How to Build $1.5B AI Startup in Just 3 Years

In this conversation, we sit down with co-founders of fal, Burkay Gur and Gorkem Yurtseven. fal is a generative media platform. They just raised their Series C round of $125M, which values them at a $1.5B valuation. In 2020, during the COVID bubble in Palm Springs, two Turkish engineers – one from Amazon, one from Coinbase – started exploring startup ideas together. Starting with just image models when everyone said the market was too small, they stuck to their vision while competitors chased LLMs. They foresaw the video revolution coming. They obsessed over speed, becoming #1 on every benchmark. From two Turkish immigrants to a billion-dollar AI unicorn – this is their story.

Burkay & Gorkem: Hello we are Burkay and Gorkem, and we are co-founders of fal. Fal is a generative media platform for developers. We host models that can generate images, videos, 3D, and audio. Typically these models are very hard to host, so the problem we solve is hosting these models as APIs, which makes them very easy to consume for developers. We also have an inference engine that we built in-house, specifically optimized to run diffusion models two to three times better.

Latency kills creativity, latency kills productivity. We work with customers like Adobe, Canva, Shopify, and Perplexity. We're at a 90 million annualized run rate revenue.

We're fast at everything we do. We put out the models, and usually we have day-zero releases. The space is moving so fast, we have to be very ahead of getting these models in front of people and making it easy to use.

There's a lot of startups out there that will be stuck on an idea for months and years with no traction. You have to really take that to the extreme. I don't think people stress that enough: think of moving fast, take that, and multiply it by a hundred.

We just raised our Series C round, 125 million, which values us at a 1.5 billion valuation. We're very prepared, we're prepared to scale.

Burkay: My co-founder and I have been long-term friends. We're actually both from Turkey. I grew up in Turkey and moved to the States for college.

There was definitely culture shock. I think even a decade makes a big difference here. I moved to the States in 2007, right when Facebook had just come out. It definitely felt like I wasn't as tapped into the culture. There's a big gap between how I grew up in Turkey and what people like to do there versus how they are in the US.

Gorkem: I would say schoolwork felt a little bit easier than I thought, because we have a pretty good education system in high school in Turkey, especially with math and sciences. The biggest challenge was actually understanding the job market, how people do internships. Immediately, people start school and prepare for their summer internship, then the next summer, and they have a whole plan for how their career is going to happen. I didn't know I should be doing that. So for a couple of years I wasn't really planning my internships toward my career. That was a big shock.

Burkay: I actually did an internship at Oracle during college. I was working on some fairly boring things in the beginning, to be honest, and I had started my green card process. This is a very typical thing for immigrants in the US. You could kind of be stuck in jobs if you start your green card process.

Around 2015 was a very interesting time. Deep learning was just starting to become popular, and I started getting really into it. Around that time I had a few other friends at Coinbase, and Coinbase was a very small company back then, maybe 40 or 50 people. One of my friends told me, "Hey, we're building a machine learning team." I thought, this sounds very interesting, I can go do some deep learning at this new company and there's a lot I can learn there. But I was mainly excited about starting my own thing. I had talked to a lot of my founder friends about their experience, and I had a lot of encouragement from friends to actually go and start my own thing.

Gorkem: In the beginning of COVID, Burkay and I rented a house in Palm Springs for a while. We were talking about potentially starting a company, but we didn't have a particular angle or idea to go after. We knew that if we wanted to do this, we would have to go through a period of exploration where we found something we were both passionate about. Burkay quit maybe four months before me, and then I joined him.

It's liberating, because all my life I also had to deal with immigration, work visas, and then a green card. That's one of the reasons I stayed working at a big company, I wouldn't say that's the only reason, but it was definitely a factor. By that time, all my immigration process had ended as well. That was also liberating, in the sense that I didn't have to work for a big tech company to stay in the country. I could do whatever I wanted, and I took the opportunity then.

Gorkem: Starting with the post-ChatGPT era, it was brand new to everybody. It was such a new environment that nobody knew where things were going. We started running image workloads and saw tremendous growth in the companies working with us. That made us really excited about the space. We also sat down and thought about where this could be going.

