Roblox, the virtual world and game platform, has 66 million daily users who spent 14 billion collective hours on it in just Q1 of 2023. If you have kids, you’ve probably heard of Roblox. If you haven’t heard of it, Roblox CEO David Baszucki wants to change that.
One idea? Aging up the kinds of experiences that are allowed on its platform. Roblox recently introduced 17-plus experiences. It wants to add new AI world-building capabilities. It’s even partnering with advertisers to roll out more immersive ad experiences.
It’s been years since the number of adults gaming outnumbered kids — it seems like that’s driving a lot of growth for everyone, including Roblox. But these virtual world games seem like they all want to expand to be much more than just for kids and much more than just for games.
If you think about it, Roblox is already like a metaverse. Schools are using it for classes, companies are starting to advertise there, and people are just hanging out as avatars.
It’s already big, but the hope is to get much, much bigger.
Alex Heath, deputy editor at The Verge, got the chance to chat with David up at the Roblox headquarters in San Mateo, California. Their conversation covered a lot: why it’s now the time for Roblox to grow up, the classic Decoder questions about org charts and decision-making, and, sadly, why infinite Robux isn’t a thing. Apologies to all the eight-year-olds out there.
Also, you can watch the video of our conversation in this post or on YouTube!
Okay, Roblox CEO David Baszucki. Here we go.
This transcript has been lightly edited for clarity.
Alex Heath: David Baszucki, you’re the founder and CEO of Roblox. Welcome to Decoder.
David Baszucki: It’s great to be here. Thank you.
Roblox recently announced that, for the first time, it will allow 17 and up experiences. It’s a big shift for the platform. Big picture, why do this now?
Listen to Decoder, a show hosted by The Verge’s Nilay Patel about big ideas — and other problems. Subscribe here!
Actually, even five years ago, why do it? And probably five or six years ago, when we weren’t as big as we are today, we developed this theory [of] all ages around the world not just playing but ultimately learning or going to work or going to concerts or all of this stuff. So we developed this theory that there’s so many ways to grow, and one is all ages. And we said, as we typically do, what would be the hardest way to do this? The easy way might be to split the platform and have an older Roblox or a younger one. The hardest would be to be so good at safety and civility across the platform that we could start supporting those older, more mature experiences at the same time.
And that was the discussion. I’m so proud of what we’ve done now with 17 and up. We’re using photo ID, validating ages, kind of slowly and gently bringing this in, but I’m super optimistic that it supports that vision, that really, these types of platforms at Roblox are future communication and connection platforms. And they’re not just for six-year-olds. They’re not just for 14-year-olds. They’re for 60-year-olds. They’re for 35-year-olds. They’re really for everyone.
You guys have been pointing out for the last year or so that your user base is shifting increasingly older.
It’s almost natural. And the amount of usage we have, for example, in USA [ages] nine through 12 is pretty staggering. There’s still a lot of room for growth in USA, nine through 12, but historically, the opportunity we have in 17 and up, it’s massive, really, and there’s so many users out there. What has happened is what we’ve expected and hoped would happen is that as the quality of the product goes up, as search and discovery gets better, as developers make more interesting experiences, just like we grew organically with nine through 12, we would see that with 17 and up. And we’ve seen it 17-24 USA, 35 percent year on year. It’s really exciting.
I met someone, a 21-year-old member of our family, last week, and he’s like, “Oh, Roblox! All my 21-year-old friends are playing Roblox.”
I’m like, “What is up with that?”
I think they were playing DOORS and Murder Mystery. He knew the names of these experiences, and it was an early touch point for me of that type of behavior we want to see.
So are your users aging up? Because I mean, Roblox as a company is a young adult now, Roblox has been around for a long time. So how much of it is users aging up, which happens naturally on these social platforms, versus older, new ages?
I think we see both. We see the people who’ve been on Roblox for 12 years now, and a lot of our creators have been on the platform for 15 years, which is really, really cool. In this case, though, this was word of mouth — 21-year-old friends like, “Go try this out.” So I think they are both coming together, and as we get into 30 and up, less of those people have been on the platform before.
I was going to say, I feel like, personally, I’m one of your target users to come on the platform. I’m 30. I don’t have kids yet.
