How to Build $1.5B AI Startup in Just 3 Years | fal's co-founders

From Zero to $1.5B: fal's Milestones

00:00Language models at the time was like all the hype, crazy optimism about language models. We felt similarly about image models. Yes, this is a niche place. Where could this be going? Finding a niche market that is fast growing is the key to startup success.

00:16After these big models were released is that the number of users 10x, 100x, maybe over a millionx, we we saw this this change early on. We have to build systems that are ready for this change. We're fast at everything we do. We put the models, usually we have day zero releases.

00:33Space is moving so fast. We have to be like very ahead of getting these models in front of people, making it easy to uh use. There's a lot of startups out there, you know, they will be stuck on an idea for like months and years, right? With no traction.

00:44You have to really take that to the extreme. Like I don't think people stress that enough. Think of moving fast. take that and multiply with like 100 and move that fast. If we are wrong, we can always revisit our decision. Focusing on image and video is going to be an important differentiator.

01:04We just raised our series C round 125 million which values us at a 1.5 billion valuation. We're very prepared, you know, we're prepared to scale. Chat GPT moment for video is that I don't think we've hit it yet. Right now, we we employed, you know, language models at massive scale.

01:18We're going to have to do that for AI video, AI image, AI audio, even AI games. And we want to be the place where like all the builders that are building with this technology, we want them to do that through FA. So my name is Burkai and I'm co-founder and CEO at FAL.

01:39FAL is a generative media platform for developers. We host models that can generate images, videos, 3D audio. Typically these models are very hard to host. So the problem we solve is like hosting these models as APIs which makes it very easy to consume for developers.

01:54We also have a inference engine that we built inhouse that is specifically optimized to run diffusion models to run two three times better. I think latency kills creativity, latency kills productivity. We work with customers like Adobe, Canva, Shopify, Perplexity.

02:12We are at uh 90 million annualized run rate revenue.

How Two Immigrants Without Connections Built a $1.5B Startup

02:20My co-founder and I have been long-term friends. We've actually we're both from Turkey. I grew up in Turkey. I moved to the States for college. There was definitely like culture shock. I think like even a decade makes a big difference here, right?

02:32Like I moved to the States 2007. I think Facebook had just come out. It definitely felt like I wasn't as tapped in to like the culture, right? So there's like a big gap between how I how I grew up in Turkey and like what people like to do there versus like how they're in the US.

02:47I would say like school work felt a little bit easier than than I thought because we have a pretty good education system in high school in Turkey especially with maths and sciences. the the the biggest challenge actually was like the understanding the job market, how people do internships because immediately people start school and prepare for their summer internship and then the next summer and like they have a whole plan on how their career is going to happen.

03:16I didn't know I should be doing that. So couple years I wasn't really planning my internships towards my career. So that was a big big shock. I actually did an internship at Oracle like during college. I was working on some like fairly boring things in the beginning to be honest and I had started my green card process.

03:34This is like a very typical thing for immigrants in the US. You could kind of like be stuck in jobs if you start your green card process. Around 2015 was a very interesting time. Deep learning was just like kind of starting to become popular.

03:49I 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. It's like maybe 40 50 people. One of my friends told me like, "Hey, we're building a machine learning team."

04:01And I was like, "Okay, this sounds very interesting. Like I can go like do some deep learning in this new company and there's a lot of things things I can learn there." But I was mainly excited about like starting my own thing. I had actually talked to a lot of my founder friends seeing their experience.

04:15I had a lot of encouragement from friends to actually go and start my own thing. In the beginning of co Burka and I rent a house in Palm Springs for a while. We were talking about potentially starting a company but we didn't have particular angle or idea to go after.

04:30So we knew that we want to do this we would have to go through period of exploration where we find something that we are both passionate about. Purai quit maybe 4 months before me and then I joined them. It is liberating because all my life I also had to deal with immigration work visa and then green card.

04:50That's one of the reasons actually I stayed at working at a big company. I wouldn't say that's the only reason but that's definitely a factor. And at 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.

05:08I could do whatever I want and I took the opportunity then.

Why We Bet on Gen AI Video Instead

05:17Starting with the posthatit era it was brand new to everybody. It was such a new environment that like nobody knew where things are going. We started running image workloads and we saw a tremendous growth in the companies that are working with us.

05:31That made us really excited about the space. We also sat down and thought like where could this be going? 2 and a half years ago people saw LLMs and and Chhatra PT and and they sort of like drew out where this technology go and you know they immediately said okay you know we're going to AGI.

