You always want to be an early adopter. It's basically like the saying goes, early bird catches the worm. If you get to use AI agents, right, and use it early, you're going to be having a leg up against all the other people who are still in chat GPT mode. Effectively, agents have made us experts in everything. I've never been as busy and productive as before because of AI agents. I didn't get to productive, you know, let's call it 50 agents that are really productive for me by just saying, 'Hey, make something for me.' It took work to make it work. Start using this technology. There's a big gap in terms of the opportunity that you can leverage from that.
Yat, welcome to the show. Thank you so much for being here.
Thank you for having me. It's a great pleasure to be here.
You've been on the cutting edge of tech innovation ever since the start of the internet, right? Early to the internet, early to mobile, early to crypto and to web 3 essentially. And I kind of want to know what skill set you had to see how transformative those technologies were before anybody else. Was it right place, right time? Was it you just you could see the future? Was it a bit of luck? What was it that gave you the ability to do that at the time?
Yeah, so I think in all of these phases, including the most recent one where we talk about AI agents, the way that we were thinking about the internet and the way we think about technology always came from the lens of how we thought it would make the world better. And the other thing I also keep telling people is that the most successful and the most impactful entrepreneurs in the world don't follow a tide, they make the tide as it were, right? So in other words, and it doesn't mean that it always takes shape in that particular form. Like when you think about the more recent example of blockchain gaming, NFTs, and so forth, we didn't actually sort of imagine it to happen the way that it did. We thought it'd be a big thing, but why do we think it was a big thing? Well, because we figured that digital property was going to be valuable. And I think we all agree that digital property is important and that we have a lot of people spend money on that in games and therefore we should be owning these items. But what shape and form it would take, frankly, we didn't know at that moment in time, right? And then it needed DeFi summer to happen and then it needed all these other things that develop for the whole thing to sort of explode the way that it did during that time. And similarly back in the early days of the internet, I mean, one of our most popular products back in 1998 and 1999 was email. And of course, in the beginning, email was very difficult to sell because everyone was like, 'It's kind of clunky and honestly speaking, fax machine does it just as well.' You know, you had to go through that phase. But we knew email was faster and the image back then wasn't around, you know, is email going to be the future, but rather the image was very clearly around what is a technology that can make the world communicate better broadly and therefore basically just have more impact as a whole. And that seemed to make sense, but you know, did we see instant messaging the way that it developed? Not quite. Right? So, we knew it was communication, but we didn't really see the whole pattern around the fact that it was going to be sort of a form of WhatsApp or Messenger or Line or all those type of formats, but it ended up being again similar in style, right? So you could say that the two perspectives is really that we want to shape the world the way that we think makes sense and I think that's true for everyone, right? You have a vision, right? And then of course market forces and everything will drive it towards a different direction. And then I think the final piece is that you just have to be open-minded and really, really flexible, right? So if you have a vision and it so happens to be that the world is moving in a different direction then you basically sort of move towards that. I mean, it's hard to imagine that in 2021 and 2022 there were blockchain gaming companies that actually didn't do well, right? But it did happen because they completely ignored DeFi, right? They were the gaming companies who didn't do DeFi. They were just like, 'We're only going to do this and we're not going to care about tokens and we're only going to care about gaming assets.' And that didn't work either, right? So it could have not worked out as well.
Just on gaming then, in 2021, 2022, the bull case for crypto gaming was so obvious, so clear, and it just hasn't really played out. You know, maybe it's because GTA 6 isn't live yet, maybe it's because the best games are just not on chain yet, but do you still see that broad gaming thesis as being correct, I guess?