The thing with Roblox that I am still trying to figure out is that it’s such a social platform. It’s more than just gaming. As a gaming platform, I could see you eventually getting experiences that are compelling enough on their own as a first-person type experience for someone my age. But how do you get all my friends on there? That seems like a tricky problem and a unique problem for you all as a company. How are you thinking about that social graph component of aging up?
Yes, I think if we take a step back, arguably, just the gaming market is huge. And so there’s arguably a company strategy where we’re just going to be the best at gaming. We have all these developers; they’re starting to go from one to five to 10 to 100-person studios.
You’ve got studios merging now.
“We believe it’s much bigger than gaming.”
[There are] venture capital-backed studios. We have all of that. I think when we take a step back and we look at the history of technology, we believe it’s much bigger than gaming. And we believe there’s been a historical evolution of just how people communicate and connect. I used to send you a letter. It took four days to get there, and then I could send you a telegraph. [That took] one hour. Then we had the phone system — I could call you and start to have a bit of an emotional connection. And then, for the last two to three years, we’ve been on video. We believe the next wave of this is 3D simulation communication. And there are inherent benefits of 3D simulation that I think all the young people on Roblox figured out 15 years ago.
And that in 3D simulation, we can play hide and seek, we can go get a high school diploma and walk down the aisle. We can have a birthday party. We can pretend we’re in the office together. Those are some that I think haven’t been discovered so far. I think we’re going to see more and more of that. Say we want to have a Roblox company meeting with 2,700 people. It’s one thing to have everyone on Zoom or any video thing. It’s very different to have 2,700 people feel like we’re all in the same place, and me whisper to someone next to me, “Hey, what do you think of Dave?” That kind of thing. Walk around, have all of that social stuff. So as far as body language cues, being in the same place, audio, I think there’s just a lot of inherent benefits. I think it’s bigger than gaming. It’s how people ultimately communicate at work, how we go to school, how we go to concerts — a wide range of stuff.
There have already been some interesting reactions to this in the Roblox community. I think especially among your younger developers who have also noticed that at your gaming conference this year, for example, it’s the first year that it’s only 18 and up allowed. You’re not allowing minors accompanied by adults this year. And so I think there’s already some tension where people are asking, “Does this aging up that’s happening mean that Roblox is going to leave behind the young developers who helped get the platform to where it’s today?”
I’m super optimistic that this isn’t going to be an issue, and I’m so optimistic because Roblox is starting to be used for computer science education. It’s starting to work its way into not just coding education but robotics, for example. Not everyone has the ability to have the physical hardware to compete in robotics. Doing that in a simulated way, learning about traveling to Mars, Mission: Mars. I think there’s just going to be such a continued educational groundswell. It’s much bigger even than our developer community. I think for Roblox Developer Conference, you’re right, it’s going to be 18 and up. It’s becoming a grown-up, professional-type event. But I think our roots are so much in the community and those younger people — there’ll continue to be amazing 13-year-old creators who start to make a living on Roblox. I don’t think there’s any risk of that.
People spend, I think, roughly $3 billion a year or so on Roblox. You’ve got your currency, Robux. What are they buying? I think parents sometimes wonder this, too. “What is my kid buying?”
One of our innovations was to trust the creativity of our developers rather than being very regimented around what you might buy in an experience. [We thought] if we provide the core platform and the infra, we’ll be really surprised by it. And we continue to be really surprised by it. In the early days, we were like, “Oh my gosh, someone’s selling flashlights. That’s really cool. That’s really interesting.” Whatever our creators have sold — vehicles, clothing, VIP passes, all of that — they continued to astound us with their creativity. And I think our creators are always very wisely balancing new users, fun retention without people spending, with those fun things that some people can buy on top of it. So one of my favorites was Bird Simulator. A long time ago, in Bird Simulator, you could have all kinds of things. You pretend you’re a bird, catch animals — what’s it like to be a bird? And if you wanted to be an eagle sooner than later, you could pay a little to do that. So it’s really astonished us.
I think it’s astonished me, too. I mean, you see just things that you’re like, “Wow, it costs that much to have this virtual item in Robux.”
