Rollup TV: Money Moves Fast EP #3 & zkVerify

Transcript

Heat. Heat. [Music] Heat. Heat.

[Music] Heat. Heat. Heat. Heat. N. [Applause] [Music] Heat. Heat.

[Music] Heat. Heat.

Heat. Heat. [Applause] [Music] Welcome back to the Rob TV, baby. It is Tuesday, September 30th, the last day before our Q4 mega rally.

Welcome back to Rollup TV, your daily show for everything you need to know about crypto. There is so much to talk about. We're running a little bit behind. We got our first guest ready to come in. We've got Ran Goldie of Fireblocks. He's going to come in and talk to us here first on today's show. Guys, welcome back to Rollup TV. Looking forward to having our premier payment show brought to you by Polygon, presented by Polygon, one of the leading payments chains in the crypto space. Today we are going to be talking to Ran Goldie, VP of payments at Fireblocks.

Ran is a new friend of the show, uh, his first time on the show and so looking forward to having him up and talking about the all things Fireblocks, payments crossber remittances stable coins. looking forward to having a good episode. So, we got him in the back ready to go. We're going to take a quick pause, guys. We're going to jump into Ren. Then, we're going to jump into John from ZK Verify, their mainet, and then we're going to have some fun. I I got some stuff that I want to talk about w with you guys. So, let's take a quick break here, and then we're we're going to flow right into the show with Rand.

So, we'll see you guys in a sec. Welcome back to Money Moves Fast, the

premier live show for st for payments, RWA's tokenization. I am here today with Ran Goldie, the head of payments at Fireblocks, amongst other important roles in his career. This show is presented by Polygon, one of the leading chains in payments for crypto. Uh, tested battle tested fast cheap uh with a big upgrade coming soon. I know Fireblocks and Polygon are close friends as well. Ran, welcome to the show. Glad to have you on. >> Thanks, Andy. Love to be here. Love the energy. This is probably one of, uh, I would say my most viewed shows. So, thank you for having me.

>> Beautiful. Beautiful. Well, Ran, it's your first time. Uh, it's good it's good to meet you as well. Um, I'm glad that you took the time and yeah, man, just a quick, you know, intro about how you got into your role. You have quite the, uh, quite the background and I think it'd be helpful just to get some context. >> Yeah. No, so um, I'm I'm, you know, I'm I'm basically here because, uh, I've got sucked into this world about 10 years ago, uh, into, I would say, the world of crypto by, uh, by E Toro's co-founder and CEO. He was the one that told me to buy my first Bitcoin before I knew what it was. And Uh I I fell in love into this and I before this I was in I was in finance. I actually took one of my companies public before I fell in love

finance. But this was obviously different. Uh so I started a company that was doing payments over blockchain. That was 2017. You could basically tell a taxi driver you're doing Bitcoin. He will like give you money over right in there. And that was like the easy part. But then obviously crypto winters, some other stormy weathers. And at the end of the day, we were able um I think to create some really cool tech around payments and got acquired by Fireblocks about you know four years ago. Ever since I've been leading the payments and network divisions with Fireblocks uh which has been nothing short of amazing obviously because the last four years I think that you know crypto and digital asset have been payments right especially this year with stable coins

where stable coins are pretending to solve world hunger as well right so so it's been a great journey >> yeah yeah that's awesome no I think uh the entrepreneur to the uh head of a a uh you know branch which is a bit of a different transition, but it it sounds like it sounds like it was fruitful and it sounds like you you guys have built some really cool stuff um you know back in the day. So Fireblocks has had a very interesting journey, right? Like if you think if you think back to like 21 and you know 2019 kind of cycle, Fireblocks was like the enterprise uh kind of network uh category that was kind of hot in the 2018 bare when security tokens were kind of a thing and people were

trying to figure that out and custody was a big thing. Uh then Firebox kind of went back under the radar a bit. I feel like with the L1 cycle of 21, 22, 23, everyone was kind of gung-ho about all these other kind of crypto apps, DeFi. Now, Firebox is right back in the center of attention. People are teasing an IPO. You guys are absolutely crushing it on the revenue side in terms of your custody business as well as your other offerings. Um, and so it's kind of been a it's been an eb and a flow for the last eight eight-ish years of Fireblocks. Um, and recently you have just launched what you guys are calling the Swift for stable coins network. Um, so would love to hear about this and really just would love to hear about Fireblocks's thesis as a whole for the payments kind of space in crypto and