Yes, so I do think it's kind of pivotal. It probably won't take the shape the way that we thought it would, but it's still going to be pivotal. And there's a couple things to think about. You know, when you say GameFi, for instance, I think one of the things that we need to sort of really reflect on is that actually what's happened in most of crypto is it isn't so much that we've basically brought finance to games, but rather finance became gamified. Right? So in that construct, you could say if you look at Polymarket, Polymarket is effectively a game for the most part, right? I mean, it's taken a different shape but it's basically taken a certain type of gamified finance and taken that attention. But the other thing of course we have to also remember is market circumstance in 2021 and 2022. What was the big thematic? It was COVID. What do we say? All of us are just going to live virtually. We're never going to fly anymore. We're all going to be at home doing this, basically just having video conference calls. And I want to remind people that gaming is roughly an $8 billion industry, blockchain gaming is about an $8 billion industry at this moment in time. So it's not nothing, right? NFTs still do around $250 to $300 million of sales a month. Okay? So again, it's not nothing. Five, six years ago or six, seven years ago, this was a zero industry, right? So again, it's all relative. And we could also look back at the early days of Zoom's market cap and say or Peloton's market cap and say, 'Oh my goodness, look at that.' You know, it's but the reality is that these are still valuable companies. Bored Ape NFTs are still 10 ETH floor price, right? So it's all kind of relative as to how you look at it from that historical perspective from when it started. But you know, I guess the one thing that we were all expecting or hoping for is that we're going to go be playing the most popular games in the world and going to have blockchain technology in the back. And I do think there's a couple catalysts that will drive that. And then I'll get into the other piece around the agentic piece because I do think that's one of them. The first one is that gaming generally has been under siege. I've said this before, right? But really, you think about it, the one thing that's really taking away attention from gaming hasn't been broadly crypto. It's been TikTok, right? A lot of the sort of entertainment has been going in that other direction. So actually if you look at gaming stocks overall and gaming performance it's not been as brilliant as it has compared to other industries. So that's kind of one macro factor. It hasn't really grown, right? The second factor is the fact that we have basically now with agentic AI, actually, we joke a little bit about the metaverse, right? And people say, 'Ah, the metaverse is dead,' all that kind of stuff, right? But actually, the way that I would phrase it is we may have made a mistake in thinking that we looked at us entering the metaverse, right? In other words, the goggles that we'd be wearing and sort of the interfaces to immerse ourselves into the metaverse, when in fact actually what AI agents are proving today, and that's kind of why we're so focused with Hello Minds, is that actually AI agents are stepping out of the metaverse into the physical world and impacting our life. And for anyone who has managed more than 20 or 30 AI agents, actually it feels very much like a game. So actually we think that's going to be sort of one major shift as well that's happening broadly speaking. And finally, the other thing is that, from a niche industry standpoint, gaming seems small, right, at around sort of an $8 billion industry. But, you know, there's other sectors that grow as a result, like strategy gaming, for instance, things like real-time strategy type of games, for instance, even card battlers. They started as small niche industry games and then they started to become more popular. And the final piece I would say is clarity. So we've been talking about clarity for a while. One of the challenges that many companies have had is that when they want to tokenize something, whether it's a virtual asset like a gaming asset or even tokenize their virtual gaming currency, it's always been considered kind of dicey as to whether that could be considered a security under the previous administration, right? We have to remember that up until maybe a year and a half ago the entire crypto industry, everything related to tokens was under siege, right? Especially in America, right? So nothing could happen. And so when clarity finally emerges, which we do believe will happen at some point hopefully, then a lot of companies will have clarity, no pun intended, to say, 'Oh, if I launch a token this way, I'm going to be safe, right?' Regardless of whether it's about raising money. It's just about whether they are even allowed to tokenize, right? And I think this is where we're going to see a trigger point happening. And so we were obviously all hoping that clarity would happen sooner for a lot of reasons we might not have to discuss here. It has not yet, but I think that's going to be another big one.
Just on clarity, actually, I mean, I'm interested to know, you seem relatively convinced it's going to pass. It's only 50/50 on Polymarket at the moment. I mean, it seems like the runway is rapidly kind of like running out. I mean, have you got any information on that that potentially we're not privy to? And with the clarification that it should provide to token issuers, you know, companies that are issuing tokens, can you envision how that token structure needs to change or what is likely to be the case for things that would be deemed as okay and not okay?