Well, I think then we get out of the experience into virtual items, and I would say I have a personal thesis about virtual items, and I think it is astonishing that luxury goods can be so expensive, right? My son was telling me about a certain brand of watches that music stars wear that cost $250,000.
There’s several of those.
Yeah, I didn’t even know that. But what it highlights is that these behaviors in the digital world, once again, may start to mirror what happens in the physical world. And I think if you and I, for example, long-term spend one or two hours in the digital world and 16 hours in the physical world, it’s arguably possible that our clothing, what we wear, we would pro rata value that, and if I’m spending 5 percent of my time in the digital world, I might spend 5 percent as much digitally on what I’m wearing and the brand as what I spend in the physical world. And I think that supports some of the behavior we’ve seen.
On the developer side of making experiences, I’ve seen some analysis that the platform is trending toward fewer devs making more of the money. I’m wondering if that’s something you’re seeing in the data and if-
That’s a fun, I would say, misconception. And I’ve recently looked at the aggregated growth rate of top 10 devs, top 100 devs, top 1,000 devs. And over the last three to four years, the growth rate in the spending of dev number 1,000 is faster than dev number 10 and dev number hundred is in the middle. Dev number 10 right now is growing faster than our bookings year on year, which is super exciting. We want those devs to be growing their earnings. Dev number 1,000, I think, went from very small numbers three or four years ago to, I think, easily making a living now and growth rates in the 2, 3, 4x over the last few years. So that’s really a sign of a healthy dev ecosystem.
So that long tail is still important to the platform.
It’s almost like you can imagine everything’s accelerating and that long tail is accelerating even more rapidly, which bodes well for interesting content, and bodes well for new experiences popping up. If we ran the number, our ecosystem, which is going to distribute over $800 million this year, supports a lot of developers. There are a lot of people making a living on Roblox right now.
And you announced ads recently, immersive ads. There’s going to be a developer component of that where they can also tie into the advertising system.
“Why would we control the ad marketplace? Why wouldn’t we build a developer-powered ad marketplace?”
This is so true to the Roblox vision of “Why would we control the ad marketplace? Why wouldn’t we build a developer-powered ad marketplace?” We had the thesis that the more immersive the brand advertising, possibly the more memorable. We had the notion that native is very interesting. And so when we look at the history of advertising, print, web image, web video starting to get integrated in various platforms and native formats — you name it, YouTube, TikTok, Twitter — they all have their more native formats. What’s really interesting for us is if we’re walking around here and pick your favorite brand, it could be Gucci, Nike, Vans — you name it. If we see a portal that says, “Hey, stop by the store and check it out,” we literally go to that place, and we go to the Nike or the Vans or the Gucci experience for a few minutes and come back to where we are.
We’re very bullish on a new ad unit that I would call a portal ad unit where if I’m a developer and I put a blank portal somewhere in my place, it’s dynamically served. The advertisers can choose which developers they want to be with, and I participate as a developer in those earnings. I think the potential here is amazing, massive, because, in 3D, just like when we’re in a shopping center, we go into a store like, “We’re fans. Let’s hang out. Let’s check it out.” I think those are pretty deep memories. It’s also interesting that it doesn’t get in front of the user. It’s very optional. We’re out starting to talk to [an] advertising agency and the early signals we’ve had are that there’s a lot of demand for this. It’s very interesting.
How much of your aging-up strategy is tied to this ads push? Because you can give marketers a guarantee that they’re reaching an older audience.
I would say it’s actually a byproduct rather than a weird intentionality.
The strategy wasn’t to do them together?
I don’t think our strategy would be, “Oh, we got to do 17-up so we can make more money.” I think our strategy would be, “We believe this is a platform for all ages. Oh, by the way, we are doing 17-up. Advertising’s really interesting.” That 17 through 24 segment is potentially very valuable, especially in an experiential format, so it was more the other direction.
How do you balance developers being able to advertise their experiences on the platform and ensure that doesn’t game discovery and doesn’t become pay-to-play?
It’s so interesting. I think we have an amazing discovery team that is focused on organic discovery and doing it as 100 percent objectively and fairly as possible. We don’t want to be putting our finger on the scale here. I think we have some big innovations on discovery, on organic versus paid, that we’ve been talking about, but I do think our long-term vision is 100 percent fair, 100 percent objective, and an appropriate balance of organic and paid discovery.