05:46We felt similarly about image models. We thought like as the models get better, there's going to be more capabilities. Quality is going to increase and the resolutions are going to increase and the controllability is going to increase. I think finding a niche market that is fast growing is the key to startup success.

06:03There are a lot of niche markets that stay niche and never grow. But we were lucky that market we operated in was very niche and small but also was growing incredibly fast. What changed after these big models were released is that you didn't have to train it anymore.

06:20You can just pick it off the shelf and start building something useful. And that meant the number of users maybe 10xed, 100xed, maybe over a millionx. We saw this this change early on and we 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.

06:40And that's why we decided to build an inference platform early on. Another decision we had to make when the revenue was constant for a couple of months. One tempting thing we could have done run inference for LLM models as well. Focusing on imu and video is going to be an important differentiator.

06:58We already have a technical advantage because we've been working on on this type of models for a while. If we are 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.

07:14I 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 very similar things.

07:29So we were able to focus on it, get really good at it and differentiate ourselves from others. So our ultimate vision is basically we want to be the infrastructure layer for this new technology. Chat GPT moment for video is that I don't think we've hit it yet.

07:43I think there's a lot of signs like we're getting very close to it. Like if you've seen V3 it's close to the chat moment. It's a very capable model but I think I think we're still not there yet. But interestingly like you know now if you go to your Instagram Tik Tok feed like third half the the videos are AI generated right?

08:03it is already happening in a way. It's just happening in like a little bit of a slow motion. There may be a point this year is that we see like even better models that can actually like be edited uh real time and you can interact with the characters that are in the video and and you know generate very very interesting content.

08:19And we want to be the place where like all of this infrastructure is being hosted and all the builders that are building with this technology we want them to do that through fall.

Kill Your AI Product If It Doesn't Sell on Day1

08:34I 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 that revenue is going to be. I think AI markets are are incredible markets. Generative media is is one of those things where that it can be monetized right away.

08:51In the previous versions of internet businesses, people waited years and years to monetize their their business. First built a user base and then maybe try to monetize with subscription or ads. But with AI, people are willing to pay for it right away.

09:07The MVP you are building should be good enough for people to start paying. And it is really easy to get signs. the revenue numbers are increasing or not. Now monetization should be something a priority from day zero and it's actually easier for the founder to see if this is a good idea or if this is a good product by the revenue they are making from from the first day.

09:34Um we are very particular about what models we want to put because there's a lot of models out there. There's a lot of projects out there, research projects, even things that like big funded companies that put out there that are cherrypicked.

09:46Basically, like you take the results and you look at the good ones and you just use those for your demo or like for your launch. So that's called cherrypicking. So there's a lot of cherry picking happening in in models. When you when we look at the model, first thing we do is we take the model, we run the model and we run bunch of queries to understand like is it actually doing the thing that is advertised and then we we will go and optimize it and make sure like it can run faster and faster especially if there's a lot of demand.

10:12Developers like they spend so much time optimizing their like iterative loop, right? Like making sure that like once they do something they can see the result, they can see the tests and go iterate. So no one wants to like sit and wait around 5 minutes for a video to generate in the future.

10:29This is going to be seconds. It's going to be real time and we're we're preparing ourselves from infrastructure standpoint for that future.

Stay Small to Grow Big

10:40Scaling the company has been one of the like most exciting things about this job to be honest. We were like very small team like six people for the first two years almost. I think small teams is very important before product market fit. You actually do want to have like the smallest team that you can, right?

10:59And and and experiment and like have a small group making decisions and move really really fast. I think this like alignment with the company's mission is very important. This is something people talk about. This is another thing like I really learned from Coinbase.

11:11Like Coinbase early days like everyone was a crypto head. you would not find anybody that is not, you know, just insanely excited about crypto and and that created the foundation for the company that that is just so it's just so specific and so like, you know, just by default people are just excited about what they're working on, you know.

11:32But one of my criteria was that like I had to love I had to love it, you know. That was like super important to me. like the intersection of creativity and AI. I mean, there's like unlimited fun there. At least for me, I wake up every day, I'm very excited about like the next models that are released, what this tech where this technology is going, like what amazing things are people building.

11:56Uh it it is it is literally like the most fun thing I could be I could be doing. If I wasn't doing this, I would probably go play with these models myself. I I love this technology and and that's the thing that's like that that you know gives me a lot of drive [Music]