Astrocade hits 5M MAU and 75K user-created games — building the TikTok for AI-generated playable content

Speaker 1: Marin maybe, but I still think the total amount of views will be higher in Los Angeles than Marin. I think there are some maybe But there's way more celebrities and whatnot. Like, there's But congratulations

Speaker 5: to the Laurel Supply team Yeah. On the launch. Welcome on the show. We love talking I

Speaker 1: want have the CEO on the show. Anyway, we have our next guest. We've been keeping waiting in the waiting room. Here he is in the TBPN Ultra down the mirror from Astrohade.

Speaker 6: How are

Speaker 1: you doing?

Speaker 10: Hey, guys. Thank you for having me.

Speaker 1: Joining. To the show. It's great

Speaker 5: to meet you.

Speaker 1: Great to the progress. Introduce the product, then we can talk about the fundraising. But how are you describing what you're building right now?

Speaker 10: Yeah. Astrohade basically is a social entertainment platform where anyone, and I mean anyone, can create games via just natural language. Very similar to how you problem chat GPT, you can make games. For example, you can say a car racing that is happening under the water and the driver is an elephant or a candy crush that instead of popping the balls with color, you can form basically words. Something like that. And then with a few clicks, we can just publish it to our feed and then there are millions of players right now that play your games. Yeah. And I think the part that I'm the most excited is our mission, is sharing fun with the world. Mhmm. I think what we are building is pretty cool for people to use and be happy.

Speaker 1: This seems Love it. I don't know, like almost obvious. I I don't want to degrade it but it feels like there's been this vibe coding boom. I see a ton of really fun games. We've made like simulators and whatnot. We always buy a domain name and then the question's like how do you drive traffic to it? You can't really go viral. So then you wind up making a viral video about the game and you try and get people to go to the.com.

Speaker 5: Well, here's the thing. So games that we've been creating

Speaker 1: Yeah.

Speaker 5: Are only funny to like a very and fun Yes. Like a very small group of people.

Speaker 1: Yes. Yes.

Speaker 5: But they're not really fun to play.

Speaker 1: That's true.

Speaker 5: And so what I wonder what I wonder is like is the is like reskinning a bigger opportunity than like generating entirely new games because there's certain game mechanics that are just sort of tried

Speaker 1: I would argue that that's how that's how Instagram started. Like the original Instagram was like share your picture of your run with your close friends. That's what we're doing with the simulators that we build. But eventually there's game mechanic

Speaker 5: at all.

Speaker 1: No. No. No. What I'm just saying is like in terms of the actual application like like we we like we are doing the we are doing the you know, the the Instagram post that's just like, hey, I ran a marathon and 10 friends like it. We are not creating the viral Instagram that will be like loved by everyone.

Speaker 6: We're

Speaker 1: not Yeah. Professor Sendy making wombos All all of

Speaker 5: I'm saying is like the the social aspect, the sharing aspect is very cool. But if I'm making a first person shooter, there's probably some like out of the box mechanics

Speaker 1: that would be

Speaker 5: great and reskinning like putting the team I

Speaker 1: think probably a lot of harness. Like, yeah, tell us about how people are actually using it.

Speaker 10: You are actually making all the good points. First of all, and you mentioned this, it's actually obvious. When we started three years ago, we also I was like convinced that such a thing should happen. Yeah. Right? It's very hard to also predict the future, but if you look at the past, the past twenty years, Twenty years ago when Twitter launched, basically, what happened is they made creation tool of text accessible to everyone with the keyboard, all these things, right? Again, and almost everyone is the writer now on Twitter. They do a lot of rage baiting on a day to day basis or whatever. But everyone is right there. And the pattern, what we see here is, once the creation tool becomes accessible to everyone, great generational companies appear, right? Same thing. Four years after Twitter, Instagram. When the camera, iPhone four launched with a good camera, everyone became a photographer.

Speaker 1: Yeah.

Speaker 10: And it's great, but also, like, thanks to my wife, I don't have any more privacy and I think you guys would do that too. TikTok, same thing. And like with the videos, all you see the pattern basically is a creation tool became accessible and what you guys pointed out is like creators from diverse backgrounds started creating this content. First person shooter thing. A lot of people remix new games, but also, the beauty of large numbers is a lot of people actually invent new kind of mechanics. It's like one of the interesting genres for us that we saw that got super viral is you have to wash wash things. So you have something, you go wash Mona Lisa's painting, then you go wash a car, you have to clean it. And like, it's actually pretty addicting when you play it. It's so shocking.