It sounds like there’s potentially big changes coming to how discovery works on Roblox.
I would say long term, it’s just going to get more and more fair. It’s already very fair, but I think we’re thinking at a scale of 10X when there’s both brands buying for Discovery as well as young developers who just launch their own [thing]. How do we really balance a big brand wanting to spend a lot of discovery and some up-and-coming dev in a really interesting way that gives us all of what we want? And I think we all want the same thing. We want that classic story of a young developer who just made their place [and] there’s 200 people there. We want our organic discovery to bubble that up. It’s sometimes called One-Armed Bandit or whatever. Get more and more people to see that. And we also want brands to be able to advertise. I think we’re going to be very thoughtful in making both of those possible.
Roblox is kind of at an inflection point. You all recently announced some big news that you’re opening up experiences specifically for older audiences, 17 and up, which is a big shift for the platform. Talk a little bit more about how you got there. You started as this 2D educational platform. When did Roblox become more of what it is today, and how did that happen?
“In the early days… it was just fun to build stuff. It started migrating into people being able to make a living on Roblox.”
I think it’s interesting because we have two values that work together. They almost signify how we’ve worked through this, which is, one, to take the long view. If we can imagine it and if we think it’s technically possible with the compute power and the devices that people have, we want to aim for that. But constant iteration at the same time, getting stuff done and executing, has been a big part of it. And literally over the last 16 years, through millions of small steps all aiming in the right big-term direction, Roblox has evolved to what it is today. It started out very young. It started out much less realistic. It started out with a developer community that was 100 percent volunteer. In the early days of Roblox, it was just fun to build stuff. It started migrating into people being able to make a living on Roblox, studios being able to make money, technology getting better in every dimension, the avatar system, the cloud system, being on more and more devices. This has just all been constant iteration over 16 years.
You were a private company for most of your history, and then you recently went public in 2021. I always like to hear, on the other side of that, what that experience was like, as the CEO. I think there are always some lessons on the other side of something like that.
My message to future CEOs of public companies would be you do a lot of your work before the IPO. I think we were fortunately running the company almost identically pre-IPO as post-IPO. A couple cool things: generating cash pre-IPO is super nice. Having a lot of cash in the bank — super nice. Having a business that we could imagine being 20 times larger without some big pivot — super nice. Having a philosophy of hiring amazing people, trying to drive innovation — super nice. Having very good modeling forecasting tools — super nice. Having a transparent, open company culture that’s pretty established — super nice. So for us, we had all of these big things in place. It made the IPO almost less of a big event for us.
Your revenue has continued to grow even out of the pandemic, but you’re still not technically profitable, and you’ve managed to avoid doing layoffs when a lot of your peers in tech have done that. I’m curious to hear your philosophy on cash management and hiring. Has that changed at all, given just how the market’s changed a lot coming out of the pandemic, and now we’re talking about new metrics and goals for public companies?
I like that you asked the question about cash management because that’s how we run the business. Cash is really king. We’re in this interesting position right now where, yes, we show a loss on gap accounting. It’s interesting that we show that loss because, just like in the physical world, in the digital world, we defer a lot of revenue. When someone on Roblox buys a shirt, we don’t book it as revenue. We book it over 29 months. On a cash basis, we generate a lot of cash. We’ve invested over $400 million in infrastructure in the last year. We’re essentially at the same place we are with cash now as we were a year ago. And so, from a bookings basis, an ability to bring in cash and fund our growth, we run an amazingly solid business.
And as you correctly note, we continue to hire through these interesting times, I think, through very thoughtful modeling and a constant approach of bringing great people on. We haven’t slowed that down at all. I think we announced recently in our earnings call we’re starting to grow bookings more quickly than cost of sales. We think, in Q3, our bookings are going to grow more quickly than our infrastructure costs. And next year, we’re going to start to see bookings grow more quickly than our headcount cost. Those are all really good indicators of increasing cash margin.
How is Roblox structured?
