Roblox CEO David Baszucki on 150M daily users, a $1B creator economy, and the path to 10% of gaming

are live for this one. Um but since it is your first time on the show, I would love for you to uh just give us a little bit of an update on where the business is. Uh what's the newest? What's the latest and greatest in your world? Hey, so first off, thanks for having me on. I am a huge fan.

You're reinventing really the future of business news. So, I I love your multimodal streaming strategy and it was great seeing you guys two days ago.

You know, every every day about 150 million people now come to Roblox and they play, they communicate, they hang out not um in games made by Roblox, but literally in experiences made by a huge creator community. So, everything on Roblox is built by the community.

And we've shared about a year ago um we believe Roblox will get to 10% of all gaming running on the platform. We're at about 3% right now. Just came off a great earnings call last week. Our year-on-year bookings is growing 70% and our year-on-year DAUs are growing 70%. Wow.

So, I mean I I I asked you this before, I'm gonna ask you again because it's a funny question. Uh it feels like you have built such a powerful platform. The business is running so smoothly. You have this amazing cohort of particularly young gamers who are going to grow up and probably stay with the platform forever.

Uh can you just do nothing and win? Like if you just if you just keep the servers online and keep the platform flexible, it's not like you have to come up with the next greatest piece of intellectual property because you are the platform.

So if if I if I misunderstand that, what what do you have to do daytoday to actually move the ball down the field to get to that 10% that your goal is?

I like the way you think of that and we, you know, for those of us that are mathematicians out there, we sometimes think of the second and third derivative, not just the first derivative. And and what you shared is exactly right. If we walked away, just kept the lights on. There's enormous native growth in the platform.

This is a a highly technical platform. Creators make experiences. They auto translate. They run anywhere in the world. We've now seen peak concurrency records for gaming all around the world. In August, Grow a Garden, one of the more recent big games on Roblox, hit 25 million concurrence.

So I I would say we cannot walk away. We need to support infra and support scale. But if we did walk away, I I I think there may be some native growth that keeps going. What we think about when we're running the company though is the exact opposite of that. We think about are we constantly layering in systems?

Are we building a machine that can keep growing? And uh we building a machine that as we hit 10% of the gaming space, we're we're growing right by that. And and the gaming, you know, the gaming market's pretty interesting. It's about 180 to 200 billion.

It's arguably really um sitting right there for technical innovation, for new ways of distributing, for games that run really well on low-end Android phones and on high-end PCs, for games that incorporate AI uh within the experience. So, we just think there's enormous headroom in this market.

What's exciting to you about AI as a creative tool or or or a tool that enables people to build games faster? And what's most exciting to you about AI that can be integrated into a gaming experience? Because I feel like they're kind of two distinct buckets.

I I think there's several buckets behind the scenes over the last four to five years. We've built out a real substantial offering of models we've built ourselves.

We have over 400 models uh teams that are building everything from cutting edge voice and text safety models, teams that are building models that auto translate, teams that are powering our search and discovery models.

We are starting and and very shortly we're going to offer what we call inexperience 4D generation which is it literally the ability to make a game that you or I are playing in and we can prompt and ask for some type of vehicle for example or some kind type of dwelling and literally build it as we're playing.

So, it's a it's a little bit um of a wide open field here as far as what what types of games are people going to build when they have unlimited AI not just for the creators but for everyone who's playing in that game.

And and I do think the future spec which we think about which is an enormous technical spec is can we host one to 100,000 people together? Can they be either photorealistic or cartoony? Whatever the developer wants. Can they all be talking, listening, interacting with each other?

And can they dynamically change the environment? Can one or two or many of them essentially be dungeon masters and and changing the environment dynamically with AI? That's a huge technical lift, but it's it's what we're working on.

How much how do you so so when I think of the Roblox business and and you know people are very excited about AI and gaming uh I view uh not not to kind of gas you and the team up too much but I view you guys as having the kind of communication layer the user base the distribution it gives you such an insane advantage because if I'm somebody who wants to create a game I can go to I can create a standalone game I could go to another platform or things like that or I could go to Roblox where I know there's hundreds of millions of people that want to play games.

I I how uh yeah talk about uh yeah talk like do you think there's a what percentage of game when you when you talk about like owning 10% of the gaming market? Do you think do you think game developers are in of the future will just be kind of defaulting to Roblox?

Is that a goal of like if I want to build a game, it's like why why go through the effort of like trying to build this build my own kind of like community and player base when I can go to where an existing player base is and have a much higher likelihood of of success. Yeah, I think we don't rest on our laurels.

Um, obviously, but I do think the gaming market is a bit different than, for example, the video market. And in the video market, whether it's short, um, UGC, you know, Tik Tok, YouTube, Netflix, you name it. That's relatively mature technology. The the cameras are relatively mature.

