Roblox CEO David Baszucki has a goal that even he admits is “a little audacious”: to grow the company’s share of the global gaming market’s revenue to 10 percent in the coming few years.
Roblox’s CEO on getting to 1 billion users
Roblox may already be bigger than the entire AAA game industry. What’s next? Plus: Telegram’s mea culpa and Amazon’s latest sneaky AI deal.
Roblox may already be bigger than the entire AAA game industry. What’s next? Plus: Telegram’s mea culpa and Amazon’s latest sneaky AI deal.


For comparison, the company’s investor guidance of $4.2 billion in bookings this year is about 2 percent of the global gaming market, which Roblox estimates to currently be between $160 billion and $180 billion. Baszucki shared this goal, which ladders up to his ultimate mission of reaching 1 billion users, during the company’s annual developer conference on Friday.
The scale of Roblox is already staggering relative to external perception of the platform still primarily being for just kids gaming. There are now 25 experiences with over a million daily actives, and as Matthew Ball recently highlighted, “it’s likely Roblox has more monthly users than the entire AAA gaming ecosystem combined.”
I spoke with Baszucki in an exclusive interview earlier this week about his growth plans, AI efforts, response to recent criticism about poor moderation on the platform, and more.
The following conversation has been edited for length and clarity:
In conversations with me and others, you’ve really been clear that you see Roblox as more than gaming, that it’s a social platform. So, I’m curious, why highlight this revenue goal in relation to the gaming industry?
Yeah, the positioning of the company is broad. I think it’s a big step to go from 79.5 million DAU to a billion. Seeing a path to 4x-plus growth as we continue to bring social forward as well, it just adds more color and clarity to the journey to a billion.
By the time you’re at a billion, do you still expect most of the use cases to still be social gaming?
We’re not sharing the expected mix. I think getting to a billion has more social. Gaming is probably not 80 percent of a billion. I think we expect more social, more entertainment consumption, more utility-type uses. We are internally building out a Roblox content thesis of what we look like at 300 million [daily users]. How big is social? What types of experiences are powering that?
Do you see any big platform shifts that need to happen, either with the platforms you’re on today or new ones, to get to a billion?
I think we’re on all the right platforms right now. Moving to 70 percent rev share for paid gaming is an expansion, which is really going to be big for us. I think rewarding creators with an affiliate program — so when someone makes a great experience and does a lot of social [marketing], they can say, “You know what? This feels fair to me because when all these users come to Roblox, I’m going to share the revenue of them in addition to what I get in my game” — that’s going to make it a lot easier for creators to be comfortable leaning into heavy social marketing on our platform.
We’re going to announce a huge effort within the company to support 100-plus-player, open-world sports or battle royale-type experiences on low-end devices with two gigs of RAM in difficult countries. There are some genres we’re very, very strong in. There are other genres that we feel there’s huge opportunity in: sports, battle royale, first-person shooters.
I know there have been paid experiences on Roblox already, but the increased rev share feels significant along with the switch to USD for desktop. Why are you doing that?
We run all of Roblox infrastructure now at less than a penny an hour. We’ve made incredible progress on driving the cost of that down. The cost is getting better because we’re using more and more AI in our safety and civility stack. We run on our own, 27 data centers. We have our own compute. We have our own bandwidth. We have a little mini Azure/AWS for Roblox gaming, and that’s getting extremely cost-effective.
So when we run the numbers on expected lifetime of the paid experience gamer, we all of a sudden see that, if we get $19.99 for a paid experience, we’re very comfortable. We can support that user on our infra for as long as they might expect to play. And so that’s part of the calculation, that when we move from freemium into paid access upfront, we can easily support all those users efficiently on our infrastructure.
We’re also doing this initially on desktop with credit card. We do plan to make it available on Apple, Google, and other platforms as well.
Is part of this also about getting more premium-feeling, higher-fidelity experiences in Roblox?
I think that is an organic cycle where if the technology is right, if the economics are right, if the user acquisition is right, and if search and discovery is right, we will create a fertile place for an expanded range of type of experiences. And so, yes, we believe we’re setting the groundwork for a more expansive range of type of content on the platform.
The EU has started mandating alternative app stores on iOS. What are you thinking about with that? Would Roblox do its own store?
We have analyzed it, but we have no announced changes there.
How do you feel about the possibility?
You know, it’s an interesting world. Part of our whole philosophy is being great partners with Apple, with Google, with Microsoft, with Sony, and all of our platform partners. And at the same time, of course, we want to offer the most value to all of our users. So, we hold both of those values at the same time.
This 3D, multimodal AI model that you guys are building, why open-source this and allow it for off-platform use? It seems like your incentives would be to keep this as something exclusive to Roblox.
What we’re doing and what we’re going to be supporting with open source is core foundational 3D creation. And 3D creation is stuff we use all over. It’s meshes, it’s textures, but it’s not necessarily the stuff that drives a really immersive video game. So we believe by open-sourcing and working with partners, in addition to all the amazing data on Roblox that we can train on — and I would note that the Roblox community has embraced the notion in a privacy-compliant way of having their data contribute to the quality of AI — we’ll be co-developing the ability to turn that 3D information into what we call functional, 4D information.
Assets on Roblox all have material properties, have code embedded, are things that can have physical properties that can be dropped directly into a game. So, we’re open-sourcing core utility, 3D creation. We believe there’s a second step, which is making that into game-ready assets that are more emulating the real world on Roblox.
Have you had conversations with Roblox creators about this AI model you’re training and how they feel about it? I’m sure you’ve seen there’s a lot of debate about how creators should be compensated for contributing to these models.