I love this question because we really think about people and systems as the primary product we’re building here. It’s almost like the primary product [is] great people running in somewhat autonomous units that can innovate as quickly as possible. So our org chart inside of the company is almost as if we’re running eight smaller companies that very much align in terms of vision and areas of innovation. As you mentioned, our infra and our platform — maybe like AWS or Azure Cloud — is a little autonomous unit inside Roblox trying to build the world’s best infra for 3D co-experience and run it as efficiently as possible. We have some great leaders. We have another area, which is our 3D simulation engine. We had a great person join us there recently. That’s almost like a company, building the best cloud 3D simulation area. Another area, our economy, that’s like a whole company, how we run a great economic model and all of that.
So you have literal economists on staff?
I was in meetings two days ago talking about all kinds of really deep economic stuff, money supply, inflation. Advertising is coming to Roblox. It’s super exciting. We’re redesigning our marketplace and expanding it so it’s very rich with a wide range of UGC items. We really focus on as much autonomy for each of the groups within the company so they can innovate quickly, and then we keep them nicely coupled at the same time.
And when those units need to collaborate and there’s disagreements, maybe at the high level, about how that actually plays out, are you the tiebreaker? Is that when you get involved?
As CEO, you want to be tiebreaking as infrequently as possible because then people feel that level of autonomy. In certain cases, we have a huge initiative going on this year about communication and connection — really, what’s the future of how people communicate on our platform? And it goes into some of the things we’ve talked about — facial animation, voice, our simulation engine. In that case, we actually have two of the leaders of those groups, the engine group and the user group, acting as shepherds really around keeping that all connected. It reduces the CEO tiebreaking.
Got it. How do you delegate?
Always trying to delegate more, right? Isn’t that the lesson of a CEO? The more you can delegate, the more individual people can take things over. I think, by this structure where the leaders of these individual groups effectively are running small companies, it creates amazing opportunity for delegation. Our individual groups get together on a monthly basis and go through all the metrics. Their metrics include cost of headcount and cost of infra, and then, for those individual groups, they may have individual metrics. For the economy group, [it’s] what’s our bookings per hour? What’s that look like in all different regions? What’s that look like in individual cohorts? So I think I’m always trying to delegate more. I think every CEO is. In this case, we’re trying to combine that with structure and the way we run the company that almost forces delegation.
How do you spend your time in terms of where you want to spend it? I think every CEO has a passion. You have product CEOs vs. strategy or more financial-minded ones. Where do you want to be spending most of your time?
We have an amazing CEO team now. We have three members on the team who are exceptional. We have active dialogues because they’re trying to schedule my day. We used to be scheduling 30 [minute meetings] — we’re now into 15 and 10 and 5 [minute] kind of stuff. They are actually pointing out to me places where the time management could be better. Time is so valuable. We’re always trying to balance leadership time, recruiting time, engineering time. We’ve almost formed this amazing partnership where we’re designing the schedule around that time management because time is very valuable.
Do you find yourself gravitating to one area the most? Is it like, “I really want to be in product reviews if I can”? Or do you have an area like that?
I mean, in the early days of Roblox, it was two of us writing code. There are some meaty complex systems problems that are very technical that you just think, “Oh my gosh. It’s so fun to work on that.” And those things are super fun to work on because, you can imagine, early on, it was us designing these systems; now, we have amazing people designing these systems. So it’s much more of a mentorship thing, or it’s just a little [bit of saying], “Here’s what we’ve seen before.” But time and time again, when we build these really good systems, like our virtual economy, if they’re built right, they work for 10 years, and they keep going. So building those kinds of systems is super exciting. I think the recruiting time has gotten more and more fun for me because it turns into not really recruiting time — it’s educational time.
I think people underestimate the amount of time that CEOs spend on recruiting these days. It’s a huge focus.
Before this meeting, I just spent the daily 30 minutes with our executive team going through all of the people I’m talking with on a recruiting basis, but that recruiting time has become a lot more fun because it’s so educational and mind-blowing. I get to go hang out with amazing people, and I learn a lot in recruiting time. So arguably, half or three-quarters of what I know about AI is through using recruiting with some of the world’s best people to learn from them.
Sign up for Command Line, a paid weekly newsletter from Alex Heath about the tech industry’s inside conversation. Your first month is free!
A flexible plan you can cancel anytime.
A discounted plan to keep you up to date all year.
