen 1.0, I don't know what the name was, up to the latest model you had at that time, which looked like something a bit intangible, that was gradually taking shape. How have these almost, how many? 8 years of Runway been in terms of development and in the stage you're in today?
They've been radical and fast and different. I think what we talk about in the company is that each year feels like a decade in a way, and when I told you that the last time we met, for me it felt like 20 years ago, that's a bit of a reflection of the speed at which I believe this technology will keep advancing.
One of the things I constantly realize when I talk to people, even now I just had a conversation with several advertising agencies here in Chile, is that we have collective short-term memory loss because we've forgotten that a year, two years ago all of this was impossible. All of this was a dream, doing the things that models do today, the capabilities of language models that have been able to solve mathematical problems that had been impossible to solve for the last hundreds of years. The models like ours that allow you to generate an impressive video quality. 10 years ago it would have been unthinkable. Any person 100 years later. And I think one of the things we've naturally become better at is understanding how to operate at that speed within a company, because that speed within our company feels, I mean, it's literally like a rocket going 1000 km per hour.
And I also think you feel it more because you're not a team, because one says, I don't know, compares you with the other labs, OpenAI, Anthropic have thousands maybe of employees and you guys I remember were always very small, like 100 something, right?
Yes, now we're growing. We'll probably end the year at 400 people, 350, 400 people, that's like the goal. So we're growing, we're still not thousands of people. But it's necessary as you grow, not only the team, to grow the culture, this culture that lets you know a bit that things can change next week and that you could have been working on something for the last 5 months and it doesn't matter because we made some discovery or something happened in the market that allowed us to do something different. And that adaptability becomes fundamental in the company.
Hmm. Hey, and what's something that, I don't know, a couple of years ago maybe you would have found very difficult and that thanks to these same advances you've had today becomes faster or more efficient? Maybe that also perhaps surprised you, the speed at which things move now.
Maybe we can talk about real time.
Yes, I think that's something we've been thinking about for a while now, which is being able to generate videos or outputs in real time. And it's something that solves many problems in the creative process, it solves the iteration problem, of how to explore different ideas. Nowadays, to give context, to make a feature film or an audiovisual piece, you have to work with these models, you have to wait three, five minutes for an output to come out, and that output maybe isn't the one you like. And you have to keep changing it, and that ends up being a process. We're doing all of this to eliminate tedious processes, to eliminate all those things, but with long processes it still becomes a bit difficult. But with real time you can iterate immediately on what you're thinking, what you're looking for, and it also opens new possibilities, new capabilities to use that technology in things that weren't possible before, like for example, one of the things we're doing is real-time video that allows you to change the frame immediately by giving it new information. And there you can start thinking that tomorrow there could be applications like, for example, you're watching a football match or something live and you can give this model to enhance that experience in some way according to the show you're watching or the thing. And also for new cases like video games, like interactive experiences, immersive experiences, and for everything that's coming with robotics that will need to simulate, imagine a certain scenario in real time.
Yes. One of the main or first industries you were aiming at, I remember you always talk about media and entertainment, and one of those was the film industry. How have you seen the evolution this industry has had in adopting and also accepting this technology? Do you see them more open? Are some still more reluctant? Because you already work with several Hollywood studios, for example.
Yes, I think we work with all the Hollywood studios and it's been a bit varied. I'd say two, three years ago there was a feeling of uncertainty. Basically an alien arrived to work at their company and they have no idea what it means, what it does, but it seems it can do whatever they want it to do. And during the last two years all the Hollywood companies have tried to understand this alien, they've tried to understand why there's a technology now that allows them to do things they never could have done at the speed it allows them to do. And I think the last two years have been a bit of understanding and comprehension of this technology within the studios, especially Hollywood's, what does this mean for the actors, for the musicians, for the directors, for the writers. And nowadays, I mean, we work with all the studios, I mean, we've reached the point where finally everyone understood that this alien wasn't an alien, but rather a quite powerful technology that allows you to do many things and that if you knew how to use it well, you could go very far. And I think now that same version and that same process has been happening in other parts of the world. I think in Europe too, several filmmakers and directors we talk to are going through a similar process and are also arriving at the same conclusions. A quite creative tool, quite important and quite necessary. And I think in Chile, from the interactions we've had, well, we've worked, we announced something with Fabula a couple of years ago that we're working closely with, I think they're also heading that way. I think it's a function and they're in direct correlation to how much experience you have with this technology and if you've experimented, if you've tried it, if you've used it, and if you've seen what it does and what it doesn't do, that helps you form an idea and overcome your fears. And I think Chile today is in a position where we've seen many more people trying, being much more open to discovering, trying and experimenting, and that's changing the way people in the media industry, in Hollywood, in film, in television have become much more open to using it.