Yeah, I mean, I think under the act obviously, right now the way that it looks like is most tokens won't be considered a security, but the key real aspect here is that now you're able to launch a token under a certain framework, especially if you're launching it for the purposes of proof or maybe even if it's used as a kind of gaming-related currency that could be traded but not necessarily be sort of used as an investment vehicle, for instance. Like these are things that people have tried before but no matter what you did it was like, 'Oh, token can trade, security,' so basically you get hit over the head. Whereas with this, it allows a game company to experiment. In particular, I think medium-sized companies are more prone to experiment than the large ones. But the large ones, this is the thing, the large companies are actually looking at this saying, 'This is interesting, we're just not sure what to do yet.' And nobody there, it's just too much to lose if you're out there basically trying to experiment with a token that you could end up getting in trouble with because there's no clarity, right? So once that comes out, it allows people to start actually really sort of, because it's not just are the companies open to it. It's also other companies and individuals inside the companies going, 'Hey, now that it's possible, here's a framework.' You know, your legal department isn't going to be all over it saying, 'Stop, stop. We can't even blink token, right? We can't even think about it.' Now you're able to sort of say, look, because the technology of blockchain and tokenization, I mean, everyone talks about RWAs, right? Well, you know, when we talk about real world assets, I can assure you that for our children, gaming items are as real as they are to basically how maybe the adult population might think of physical items, right? It is very, very valuable and important for them. We spend hundreds of millions a year on these virtual assets that are very real to those users. And so at minimum we should be able to tokenize them so that they can own them regardless of whether you trade them or speculate on them, you know? And I think nobody objects to the idea. I mean, every year Blockchain Gaming Alliance goes out and makes a survey. What's the number one reason why you believe or you like blockchain games? It all comes down to ownership, right? So, at the end of the day, nobody disagrees that ownership isn't a critical item. And we talk about 3.6 billion gamers in the world who ultimately, if they were given the choice to own something, would probably prefer it. I mean, even look at Pokemon cards. I mean, Pokemon cards started as something funny and now it's a very valuable asset and I can assure you that it will be the same for gaming items. Already in the black market, on eBay, on Amazon, you can see that. It's just that with Pokemon cards, look at today. I mean, it didn't happen overnight. It didn't happen like, oh, in the 80s and 90s, it exploded the way that it did. It's really a phenomenon that interesting enough became really quite big over the last couple of years, you know? And then because of tokenization, by the way, which is one of the more popularly traded NFTs, the value of Pokemon cards has increased because now there's a market and there's transparency and there's clarity and all the things that blockchain is good for. Now quickly on clarity itself, obviously I don't have any inside information. So you can call me a relentless optimist on the matter. However, I would say the key thing is that I don't think America wants to fall behind the rest of the world. Yeah. So our joint venture, Anchorpoint, just recently received the license for the dollar stablecoin. We passed our law in Liechtenstein last year a couple weeks before Genius. And look at the US, Genius Act was passed last year and the stablecoin licenses probably won't emerge until next year. Clarity is obviously one thing that has to be resolved, you know, and the rest of the world is doing things as well. So there is a competitive pressure that's coming that isn't just about what's happening in the US but also in other markets. They're going to see that as well, right? And then I think there's going to be many, many aspects. Of course, there's horse trading and then there's negotiations, but I would say there is a sense of urgency that Clarity Act does need to pass, but there's many things that people want in between. So, I don't think it's a case of before where you had a war camp that was like, 'We can never have this, we should never ever have this.' I don't think we're there. If it was like this, I'd be more skeptical. This 50/50 scenario is really because the parts of the horse trade as to the Clarity Act is obviously uncertain, particularly things related to the ethics part is probably the biggest stumbling block, right? Ironically, however, in terms of the substance as to requiring a bill that requires clarity and supports America's role in being a leader in crypto, I think there is more support than not.
Awesome. So you said in the past you think blockchains are built for agents, basically. And obviously you guys at Animoca are now focusing, I mean, you've gone from being gaming focused to basically everything crypto focused, right? Animoca, you know, like the like figure markets, right? It's like, you know, onchain RWAs, agents are essentially the, well, that's AI is essentially pulling up the entire of the stock market right now, taking everything along for the ride. And this is clearly where the most innovation is happening. Arguably, it's one of the reasons why crypto has been so kind of like lackluster, you know, for the last few years because all of the attention has gone directly into AI and AI-adjacent stocks. Whether that's people building GPUs, whether it's people developing, whether it's even in the buildout, you've seen this all trickle down through because the reality is for these models to run, you need so much infrastructure. It's arguably a lot more beneficial to economies generally than just like software doing well. You kind of, this needs loads of infrastructure, right? It trickles down to transportation, data centers, real estate, everything. Yeah, exactly. Right. You've got basically everything all tied up in what used to just be sort of like hoovered up by just these few software deployers. So I guess like twofold really. Why do you think blockchains are built for agents? And where, if blockchains are built for agents, does that value end up accruing to?