Keep your team informed on the inside conversation.
We accept credit card, Apple Pay and Google Pay.
I want to turn a little bit to regulation and moderation, two fun topics and they kind of intersect. There’s this huge push right now to really keep kids off the internet, I would say, in Congress, especially in the US. There’s like a COPPA 2.0 bill. There’s all these talks that we should ban social media for kids. Roblox is right in the center of this debate with such a young audience. What are your conversations like with lawmakers on this? Is this something you guys talk about with them? How do you think about this?
I think we’re in an exceptional position both from a company culture point of view as well as from strategically what we’re building as a platform. People are out speaking in various forums, trying to take a leadership role in what’s the vision for kids on the internet, and if you scan some of the places we’re participating, you’ll see us actually trying to drive the discussion here. There’s an interesting thing around my belief of what we mean when we say social media, and social media means a lot of things. It can mean you and I sharing photos about our life and everyone else looking at them and comparing them. It can mean you and I consuming short form video a lot, but it can also mean you and I hanging out and communicating in a virtual environment and doing some of the things we do in the real world.
What we saw over covid: how do you stay connected with your friends or your grandparents? How do you have authentic communication? I think legislators are getting more and more savvy that social media isn’t just one big thing, that it’s a variety of things. I’m optimistic about the direction we’re going. I think I’m optimistic about a direction around civility, around safe communication, around learning with your friends. I think a lot of the things that people are starting to build on Roblox, I think, are very positive things for the world.
And it seems like you all are leaning heavily into age verification as a way to keep the platform safe because you do have the potential for strangers to interact.
I think it’s safe to say that in many areas, we lead with company values and vision ahead of legal restraint, and I think we self-correct well before legislation in many ways. Our voice rollout: we’re being very, very thoughtful on that. Our 17 and up, as you noticed: full-on photo ID verification, which is beyond probably most apps out there. But we believe, I think, typically by doing this hard stuff that we do create this opportunity for this billion-user platform that we’re optimistic about.
Let’s be honest. The verification is also a business growth barrier in the near term. If you were being purely capitalistic about it, you wouldn’t try to do this.
We have this saying: “Optimize five-year enterprise value.” If we could say, “How do we optimize Roblox’s enterprise value in 2028?” probably what we’re doing is going to optimize that because we are eventually going to get to larger and larger numbers of 17 through 24-year-olds. We generate more than enough growth and cash to say, “Oh, we’ve got to let every 17-year-old on right now.” And so I do think it’s a little bit more of a long-term optimization rather than short-term.
Like I said, I don’t have kids. I did talk to some friends who have kids on Roblox ahead of this to hear about their experience as parents with kids on the platform. I heard a recurring thing from several parents and they were like, “You should ask David about this.” What do you say to parents who have already witnessed their kids experiencing things on Roblox that are not age-appropriate, that are not up to your standards as a company — obviously, you’re not going to catch everything — but who are now worried, saying, “Oh, now they’re trying to age up the platform and get more mature experiences on it”?
I would say, since day one, we’ve had zero tolerance for that kind of stuff. There’s sometimes bad actors on our platform trying to get some content out. I’d say it’s continuously getting better on our platform. We’re not 100 percent perfect, but we’re getting very, very, very good at that. I think 17 and up is going to be very rigorous as far as who can access it. Any 17-and-up content, once again, is going to be gated by photo ID. [Younger people] are not going to see that. But I would say we have zero tolerance, and I think it’s our top priority. Our safety group and team is probably the largest group in the company right now, if not. And so it’s a big focus for us.
Do you think the more mature things like nudity will ever be allowed on Roblox in any way?
We would lead first and foremost with safety and civility and then second with mirroring what happens in the physical world. And so I would never rule it out. I would say getting to that kind of stuff probably isn’t in the short-term future for us.
With Roblox being such a UGC-heavy platform in all ways from people creating items, creating literal virtual objects, to the actual interactions that human beings have as their avatars, what are some of the moderation challenges there? I imagine they’re very unique to a platform like Roblox, and you’ve built a lot of this in-house over the years. What have you observed, and from when you first started to really do moderation to now, was there a moment where you were like, “We have to approach it this way. This is the only way we’re going to fix it”? Walk me through that journey of how moderation started.