Video distribution, playback is relatively mature. But the the spec I shared with you, 100,000 people in a you know multiplayer immersive uh world with AI is far from technically figured out. And that kind of goes to our vision that the next future of gaming is much more client server cloud integrated.

It's why Roblox has many many data centers supporting this. It's why behind the scenes it's not just the game engine. It's the persistence engine and the analytic system and the e economy system and as you correctly noted the social graph, a friend graph, where are your friends? Can you find them?

As well as the client system, which is does a game run on iOS, Android, PC, Mac, Xbox, PlayStation, Quest, and more to come. Um, I I would say there's one other angle that we really focus on, which is the economics.

the more efficiently we can build this whole platform and the more money that can flow back to the creators, the more that combination of technology and discovery starts to make a lot of sense. And so we do focus on running a very efficient company that that pushes money back to the creator community.

Well over a billion dollars will flow back to Roblox creators this year. It it reminds me of Substack.

Like there was a time when you if you wanted to write online, you started by spinning up a blog website like writing the code or or at least forking the code or or managing your own server and making sure your website kept like like stayed up and now uh you can just build an actual game in Roblox.

And I I understand that it's early, but that does seem like where it's What are the keys to uh managing an economy?

because I feel like you have the unique challenge of managing you're running you're running and building roadblocks the business and the platform and the infrastructure and then there's the then there's the economy of you have a chief economist like a whole economics team like how how serious is that discipline it's really serious uh we have a lot of PhDs um we have great econ brains a lot of times kids will say hey why can't you just give me a trillion Robux and it's a great opportunity to tell kids that if we give you a trillion Robux, the average Robu will be worth about a billionth of what it used to be worth.

Y So we we actually know about money supply. We know we can't print currency. Are you the Jerome Powell of the Roblox ecosystem? You get DMs directly. People ask you to print to turn on the money printer. Uh well, imagine I'm a young kid. Why can't you just make a trillion Robux?

and and so there's a lot of learning there that the value of the ecosystem doesn't change when you print more currency. The currency is just worth less. So we're very careful with the value of our currency.

We um we've done something really right I think from the start and there's a real temptation to do uh nonlinear functions accelerators. uh if a dev gets really big, should we pay them, you know, more per hour, per minute, per robs? We've really stayed away from all those nonlinear systems.

And I think that's added a a fairness to it. We we I do believe we have aligned incentives of creators and users in a good way so that if the system is working, it keeps working year after year.

Is there any world where there there's an independent Roblox Fed uh either through crypto or some sort of spinning out or some sort of nonprofit or some sort of you know advisory group where like the monetary policy of the Roblox ecosystem is not set by anyone with uh Roblox stock. Sounds like a nightmare.

It sounds like a nightmare. I'm not advocating for this. I'm just wondering. Have you done the thought experiment? Yeah, absolutely.

I mean there's [laughter] there there are some fun thought experiments around if I'm holding Roblox currency is it backed by a digital vehicle and is that digital vehicle stable and backed by a dollar.

So there are you know as we get bigger and people start to hold larger amounts of money and we we have both Robux and fiat currency it is worth thinking what gives confidence to consumers. Yeah. No, that makes a lot of sense. How how early can you tell that a game is a is a home run? Like a hit? Oh, interesting.

Like is there like within like let's say like 10,000 people try a new game. Is that enough? Can you get enough of a read from that to understand that that that game could someday have a million concurrents? I I have both two answers to that.

One answer I would say is a more of a some companies work intuitively and they design great stuff. Other companies work hyperanalytically and they they have systems and whatever.

On the intuitive front, a couple years ago, I was I was just playing on Roblox and there were about 800 people in a experience called Dress to Impress, which was uh it was like a fashion game with their own avatars, their own clothing style, uh competitively put on your clothing and compete in a runway show.

And it it felt really good and I I forwarded it to the the team just because there's a lot of attempts to really get fashion good. Simultaneously, we have a great team of people who work on search and discovery.

And our our our general philosophy is let's make search and discovery as predictable as possible, as transparent as possible, and as aligned as possible with both the users trying to find cool new stuff, as well as the creators, which is bubbling up great new stuff on the platform.

Dress to impress got picked up by that algorithm and got forwarded a bit and that helped drive the speed of the viral success. So both I would say the users themselves share with word of mouth as well as our algorithms surface great content.

It's why I believe in the last year we had seven um experiences hitting 10 million daily active users at some point in Q3 and I believe five of them were new in the last 12 months. How do you think about the layers of the Roblox economy?