Yeah. So first, when we think about, you know, maybe if I ask you this question, right? How many AI agents do you think you'll have in the next 3 years?
I mean, 50 to 100, I guess.
Okay, so I mean, it's interesting, right? When I asked this question to people like literally 3 months ago it was like 3 to 5, right? I mean, I have over 230 myself, but the point is that it's going to be way, way more. It's going to have to, it's way, way, way, way more than sort of plus one to the human. In fact, it's probably going to be more than 10, 20, 30. So we're going to have hundreds of billions of agents. That's, I think, I don't think anyone will have strong disagreement to the fact that we'll have at least 100 billion agents, probably more in the future. Okay. So fine. So you know, how do you, and then we talk about autonomous AI agents, right? We all agree that they have to be really autonomous. So how to make them really autonomous is we have to give them some kind of power, right? Beyond taking an action or what, but what is needed to make them powerful? They need to have money, is power, right? The fact that I have it means I can do action. I can't be truly autonomous if I can't pay because everything is gated through payment, right? And so are you going to give your AI agent a credit card? I mean, you could do that, but are you happy to have the other side spend 2.5%? Is it designed for that kind of microtransaction? I mean, we obviously know the answer is no. A lot of people outside don't know that and don't understand that.
Just, sorry, just on that one point actually, I just want to play devil's advocate because I think it's not very hard, right, to spin up a card that just goes, 'Here's my agent's card, it's got access to spend, you know, $200 or it can spend 20 bucks a day.' You can just spin up, I'm sure Visa, Mastercard, etc. are thinking about this, right? They're thinking about, okay, how can we capture this market? And I'm also sure that they're probably thinking, okay, do you know what, 2.5%, there is innovation technology that's going to pull market share from us if we don't actually try and capture this. Like if you think the vast majority of transactions are currently done via these issuers at the moment, it's much easier to go and say to my dad, 'Hey dad, look, there's this thing that will just help you out a little bit.' And all you've got to do is just tick this little box here that says you're happy for it to have access to, you know, $100 a day or whatever the figure is. It's very easy to sort of put that in a box and, you know, what if you lose it and something goes wrong, they're going to reimburse you. What I, like, I'm blockchain of course as can be, but I just sort of like when I'm thinking through this I'm thinking, okay, it's still not a direct, it's still not a layup, I don't think. There's still things that need to happen before that. Could you go through those, please?
Yeah. So I mean first of all, the one important sort of factor to consider is that any kind of Visa and credit card system is a centralized banking system, right? And so yes, you could see, to me I see these as sort of gateways into it. So I'm not saying it won't happen. However, this is the gateway that's in there. At the end of the day your agent is going to do things better, faster, cheaper for you, right? They don't care that it's Visa and they don't care whatever. They're going to find out, okay, how do I save you 5% or 10%. How do I make it faster? And I think the other thing that people often forget is that, you know, the AI agents and blockchain is designed for microtransactions and the Visa and Mastercard and settlement system is not. I mean, they're not designed for 0.01 fractional sort of divisional payments. They're also, by the way, unable, beyond the payment, because I think we're all looking just as payments as a transfer between point A to point B. We're not thinking about collateral. We're not thinking about all the elements of DeFi. We're not thinking about the fact that, hey, I want to get credit. I mean, this is the other thing. How does an AI agent get credit to basically do business with another AI agent? Oh, let me see what assets you have. Let me have a proof on that. Or can I put this into an escrow? All right. I mean, so I think a lot of people who make an assumption that we're just going to use Visa and Mastercard completely ignore the programmability and smart contract element that allows all these new dimensions of money to take place, right? Because