Two and a half years ago, people saw LLMs and ChatGPT, and they immediately drew out where this technology was going: we're going to AGI. We felt similarly about image models. We thought that as the models got better, there would be more capabilities: quality would increase, resolutions would increase, and controllability would increase.

I think finding a niche market that is fast-growing is the key to startup success. There are a lot of niche markets that stay niche and never grow. But we were lucky that the market we operated in was very niche and small, but also growing incredibly fast.

What changed after these big models were released is that you didn't have to train it anymore, you could just pick it off the shelf and start building something useful. That meant the number of users maybe 10xed, 100xed, maybe over a millionx. We saw this change early on and decided, okay, this changes everything. Now that these models are going to be used by millions of people, we have to build systems that are ready for this change. That's why we decided to build an inference platform early on.

Another decision we had to make was when revenue was constant for a couple of months. One tempting thing we could have done was run inference for LLM models as well. But focusing on image and video is going to be an important differentiator. We already have a technical advantage, because we've been working on this type of model for a while. If we're wrong, we can always revisit our decision, but it's going to be harder for us to go from general to specific. So we tried to stay specific.

I think if you focus on a specific market, you get to work with your users in a closer manner, you understand their problems better. For us, this was image models and fine-tuning image models. In the beginning, all of our customers were doing very similar things, so we were able to focus on it, get really good at it, and differentiate ourselves from others.

Burkay: Our ultimate vision is basically that we want to be the infrastructure layer for this new technology. The ChatGPT moment for video, I don't think we've hit it yet. There are a lot of signs we're getting very close to it. If you've seen Veo 3, it's close to the moment. It's a very capable model, but I don't think we're there yet.

Interestingly, if you go to your Instagram or TikTok feed, a third to half of the videos are already AI-generated. It's already happening, it's just happening in a bit of a slow motion. There may be a point this year where we see even better models that can actually be edited in real time, where you can interact with the characters in the video and generate very interesting content. We want to be the place where all of this infrastructure is being hosted, and all the builders building with this technology, we want them to do that through fal.

Burkay: I think there are two things happening with AI. People are willing to pay, but there are questions about the quality of that revenue, or how durable it's going to be. AI markets are incredible markets. Generative media is one of those things that can be monetized right away.

In previous versions of internet businesses, people waited years and years to monetize, first building a user base and then maybe trying to monetize with subscriptions or ads. But with AI, people are willing to pay for it right away. The MVP you're building should be good enough for people to start paying, and it's really easy to get signs, whether the revenue numbers are increasing or not. Monetization should be a priority from day zero. It's actually easier for the founder to see if this is a good idea, or a good product, by the revenue they're making from the first day.

We're very particular about what models we want to put out, because there are a lot of models out there, a lot of research projects, even things that big funded companies put out, that are cherry-picked. Basically, you take the results, look at the good ones, and use those for your demo or your launch. That's called cherry-picking, and there's a lot of it happening with models.

When we look at a model, the first thing we do is take it, run it, and run a bunch of queries to understand whether it's actually doing the thing that's advertised. Then we go and optimize it, make sure it can run faster and faster, especially if there's a lot of demand. Developers spend so much time optimizing their iterative loop, making sure that once they do something, they can see the result, see the tests, and go iterate. No one wants to sit and wait around five minutes for a video to generate. In the future, this is going to be seconds, it's going to be real time, and we're preparing ourselves from an infrastructure standpoint for that future.

Burkay: Scaling the company has been one of the most exciting things about this job, to be honest. We were a very small team, like six people, for the first two years almost. I think small teams are very important before product-market fit. You actually do want to have the smallest team you can, experiment, have a small group making decisions and move really fast.

I think alignment with the company's mission is very important. This is something people talk about, and it's another thing I really learned from Coinbase. In Coinbase's early days, everyone was a crypto head, you would not find anybody who wasn't just insanely excited about crypto, and that created the foundation for the company.

It's just so specific that by default people are excited about what they're working on. But one of my criteria was that I had to love it, that was super important to me: the intersection of creativity and AI. There's unlimited fun there.

At least for me, I wake up every day very excited about the next models being released, where this technology is going, what amazing things people are building. It is literally the most fun thing I could be doing. If I wasn't doing this, I'd probably go play with these models myself. I love this technology, and that's the thing that gives me a lot of drive.