Speaker 1: That's funny. And and and on Instagram, you'll see viral videos of people power washing old rugs that are so dirty. There's a whole genre of this time lapse of people cleaning rugs and then it's very satisfying because you're you're watching something they cleaned which I think is satisfying but then also it's revealed to you what is the what the rug looks like and it's very satisfying Exactly. To

Speaker 10: When you see things happen, you're like, oh, oh, it's so so obvious, but then like, you know, you need people to create these things.

Speaker 1: Yeah. And I

Speaker 10: think that was exciting. The the biggest the I think the biggest advantage of Astrohade, which I really love, is the core of the experience. So number one, first of all, like our platform compared to other platforms is it's You play things rather than consumption. Majority of the social platforms right now are consumption. Engagement is just so much better. It's like hearing a story versus experiencing a story. Yeah. Right? Number two, which is kind of also our mission, just it's more fun. We are living in a world of a lot of the brain rot happening and I think this is just more creative. You come out of it super happy when you play Astrohid and, like, you use actually your brain. So that's very, very important for us. And I think the reason Azhang I believe has the possibility to become a generational company is games are so innate to humans. I have a two year old and he's just always playing. He's learning to play, Right? They were like, what? I think the latest stats I read was the 3.4, 3,500,000,000 people playing video games right now, just almost half of half of humans.

Speaker 1: Wow.

Speaker 10: Right? And I

Speaker 5: think That's not this second, right? That's like day to day, generally, 3.4.

Speaker 10: Overall, I think that no.

Speaker 5: It's not be pretty worried.

Speaker 1: Speak for yourself. I'm gaming right now.

Speaker 5: Half of humanity is is live playing first person shooters right now.

Speaker 1: But brain rot's not gonna be an issue. Don't worry about that.

Speaker 5: There must be is there a crazy power law with games? Like, YouTube, the power law would be like most videos

Speaker 1: Gangnam Style has like A views or billion views that the average YouTube video gets. The medium YouTube video is like 10 views.

Speaker 10: A 100% there's always. With all these creation platforms, there is a power law. Yeah. A lot of like like some portion of creators, like 12% of the creators get a lot of plays and like they get super viral and then there are other creators that get they don't get viral but they get at least thousands of plays on their games. Yeah. We actually, like, we launched the platform like publicly around seven, eight months ago. The numbers are actually crazy, guys. The stats is our MAU is like 5,000,000 right now. Wow. Hundreds of thousands. Hundreds of thousands of creators are now creating games. Hundreds of thousands. Obviously, like, you guys are part of the Internet, you guys are happy. We we we actually processed trillions of tokens Hey. Make this happen. Like, it's completely free. Yeah. People people

Speaker 1: Oh, okay. Okay. Free. Free.

Speaker 5: Free.

Speaker 1: That doesn't sound like it's going make any money. How how will you make money? Ads or subscriptions or both?

Speaker 10: Potentially, we want to do ads. Right now, our creators are, like, it's completely free. Between all these, like, all the vibe coding platforms, we are we are the only one that is completely free.

Speaker 1: Wow.

Speaker 5: Yeah. Yeah. Won't won't you want to make allow creators to turn on monetization at some point? Like, I I feel like you're going to have people that are like nine years old that create some hit game and and, you know

Speaker 1: It's the it's the Alex Zhu from Musically philosophy of like social networks are two sided marketplaces. You have to create liquidity and like the faster you can start paying people to post, you get more professional creators and I'm sure you're thinking about that, but Yeah.

Speaker 10: We're we're exactly. Like, I think we get inspired a lot by YouTube also. Like, almost of them, almost all these social platforms are doing that. Right? They they don't monetize their creators, actually build tools for their creators to monetize their games. Sure. And we are gonna do something similar. We still haven't started monetization. Are focusing on the user growth. But we are gonna build tools that companies can do ads on Astrocade, also creators do microtransactions and the regular type of business model that gaming has been doing all these years. I think that gaming is just generally a good business, it's very profitable, it's one of the largest Yeah. Like segments of entertainment, larger than movies and all these things combined. Right? So I'm not so worried about monetization. We just wanna make a great experience for everyone. You guys should actually play a game.