one thing that will happen in the future, and this is what Visa Mastercard doesn't do. The reason by the way we have these insurance policies around Visa Mastercard is because if someone does fraud, you basically can say, 'Oh, okay fine. I can withdraw that.' Then you go to the police, you do all that kind of stuff. You know, by the way, I've done this many, many times and it's not as fluid because I have to go to police and I literally have to go through every statement item and that kind of stuff. Sure, I'll get my money back, but it takes a long period of process. However, on a blockchain basis, I can basically spin up a smart contract. I can hold it in escrow before the transaction happens. I can prove the asset is there, right? While I do the transaction, and nobody can basically do something. And this happens in real time and super fast. Like that validity will happen much, much faster and with much more certainty than you would have for something like Visa Mastercard. So there's all these elements and again, as humans, I think when we look at the lens from a human we say, 'Oh my goodness, but I don't want to verify the smart contract part, I don't want to take a look at all the escrow, that seems extra verification.' Remember, for an AI agent, this is basically something he just does as a parlor trick, okay? For an AI agent, basically analyzing what's on the smart contract is normal, right? Having 10,000 conversations with different agents is okay, right? So often when we take that lens we think of it from the perspective of our own human limitation. Now another example, a lot of people think like, 'Oh, you know, we're not going to disintermediate the major platforms.' We're absolutely going to disintermediate the major platforms because the major platforms have always been able to basically capture our attention out of convenience and through the network effects that were built, like the App Store or Steam or Google, you know, we go there because we're used to it and because we've got the network effects. Do you think an AI agent cares about the network effects inherent? If I want something to buy in the cheapest way, he's going to find it. He's going to find it no matter where it sits all the way over in Timbuktu, for lack of an expression. If he can save you $10 to do that, he will do it because he doesn't care, right? This is something that a human will probably take hours to do. So, it's not worth his time. For an AI agent, it's probably seconds of work and it gets faster and faster. So, the entire framing is going to change. And that's why, for payments as well, this is the other thing we do. We expect there's going to be one payment rail. Absolutely not. There's going to be thousands of payment rails that are going to be interoperating with each other, right? In the same way that DeFi has with their pools. Everything we've seen over the last 6, 7, 8 years in terms of DeFi and blockchain interoperability and scalability and stablecoins, these were all the prep for this machine economy to come, right? And now we're able to do it, right? Which means, by the way, you can make a bona fide payment with your silly memecoin if you want and the other side will never have to know it's a memecoin. He just receives it, okay? There's no risk that he takes. The AI agents negotiated in real time, he basically gets a transaction, he swaps it out in some kind of DeFi pool, he gets his whatever currency he accepts, whether that's a US dollar or whether that's something else, right? And in the back, this composability just took place, right? Which is the power of blockchain, right? But humans can't compose fast enough to make sense of it. That's why we're like, 'Oh my goodness, all this blockchain is so difficult for us,' right? But for an AI agent, no problem, right? Throw me 100 tokens, I can do whatever you want with it, right? So I think Visa and Mastercard are frankly, I mean, by the way, these guys are looking at stablecoins too, right? And there's a reason Stripe is doing a stablecoin. Why? Because just, right? So, the biggest companies in the world all realize where it's going. So, it's silly for us not to think that. Quickly on the value capture, first of all, Ethereum and all these blockchains, sort of other like Solana and L2s, these are all great and they're important and they basically help to scale the industry as it is, but my mental model is a little bit like telcos, right?
A little bit like what? Sorry.
Telcos. Telecommunication companies.