And I'd love for you to explain just a cut deeper maybe on the grow a garden uh example, but um are there are there creators or do you envision in the future creators that are would would identify not as a Roblox creator but as a grow a garden creator and you're basically creating like a subeconomy within the overall economy because the tools of game design are so robust that there's exchange rates like how do you think about the layers of the onion?

on the platform that you're building. I I like how you think about this layer thing. In a way, gaming is very different than video in that on a platform like Roblox, you can create an experience on Roblox, so you're a creator and that experience can have creators within the experience. And so there are two layers there.

There's a little bit of a long-term futurep proofing in that there there are more and more experiences where you know they started initially as experiences like lumberjack tycoon where you would build a lumber mill and you would be a creator in lumberjack tycoon and compete as a creator within that experience.

I think where this goes is there are quote games on Roblox that have individual creators within them that may actually make money being a creator within an environment made by someone on Roblox.

That's a really good long-term, you know, essentially dual layer strategy that gives us a lot of capability to support a wide range of creation styles. We've seen that with some experiences like Clipit, for example, was an experience on Roblox where you make short little 3D vignettes.

And there are actually creators in that experience itself. Sorry, before we move on, I have two wild card questions. Uh, first, everyone in the chat wants to know because we saw a lot of people yesterday with lots of fish in the background. What's the biggest fish you've ever caught? Oh my gosh, you may be talking.

I think it's um there's an experience that's gone live really big in Indonesia. Yes. And we're really happy about the notion of more experiences um being different in different countries rather than one experience being hot anywhere. So fish it literally uh we were thinking about why is it big in Indonesia?

We were talking in our executive meeting on, you know, Tuesday. Maybe to go big in North America, you need like a giant bass boat with a big motor. Cultural. Yeah. Yeah. Where whereas Indonesia, it's a little bit more traditional fishing. What about personally? Have you ever been fishing?

Have you ever caught a big fish personally? I I have been in fish it, but I've never caught a big fish. What about in the real world? In a big world? No. like like like not in Roblox at all, but in the like on a vacation, have you ever in in the physical world caught a fish?

In the physical world, I grew up in Minnesota and I caught one day a bunch of small little sunfish sitting in a small little boat on a lake. There we go. Okay, that's something. Uh [clears throat] second second question, uh how have you engaged with do you believe in the simulation hypothesis?

Are you familiar with this? Right. Are we So, so are we in a simulation? If we are in a if we are in a simulation, are we in a simulation within a simulation? I I have two different views of it. Uh view one I would sure love the answer. Can you give me the answer please?

Because knowing the answer to this is probably like that that is essentially the answer to the creation of the universe. Um depending on what religion you are, what is God? like that is an answer to about a million things. So I would love the answer.

The other one is it's a little unfathomable to me but we're thinking about because a computer that could simulate the universe we are in starts to be a computer that is you know 100 to the 100 to the hundth power of processing and all of those types of things and that just blows my mind.

Anyway, so if we are in a simulation, the next layer universe that can support this simulation is pretty crazy. [snorts] Yeah. Yeah. It's a it's a fascinating thought experiment. Absolutely wild. I I wouldn't be shocked if Yeah. this conversation that we're having right now is actually happening in Roblox. Yes.

And some dimension. And it's just a it's a simulation itself on something. Actually, in the next layer universe, Roblox has been very successful. Yeah. And we don't even know it, but we are in the next layers Roblox hanging out right now. I believe it. I believe it. On something way more tractable.

Uh vibe coding for Roblox games. Uh is that a separate company? Is that features that you'll build? Does this already exist? Or are is is like the ability to generate a website or the ability to like vibe code generally is just remarkable.

It's lowering the barrier to entry on designing uh mobile apps and designing websites. Uh I imagine that the floor is going to go even lower on what it takes to build a game in Roblox. What are you seeing?

I I think uh vibe coding is maybe an intermediate step to what we would call real-time 3D or fourdimensional dreaming with other players which is literally you and I are walking around an environment talking or doing things and as we're chatting over there a 3D experience or game or object appears.

So, I think I think there's something even further than vibe creating a Roblox experience, which is how do we support um in a Roblox experience a command line where you can basically ask for anything with a friend and you'll see it pop up whether it's a game or an object or a piece of clothing.

Jordy, uh we can talk about AI buildout. I'm always fascinated by that. Yeah, I wonder. Yeah, I guess is that affecting you at all?

I mean, we talked a little bit about someone as someone who's developed and manages data centers, like how have you been processing the broader AI buildout and some of the timelines that some of these projects are happening?

Yeah, I I would say you know the good news is I feel for Roblox very early on we started with a build our own data center philosophy both core and edge data centers. We have data centers all over the world in Brazil, in Singapore, in Poland.

And when we hit these big numbers like 45 million concurrency, um we're being able to do that at a reasonable cost because we've done that. We um we are building out our GPU fleet as well.

We already, as I said, have 400 models running and we'll do that in the exact same thing we do with our our core and edge data centers. run the majority of the time on our own bare metal, complement it with burst at peak times to give essentially the optimal cost and reliability.