The journey is pretty interesting. Imagine four of us in a small office in Menlo Park being live for about a month and peak times, maybe 50 people on the platform and the temptation, more engineering, more engineering. There’s so much more stuff we can build. And then my partner Eric saying, “We have to build a moderation system.” And so it was actually a real values moment for us to say, “Okay, yeah, this is what we’re going to work on a month in as we’re growing.” All the way to the point, probably for two to three months, the four of us, we were the moderators, and so we rotated being moderators every day as part of it, as we gained experience there. It really started very early. I think it’s always been a top focus for us. The challenges just get more and more complex all the time, and we get better and better all the time. But the core value has always been there.
I don’t think we’ve said the word metaverse yet, which is frankly amazing this far into the conversation. I think if we were having this conversation a couple of years ago, we would’ve said it in the first 60 seconds. And then there’s this whole generative AI thing, which is kind of the new metaverse in a way, in terms of being a buzzword. I think there’s a little more to it than that. How have you watched the conversation change into the AI wave that we’re in now? Roblox is putting out some really interesting stuff in generative AI. Is Roblox a metaverse and generative AI company? How do those two ideas intersect?
“Metaverse is a word that actually was hot 14 years ago, subsided, got hot again three or four years ago, and I think we rode that.”
I like the notion that irrespective of the news cycle or what’s been going on, our product development trajectory has been relatively consistent, smooth and straightforward. I like the notion that what we were doing five years ago blends into what we’re doing today in a smooth fashion driven by our own vision rather than the news cycle. We’ve always called ourselves a human co-experience platform. We’ve always talked about being a communication and connection platform, a place where we reimagine the way people come together. Metaverse is a word that actually was hot 14 years ago, subsided, got hot again three or four years ago, and I think we rode that. That said: on the metaverse angle, we do believe this new emerging category runs on all devices, which is really cool. It runs on phone, tablet, computer, console, VR / AR devices — all of that. I think we’ve been very consistent there. Same on the AI thing, in that our Trust and Safety team’s been running all kinds of AI models for the last three or four years, both on quality and on cost. And we’ve gotten to the point where I think over 98 percent of the 3D models, for example, that are submitted go through AI scans that are more accurate than human scans right now, so this is both a quality and a cost-benefit.
We started rolling out more creation-type AI, so both material generation and code generation, but there’s so much opportunity here. I mean, we can imagine what it could be someday. There’s that Westworld episode where someone sits down and says, “Do this and do this.”
The narrative storyteller role.
That’s all going to happen.
Absolutely. And it’s going to happen, not just in Roblox Studio — it’s going to happen in all of the experiences made by Roblox creators as well.
“You really do think we’re going to get to that point where people are speaking things into existence?”
It’s so funny you brought up that because I’ve been thinking about that scene and that concept in Westworld, myself, with where the tech is going. You really do think we’re going to get to that point where people are speaking things into existence?
Absolutely, and I think what’s unique about Roblox is those will not just be static things. Those will be functional scenarios. Those will be NPCs backed by AI. Those will possibly have logic in them.
How many years out do you think we are from a creator being able to type or speak a world into existence on Roblox?
I think we’re very close to [that]. We’re less than five years out.
Wow. On competition, I’m really curious. I don’t think I’ve heard you talk about this yet. Epic Games has their Unreal Editor for Fortnite and are doing some interesting things on the economics side, too, in terms of how they’re sharing revenue with creators, which is, I think, a net positive for the industry. Creators getting more money. They’re approaching it differently than you all are, and they have kind of different use cases. What do you make of what they’re doing with Fortnite and trying to turn it into more of an open-world creator economy?
I would say we’re so proud of this leading platform we’ve arrived at, with all of our developers, and super proud about that. We are also really proud of the notion that early on, UGC for us was so important, and it has proven to be just such an innate thing, ultimately, that it’s a big thing for everyone, so we’re super proud of that.
We continue to just be super proud of the economic opportunity on Roblox as well. We’ll probably distribute over $800 million this year. We are starting to see these VC-backed studios, and I do think we invite all of the creators on Roblox. There’s such a range of awesome creators out there. If I think about a creator and what they are thinking about. “Okay, I’m 13. I want to try making something. Can I push it live and have it run in every language, and can I go viral?”