Speaker 1: People people people the video game industry to the to Hollywood or the movie industry and and it does dwarf it. But I wonder what the actual revenue split between like passively consumed content is versus actively played games. Because I think if you take YouTube and Netflix and Spotify and TikTok and Instagram, that's probably a bigger pie. To be clear, massive, massive market, massive opportunity, very bullish on this. This makes a ton of sense. It it is interesting that there is some there is some sort of dynamic where, like, yes, 3,000,000,000 people are playing games, but probably more people are still watching passive things. There's probably more watch hours on passive content. But that doesn't make, you know, playing games any less important. People enjoy world.

Speaker 10: You're you're right. Yeah. There so that's why I I always not I don't say normally games. We call it playable content because there's actually, like, very diverse set of content. They're not only games.

Speaker 8: Yeah. Yeah.

Speaker 10: You know? It's like Yeah. Stable experiences. Yeah. I think from the content point of view is we are heavily towards actually ultra we we call it ultra casual, not even Yeah. Hyper casual. Yeah. Ultra casual. When you go to We

Speaker 6: got casual.

Speaker 1: Yeah. What what what's the average, like, playtime? Because, like, Final Fantasy might be, like, a hundred hours. GTA might be five a hundred And then some, you know, small game might be a couple hours, but I imagine this is, on the order of minutes?

Speaker 10: Exactly. In the it's exactly on the order of minutes.

Speaker 2: Okay.

Speaker 10: So it's shorter form. Actually, my wife only my wife is one of the hardest critics. She only plays Astrogate games. She does it when she's watching a movie, she also is playing the games. Yeah. It's because it's just so like, it's it's still like engaging and interactive, but you don't need to also necessarily to focus full on versus a lot of triple a games, you have to do that. And I think that's gonna happen with the gaming industry. There is gonna be still bigger studios Yeah. Triple a studios like basically Hollywood and Yeah. There is gonna be Yeah.

Speaker 1: I mean, in same year, we got Oppenheimer and Barbie. We also got like, 12 other MrBeast videos. Are you are you based

Speaker 5: in LA? We are

Speaker 10: in Palo Alto, actually. Dang.

Speaker 6: Okay. We're

Speaker 10: down from Palo Alto.

Speaker 5: I looked at that building is so I thought I was

Speaker 1: recreated Los Angeles up north. It's crazy. Talk about the technical side of things. I imagine this is heavily powered by like WebGL and WebAssembly. But like all of this is delivered in an app. And so I'm wondering like where you're bumping into like Apple's rules of the road because there's so many cool functionalities like haptics and gyro and cameras. And like when I think about what a really robust mobile game does, it's usually using a lot of the different APIs. And some of those are probably available. Are they all available? What's that been like?

Speaker 10: Yeah. So we are completely right now web based.

Speaker 3: Okay.

Speaker 10: And the games are HTML games. Okay. Super lightweight. It's super fast to load those games.

Speaker 1: Okay.

Speaker 10: Technologically wise, I think we had multiple, like, challenges across the way. We knew this is possible, but it's actually very hard to build such a platform. Just generally, platforms are very hard to build. The first biggest challenge which we overcome and it's still ongoing is just creation tool. Yeah. How do you build this agentic system that can make everyone a game creator?

Speaker 5: Yeah, yeah, yeah.

Speaker 10: Right? Remove all the skill because these games are being like, one of the games I was showing to experts and they were saying this is a million dollar budget for a month of Yeah, the

Speaker 1: of course.

Speaker 10: Right? It is on our platform, like $20, $10. So that was one thing and it's all going and it's we call it like, it's a like, Astrohkit is a platform where we our AI is learning how to generate fun. Yeah. It's a big thing. AI learning fun is a big

Speaker 11: thing in the AI

Speaker 3: community. Yeah.

Speaker 5: So it's not a very seemingly at least very deterministic problem. Right? Because you have these games like, you know, I don't know. Why would a why would a window window washing game be fun? And yet, if the mechanics say you've

Speaker 1: never washed a window. Say you've never turned for the mines. Why would why would Minecraft be fun?

Speaker 10: You go to the floor state. Yeah. When you start washing the window, you get to the floor state.

Speaker 1: So I mean, does that mean that there's going be some sort of like, you know, crazy deal to be done between you and Apple at some point? Because I imagine that your consumers, you as a company, will want access to the full iOS API suite. And Apple's going to say, this is an app that makes other apps. We're not really cool with it and you're going to have to discuss that. Like what how does that play out? Because I want I feel like everyone wants this to be a thing but Apple's just sort of operating from a place of fifteen years ago in terms of like what the technology was capable of.