Yeah. Right. And so if Ethereum is like a telco, then Arbitrum is like an ISP, you know, basically something that sits on top of the infrastructure and delivers something at the front and they're important companies. You know, we still need internet service providers and we still need basically AT&T and British Telecom and all these companies to provide the infrastructure, but the value capture that they have is not as large as the companies that provide the front end because they own the end user. At the end of the day, yes, there's a composability aspect and there's a marketability and there's a fact that I can reach you as a customer if it's on chain. It's like an open infrastructure. However, the value capture won't be as significant. It will still be there. So does that mean that Ethereum is going to be super valuable? Yes, it will be valuable in the same way that AT&T is super valuable, right? Or British Telecom is super valuable. It's all relative. However, in comparison to the Googles or the Facebooks or the Apples in the world, it's smaller because they capture more of the value. I think in blockchain, we have some great examples. If you look at Polymarket and Polygon, I mean, Polymarket couldn't have come to life if it wasn't for Polygon, right? I mean, they needed the blockchain aspect to be able to build and have all this infrastructure that made it possible for essentially Polymarket and the thousand other wannabe prediction markets that are emerging every day because precisely blockchain makes it easy, right? But you know, Polymarket is worth what, 8 billion or more, and Polygon, well, we don't have to talk about that, right? I mean, but again, Polygon is important, right? And it's valuable infrastructure, so don't get me wrong, right? But it's just the relative value capture is always going to lag behind essentially the ones on the front end, which has always been our thesis, which is also the reason we've never really launched our own L1, L2 because we like to build on top because we see very similar to the internet, right? In the early days of the internet, infrastructure is important, but at the end of the day, the value capture ended up being the ones who ended up owning the customer, which is typically something in content or culture.
Yeah, all amazing points actually. I think 90% of all blockchain value is in the infrastructure layer and only about 10% is in the application layer at the moment. So there, and you've seen evidence of the fact that this is probably like a mismatch, misallocation of capital. Essentially, it is much more likely that you or users use applications. Whatever the infrastructure is that sits underneath is simply there so that applications can be built to capture the end user. And blockchain right now is priced, well, the vast majority of crypto right now is priced such that that value is going to trickle down to the infrastructure layer, which, you know, arguably the last sort of like four years have shown that it's not.
Yeah. But you could, yeah, but maybe just to counter that point, I actually think the argument around sort of the value capture, I think it's more forward value. And I think the forward value made the assumption that onchain usage was going to be much bigger than it is today, which it isn't. So I think the last four years, the reason why the market has been rather flaccid, as it were, is because of the fact that onchain usage hasn't grown. I mean, just think about it, the fastest growth, the biggest growth we've experienced in onchain in one fell swoop was 2021, 2022 in the NFT metaverse boom because we actually had a lot of people go onchain, right? And more importantly, they were not financial types for a lot of them, right? I mean, and I think there's a distinction that's happened, right? Those guys were there for culture, they were there for fun, they like playing games, right? And then of course you got the money guys coming in as well which obviously added another flavor to it. But then from basically late 2022, with FTX, all the way until 2025, which is basically three plus years, we've had a very, very painful drawdown. First with relentless persecution, a lot of scams, a lot of problems, right? And unfortunately last year I think everyone was hoping that Trump would make things better for crypto and in many ways he has made things good but in many ways he has made things bad as well. He's polarized the industry more in a sense that if you're pro-crypto, you're pro...
Trump, which means that if you're not pro-Trump, then you're not pro-crypto. That's a distinction that didn't happen before, right? And that is sort of a new era. And again, it's not fatal. It's just a new challenge that has sort of emerged. And on-chain usage has barely grown from the 40 to 60, maybe 70 million usage. Meanwhile, token holding has gone up to almost 700 to 800 million users. So, we're almost going to reach a billion people who own some kind of crypto, mostly Bitcoin and maybe the majors, right? But that's only speculated. They're not using it, right? If you hold crypto in an exchange, or as an ETF for that part, you're really not, you can't use it. You're only there to hold and speculate on it. Whereas, what we need to see for this market to grow is on-chain usage, right? And that's what we saw in '21 and '22. And that's what we haven't seen for the last three years for all the reasons that I had outlined. And I think AI agents are going to change that, which is why we're sort of betting so heavily on it because they're the ones who need on-chain usage. They're the ones who can actually benefit from that composability and they're the ones who are going to basically drive the industry forward. Of course, to serve humans, but at a very different ratio.
Okay. So what gets more users, agents, you're saying? I think I agree to be honest, but with regards to value capture, right? So right now it feels like something like Ethereum is already the 50th largest asset in the world, it's already very expensive. What kind of agent usage is Ethereum currently priced for, do you think? 100 times more than it is right now? 10 times more than it is right now? What would be fair value at this level, do you think, at the 250 billion or whatever?