And how are those studios using Runway's tools today? Because I remember it was editing and post-production, today that has expanded more and I don't know, you also said that an entire movie was going to be made.
Towards the use cases, I want to focus on that today. We've been saying this for a long time, I think today an entire movie could be generated by AI and that's already a fact, it's already a fact. The most constant use, I think Netflix was the one that said they had finally, had publicly announced that more than 300 subtitles had used AI, and there will be other similar companies that will start announcing and saying more about that. And I think two fundamental cases are in, as you said, pre-production, all the creative concept that comes before making the movie, the iteration, the art direction, and afterwards, of course, where we want to get into more fully.
You're getting fully into the final scene, right?
The possibility that someone can create something that only a Hollywood studio could create. If an artist or a filmmaker in Chile with perhaps not the same resources as Hollywood, today can compete with them in effects, in visual capabilities, not in creative capabilities, because creative capacity still remains independent for each of the creators, but in technical capacity we're basically leveling the playing field. And what has happened in the last 6 months has been explosive in the capabilities these models have improved in. We've gone from pre-production, from having a visualization idea to an idea to production and seeing things being broadcast on television, being scaled to larger visual productions. So that in the last 6 months has been what we call internally crossing the chasm, which is crossing this abyss where before it was like you had to wait for a final step because we didn't have the right capabilities. And what we've seen most is that the industry is establishing itself, but they're building an enormous creative industry worldwide. The clearest case for us is that we have this film festival where we celebrate projects made with artificial intelligence. We've been doing it for 4 years. The first year we received 300 short film applications from creators all over the world. This year we received 6,000 applications, and they're much more ambitious projects, much bigger products, of better quality, and from an industry being built all over the world.
Hey, and as Cris said, this is growing. In 2025 you signed this collaboration agreement with Fabula, the Chilean production company. Without going too much into detail, how are they using Runway today? I think at the minute you announced it, they had said there was a movie from a director who was using some of the tools, but I don't know if we can detail a bit what they're doing, for example.
Yes, a lot of that work comes more than anything in conceptualization, pre-visualization, creative concepts, in scene tests, camera tests. There's a lot. When you watch a movie, the big part of the work actually happens behind the scenes. I mean, you're watching 90 to 160 minutes of a heavily edited and heavily iterated piece, from a process that lasted three, four years. And the budget of a movie in general goes to that behind-the-scenes part that you never see. And if you think about a movie, for example, of 100, 200 million dollars, which is the standard for a Triple-A or high-quality movie in the United States, probably 80% of that goes simply to all the pre-production work. And so many of the parts where the best film production companies are using and knowing how to use this technology is precisely in that part. I mean, if I can reduce the cost or accelerate the speed at which you're iterating on that creative concept, on writing the script, on defining the characters, on doing camera tests, on doing visual effects, that's probably the most important part today, where many filmmakers, including part of Fabula's directors, are really getting into it. And then later on, for example, about two weeks ago we launched Gen-4.5, we launched a model, the first time it's launched and the first time this can be done, a model that allows you to transfer and transform the videos you generate within Runway into the standard quality for the big screen.
Which is very necessary. What's called HDR.
So, the format so you can watch, because if you put a high-quality camera, a very good camera, for example, a camera called Blackmagic, and then you generate a scene in Runway, a couple of months ago the quality of those two videos wouldn't have been at the same level. Today we've been able to create a new technique that allows you to take that generation made with Runway to the same level as a professional camera. So you can put a real camera, then the next scene, something generated, real camera, and the viewer will never notice. And that's going to be the change, one of the fundamental aspects in bringing this technology to a much more massive scale in terms of cinema.
How great, how great that is. Hey, you've also said that sometimes you yourselves are surprised by the use cases that Runway users create. I think it was Madonna who used it on her tour, like the images through the giant screens. What use cases did you not anticipate and are you seeing today with Runway?
Mind-blowing. There are several today. Today, I mean, we started working with the audiovisual world, with the film world, then came the advertising world, which is one of the biggest areas we have today in the company. We work with the biggest advertising companies in the world. But today the cases are, we have financial services companies, consulting services, e-commerce and retail, hospitality and tourism. Today any company that does corporate communication or marketing is using generative artificial intelligence. We have 80% of the Fortune 500 using it in some way. And any company that has an image catalog, that sees the opportunity to use it in an audiovisual asset, is seeing it. But I would say that the financial industry, banking and things like that, which are perhaps more conservative, we're seeing a super rapid adoption.
But that to promote their campaigns or from internal corporate communication to promoting campaigns?
We have an example we published recently on our site of a neobank called Chime, which is running a public television campaign using 100% generated with us.