Speaker 10: We haven't. We haven't got there yet. We haven't launched that, by the way. Know. It's fully on the web. Yeah. We have to see how when we cross. No. I think this is actually pretty good for Apple too.

Speaker 1: It should be.

Speaker 6: Because we're bringing

Speaker 10: massive distribution also. Yeah. Like, they they provide a distribution.

Speaker 1: But they make a lot money on games right now. If you're saying that we're just going to do ad supported and it's all going to be in our platform and all of a sudden it's like, well, we were making $2 off of the window washing game every time someone clicked the in app purchase and now you have a different business model that could be an issue.

Speaker 10: Yeah.

Speaker 4: I don't know.

Speaker 9: I'm sure you'll saw

Speaker 10: that. It's right. This is not our biggest problem. Right? Yeah. Yeah. Yeah. Now, the things that we are focusing right now on the technology side is like another thing, by the way, is very hard problem. It's harder, honestly, than some of our AI problems Yeah. Is the recommendation system.

Speaker 1: Interesting. I was wondering if that was off the shelf. Is it not?

Speaker 10: No. It's actually pretty challenging. I don't think anybody has built such a large scale recommendation system for this type of content.

Speaker 1: Sure.

Speaker 10: Right? We have a kind of recommendation system for videos, we have recommendations for images. Games, Sure. First of all, are a lot of them are open ended. So you don't know exactly like video, you know. I watched the whole 50% Yeah. Games is like a lot of other features or features of extracting features from the videos are

Speaker 1: Could be satisfying because you were able to finish it in ten seconds as opposed to if you're watching a ten minute YouTube video and you click out of it in ten seconds, you didn't win the goal. So you have to bake that into the weights as well. Like this person was able to destroy this game but they enjoyed it and then they might not always give you the thumbs up, thumbs down signal. So interesting.

Speaker 10: Exactly. Exactly. Yeah. That's been one of the hardest problems, by the way, you're trying to solve. Yeah. The recommendation system. And then platform, like, do you manage the platform so players are happy, the creators are happy. Yeah. Yeah. This is the first time that, like, this large scale week, like, in few months, we've got more than 75,000 games on our platform created. Wow. How do you distribute these games to millions of people from diverse backgrounds and skills? Right? So there's a lot of like technical challenges. Serving is another infrastructure you need to build

Speaker 1: Yeah.

Speaker 10: To basically efficiently serve these games

Speaker 1: to Yeah. I'm doing some back of the envelope trying to pull data out of you. You said, like, ten to twenty dollars per game inference, 75,000 games. That's like 1,500,000 in inference. Like, are are you are you actively looking at, like, what pieces of the puzzle you can move to cheaper models really quickly?

Speaker 10: We do that. We do that, actually. We we for some tasks, do cheaper models. Yeah. More complicated tasks, we do it's the thing is the cost of intelligence is coming down. Yeah. Right? The cost of these tokens are not because you always want to go with the best And

Speaker 1: the value of networks is gonna keep going up.

Speaker 5: Feature feature requests. I wanna be able to go on astrocade.com and navigate based on like popularity of different games by year. So like I wanna find all the games that were popular when I was like five Oh in like 2000 like 2005, right? I remember some of these like web based games and were like, what was that game where you're on a little bike and these Line Rider. Line Rider, you know. Rider.

Speaker 1: We didn't

Speaker 10: have that. Actually have something similar

Speaker 1: to that. Yeah. But I mean to the point about the recommendation algorithm, like is the user trained to search for that and then is your search algorithm going to match with that? If it's an abstract representation and the user didn't put the keyword line writer in their description or something. Interesting challenge. Interesting challenge.

Speaker 5: For the rest of the show today, I'll be playing

Speaker 1: Yeah?

Speaker 5: Line Rider. So if you want to keep if you keep watching after you leave, you'll see I'm completely distracted.

Speaker 10: It's actually it's pretty calming also. It like puts you in the flow state and

Speaker 5: You know what else is calming?

Speaker 9: Rolling flashbacks. The

Speaker 1: flashbacks not calming. That's extremely disorienting.

Speaker 5: It's very calming to me.

Speaker 1: Say you Just want to play tell them remake COD.

Speaker 5: Very common to me.

Speaker 10: That was true. That was true.