Okay, maybe take me through like the first big I'm going to buy a data center moment. What was that like emotionally? Were you using debt for the first time in the company? Were you nervous or was it just like yeah, of course I'm financing it. It's like I'm buying a house.

like psychologically as a CEO, as a steward of capital, there's now someone coming in potentially higher in the capital stack. How'd you finance it? What were you thinking? The the good news is we generate a fair amount of cash and we've always generated cash. So, paid for it in cash. He bought his house in cash.

So, we actually tend to buy a lot of things with cash. I think if you look at our balance sheet, you'll see billions and billions of cash and a billion of uh debt we raised a few years ago. So, so far we haven't gone heavily down that financing route.

Uh we build data centers and they last for 3, four, five years, whatever the depreciation cycle is. We're going to continue to do that. Um and I I don't think you know we're getting you're having to borrow a lot of the data centers we build. uh we buy capacity and then put our own hardware and bandwidth in together.

Uh we'll continue that. We tend to build in many many locations rather than all as at the same place. So so far so good. When when uh we've seen CEOs out there planning for six or sevenyear depreciation schedules, you said three, four, five, do you think they're potentially being a little too generous?

not not to not singling out any any specific company, but I'm curious how you're planning with your GPU fleet around depreciation. Yeah, I know we um we optimize the length and don't quote me, we should look up exactly what our depreciation is for CPU fleets. Yeah, I do think CPUs are generally aging better than GPUs.

Um, so I think it's worth taking a look at uh the speed of GPU development versus CPU development and we may see those things end up in slightly different places. Yeah, a lot of that just comes from if if CPUs are increasing 2% a year and GPUs are increasing 20% a year. That's exactly right.

You're just going to be way way way it's worth thinking about for $1,000 what kind of GPU I'm going to buy in seven years. There's a chance it's it's faster. Yep. like [laughter] just a chance small just a small chance. Wow. Wow. Hot take of the century. I love that. Thank you.

Uh how much internally with the team like what what are like how how are you how would you kind of describe some of the messaging? I mean we're in such an insane market right now. Roblox went public in 2021 during the the the peak of Zerp. Had to navigate through uh quite a lot over the last few years.

And now we're in it feels like an even more volatile, unpredictable market with the team.

Are you is is or the conversations like hey let's kind of zoom out and kind of look at you know you started the company in in 2004 so you know every uh you've seen a few cycles by now but uh how how how have you been leading uh through this through this year?

I I think earlier what I said is um you know we couldn't predict if all we did was keep the lights on how what the growth curve is.

I do feel we treat the business as if we have to be innovating at maximum velocity constantly and we we can imagine this spec and we c we can imagine what our product would be like with some of the things I mentioned earlier.

Um, and we do also think there's a future for for massive innovation in 3D communication relative to the gaming market and beyond. So, we we have our heads down uh building stuff, shipping stuff all the time. We focus a lot on how we've organized the company.

We we have it's almost as if Roblox is eight or nine companies within a company given the breadth of our tech stack. I mean we are an economy, an AI company, an infrastructure company, a game engine in the cloud company, uh search and discovery, user social, graph company.

So we have great leaders across that and in each one we have enormous innovations we're trying to bring to market. Fantastic. Uh well, thank you so much for coming on the show. Great to hang out. Congratulations to the whole team. Congratulations and congratulations.

Congratulations to you and the model and I love what you're doing. Yeah, we'll talk to you soon. Have a great rest of your day. Talk soon. Goodbye. Bye, man. Before we bring in our next guest, let me tell you about profound. Get your brand mentioned in Shot TV.

Reach millions of consumers who are using AI to discover new products and brands. Uh we have Vlad Tenn from Robin Hood in the reream waiting room. Let's bring him in to the TVPN Ultradome. Vlad, how are you doing today? We'll bring him in in just a minute. Um what else is going on in the timeline?

[music] And Dre Horowitz announced an A16Z new media fellowship. Uh Eric Torberg has this media as many no new not neo these are different things.

We described this we worked through this or as many people are saying the teal fellowship for the terminally online finding highly functional people who are also very online is hard. So we're creating a fellowship for them. Uh Tyler had a hot take on it. What would you say?

What was your what was your new media fellowship question? Uh it was like the riff on the teal question like who are who are your mutuals that are like no one else has mutuals with or something like that. Uh anyway, uh congratulations to everyone over at A16Z on the new media fellowship.

Uh and let's bring in Vladiff from Robin Hood. How you doing Vlad? Good to see you John. [laughter] How you doing Jordy? This is amazing. Good to see you. You always you always impress with the with the the camera quality is fantasticity. I also love it cuz last time you were a little bit further away from the camera