Maybe myself and three or four people want to build something really interesting and go live, or I run a 200-person studio and want to both make my own stuff, as well as brands. And can all of those work? Can all of those have a hope to not just be on a 60 million-person platform but a billion-user platform? We’re very thoughtful about all of them, and I think they’re joining us in this really long-term opportunity.
So you think you still have the most attractive revenue sharing arrangements and payouts in the market?
I think we have an amazingly attractive opportunity.
The reason I say that is because I know Fortnite is doing it differently, but the percentages are different. I’m wondering if with what they’re doing, and with Horizon Worlds coming to mobile later this year most likely, do you see there being more competition in the rev share that you give creators, and is that something that Roblox will eventually change because of the competitive dynamics, or do you think you don’t need to?
I think what we want to do as a public company is continuously grow bigger and bigger. I think we want to offer those devs ultimately not a 65 million DAU (daily active users) platform but a billion DAU platform.
And I think we want to run the company in a way so the core expenses, cost of sales, and our employees, what we spend for infra, are run so efficiently that the raw cash being generated, as much of that as possible, can go to the devs.
What I’m saying is, I think it’s bigger than just the percent distribution. I think it’s “What’s the platform, and what’s the total economic value going to those devs?” I think it’s “What’s the size of the user base on the platform?” I think it’s the variety of “What’s the quality of the tooling?” I think “What’s the opportunity for search and discovery?” So it’s so rich that I think rev share is a small part of it. And I think we’re a little bit used to this in the early days of Roblox when we continuously tried to move as much money to the developer community — the only thing to compare to was the iOS App Store, which was 70 / 30. It’s very similar. It’s like, “Well, why aren’t you giving 70?” Well, it’s because we have infinitely sharding infra, and we have trust and safety, and we have moderation, and we have a lot of these other things that make it happen. So when I empathize with devs and imagine I’m devs, I feel very strongly about our opportunity.
Have you tried the Apple Vision Pro yet?
I have not. They’re being very scarce with them.
I was one of the lucky few. It blew me away visually, and I’m wanting to ask you about that because Roblox is technically available on VR headsets, but you really haven’t embraced it, I would say, full-throatedly, as a company, that category of headsets. I’m interested — with the Quest 3 coming out later this year by Meta, Apple’s Vision line that they seem committed to — do you see an opportunity for Roblox in this category that is perhaps bigger than what you’ve done to date?
I think the opportunity is really big to imagine a platform like this that runs on every device. And I think, just like 2D HTML was early on something that jumped from PCs to phone, with the Apple iPhone, pinch and zoom, and all of that. I think 3D immersive is that same way. Once again, phone, tablet, computer console, living room VR. So, yeah, I’m bullish on VR in the sense that it offers super deep immersion. When I boot up an eight to 10-year-old Roblox experience that I might have made 10 years ago and go into it in VR, it’s a whole new deal. Like I’m here, now, kind of thing. So I would say we want to be on all platforms, and we definitely want to be on VR, and stay tuned.
Good to know. Last question. This one is actually from my colleague’s eight-year-old godson—
—who wants to know, is there a way to get infinite Robux?
I’m looking at our team. So this could be kind of a really good educational opportunity because as a parent, you could imagine, okay, there’s a company, a company has shares of stock, there’s the US, and it has the money supply, and then there’s Roblox.
What is Roblox in that analogy?
Well, there’s interesting things in each one of those when you print infinite things. So, if Roblox as a company printed another billion shares, there’d be a lot of dilution, and our shares would be now worth two pennies. So the value of the company stays the same. More shares, less value. I won’t get into the US money supply, but one could argue that if the US increased the money supply by 2X, the value of what you buy with a dollar might drop.
It might. We may be experiencing that.
Yes, so we have to be careful. That’s probably the reason why we can’t give out infinite Robux: because if we did and one person had infinite Robux, the pricing of every item, everything our developers do, would probably reduce by a factor of a hundred.
Kids, this is why it’s important to learn about inflation.
David, thanks so much for joining me.
Decoder with Nilay Patel /
A podcast about big ideas and other problems