We have, and I want to highlight we’ve adjusted our terms of service so creators feel comfortable about it. The promise of this is 3D generative experiences within any game you could imagine. And so, our hope is this is not simply a developer tool at studio creation time. This is a dynamic functionality.
For example, a developer of a fashion experience where clothing is being made would allow that clothing to be made in a generative way based on a prompt, in addition to other ways. And we believe, ultimately, many experiences on Roblox will be running against our AI cloud, whether it’s a simulated George Washington running on a cloud or whether it’s a clothing experience where someone’s building clothing from a prompt.
I wanted to get your reaction to the Bloomberg story from a couple months ago and how you felt when you read it. Then there’s this Turkey decision to block Roblox, and I think it just came out earlier today that you guys are appealing the government’s decision there.
It has been the primary focus of the company since we’ve gotten started. And it’s our top focus. Any incident on our platform is unacceptable. It’s a personal commitment. We have a lot of parents who play on Roblox, and we just keep getting better all the time with AI, with the quality, with the vision. So, that’s my take. I think we’re solidly on our way to building the world’s safest platform.
And then regarding Turkey, we respect every country, and we respect the laws and values of every country. We’re optimistic and working with their government and their courts to get Roblox back on there.
There were two things that stuck out to me in the Bloomberg story, which was the difficulty of policing off-platform behavior, and allegations in the story that trust and safety wasn’t as well-resourced as you all had said publicly. Can you respond to those two points?
When we do headcount planning or resource planning, safety is the top priority, and especially when it comes to customer service and moderation, we always make sure we have enough people there to handle all our moderation. AI is getting better and better and, more and more, is acting as a complement to those people. It’s allowing many of our employees to do higher-level functions, to dive in on special exceptions and all of that. But overall, my reaction is that all metrics are improving.
Elsewhere
- Jail changes a man: It was fascinating to see Telegram CEO Pavel Durov’s 180 turn on content moderation after spending a few days in French jail. He’s now admitting that “growing pains” have “made it easier for criminals to abuse our platform.” More tellingly, he scrubbed a section of Telegram’s FAQ that said the company ignored moderation requests for private chats. Beyond the legal repercussions, Durov has financial reasons to clean up Telegram: the Financial Times reports that the company posted an operating loss of $108 million last year and has over $2 billion in debt financing set to convert to equity “at a discount to Telegram’s IPO price if a listing takes place before the end of March 2026.”
- Cue the “can’t keep getting away with this” meme: Amazon is now the only Big Tech company to have completed two AI deals that I am calling reverse acquihires. This latest one, announced on the Friday before Labor Day weekend, will see the cofounders of AI robotics firm Covariant, along with a chunk of their employees, join Amazon. Covariant is getting paid a “licensing fee” that is actually being used to make investors in the startup whole. One investor in Covariant even called it an “acquisition” in a LinkedIn post I saw, which, oops? Based on my count, this is the fourth deal like this Big Tech has done so far, and as long as regulators don’t intervene, I expect more of them to be announced in the coming months.
- OpenAI needs more money: Lots of reporting came in just before the holiday weekend that says OpenAI is raising another mega round of financing led by Josh Kushner’s Thrive Capital. This round could coincide with a restructuring of OpenAI’s convoluted corporate structure, with the controlling nonprofit officially taking more of a backseat to the commercial entity OpenAI has already become (Apple and Nvidia certainly aren’t going to invest in an actual nonprofit). The question for anyone putting money into OpenAI at a $100 billion valuation is how that multiple is justified given the rise of open-source models, Anthropic’s recent momentum, and the immense cost to continue training frontier models. OpenAI is still burning billions of dollars a year but has managed to be very lucrative in a specific way: employees reportedly cashed out over $800 million in secondary stock sales over the past year.
Quote of the week:
“Our product strategy has been inspired in part by In-N-Out Burger… Their hamburger, cheeseburger, and Double-Double are our Snapping, chatting, and watching Stories. You can decide which parts of our product are the milkshake, fries, and secret menu.” - Snap CEO Evan Spiegel
People moves
Some interesting tech career moves you may have missed lately:
- A few OpenAI moves: Nate Gonzalez, a lieutenant of new CPO Kevin Weil at both Facebook and Planet, has joined to lead business product. Longtime policy leader Anna Makanju has moved into the new role of “VP of Global Impact,” with Chris Lehane officially taking over policy. And Irina Kofman left Meta to run strategic initiatives.
- Nick Pickles, head of global affairs at X and one of the last senior holdouts from the pre-Musk era, is leaving (and likely not vacationing in Brazil anytime soon). Renato Leite Monteiro, X’s global head of privacy and data protection officer, also recently announced that he’s leaving.
- After a stint at Bolt running partnerships, T.J. Won is back at Meta as senior director of partnerships for AR glasses and wearables.
- Coatue founder Philippe Laffont left the board of ByteDance and was replaced by Xavier Niel.
- Valeri Liborski is the new CTO of Yahoo.
Interesting links
- Bret Taylor on Invest Like the Best with Patrick O’Shaughnessy.
- OpenAI’s newest board member, Zico Kolter, on the 20VC podcast.
- Time for round two of Google’s antitrust woes.
- Midjourney says it’s “officially getting into hardware.”
- “It’s 1994 again.”
- Analysis by Sapphire Ventures on the state of AI businesses.
- Tech meets tennis.
If you aren’t already subscribed to Command Line, don’t forget to sign up and get future issues delivered directly to your inbox. You’ll also get full access to the archive featuring scoops about companies like Meta, Google, OpenAI, and more.
As always, I want to hear from you, especially if you have a tip or feedback. Respond here, and I’ll get back to you, or ping me securely on Signal. I’m happy to keep you anonymous.
Thanks for subscribing.