kind like generally how you're thinking about this space as it's evolved. >> All right, an easy question then. Okay, fine. So, let's break it down, I guess. So, you know, first of all, Fireblocks, I guess, consider itself first and foremost as you know, an infrastructure for this world, right? So if you're a payments company out there or you know a fintech like Revolute or New Bank or like I said a payments company like Wallpay or Visa whatever and you want to build on this rail that's called blockchain and again you know our view is this is another payments rail right there's there's a there's wires there's faster payment rails out there blockchain is another payments rail and if you are in the business of moving

money that's a rail you need to get connected to same way as I tell you if you're a bank you need to connect to you know the Fed or you need to connect to Swift whatever. So first of all that's our thesis and our thesis is that over time all of obviously the payment companies but all of the fintexs and eventually all of the businesses by all these layers that will be enabled every business in the world will use this tech and again I I would say that my personal thesis is you know this is this is very similar to what we have in I don't know communications right like we're still having a phone call I can call you Andy from my iPhone uh but I've been I could have called you also likeund and something years ago ago when Graham Bell invented the phone. But now we have

better infrastructure. I can also text you. I can call you with you know like video. I can send you very stupid memes on iMessage. I can do a lot of stuff that I couldn't but I can do that because of infrastructure. We could just do more things. This is an infrastructure upgrade. So that I guess that's like part of your question was what's our thesis? Um specifically about the network. You know, the problem with a lot of these companies that are coming into our space, especially in the last 12 to 18 months where there's a large, you know, very big momentum due to regulation, other stuff and more incumbents are coming in, more large payment companies, banks are coming in, right? Um, we've probably been talking to more than 200 banks over the last six months that are coming.

>> They all want to launch stable coins, don't they? >> Well, we tell them that they shouldn't, right? Like, uh, and and this really bums up people. uh like the first question is what's a stable coin? Second question, okay, should we issue a stable coin? Uh but but no, then they actually form a good strategy. I hope um and um that the thing is that a lot of these companies are coming in and you know they now want to use stable coins and send a payment to Mexico and send a payment to Africa and they have no idea who to work over there. And because Fireblocks as you said has been you know we've sort of been in the right place in the right time with the right tech and 70% of the world's institutions runs on Fireblocks. Um you know we run 15% of

all the blockchain transactions globally. We there's this year it will be about $5 trillion. So we run a lot of the infrastructure of these offramps in Mexico in Singapore in Brazil and whatever. And what we really did is like bring all of our, you know, great partners and clients together into this one network and tell every client that comes in, hey, you want to now use this magic that called stable coins to move money faster. Great. Here's a catalog of everyone in the world in a 100 countries you can actually use and send money through. So that's the idea of the network. And again the idea is that it's open you know like if if by the way if swift wants to connect to the network

come into the network right >> and when you say network it's we're not we're not thinking like consensus based node architecture here like a blockchain we're talking like a more distributed ledger kind of network or a kind of more of like a database network just like you know how like what exactly does network mean in this context? Yeah, I think that's a it's a really good good question, Handy, because everyone right now is either issuing a stable coin or a layer one, right? >> Yes. Or or both. A stable coin layer one call, which is exactly what we need more layer ones. But um no, it's it's neither of that. It's basically allowing uh Fireblocks to because all of these people run on Fireblocks then from within your you know workspace within

your API whatever you could access all of them programmatically uh over any blockchain whatever you want whatever coin you want to move through that but because they're all on the Fireb network there's a basic trust element and a basic way to communicate and move messages around and really create these payment flows end to Right. So, no, it's not a blockchain. No, it's not a protocol. No, it's not um it's not DeFi. It's it's it's just a um a simple network you can, you know, in a way mimics how Visa connects between different merchants and acquirers and whatnot.

>> Yep, makes sense. So on this kind of note around people wanting to launch a stable coin and people wanting to launch L1s, you know, it does kind of seem like Fireblocks becomes an important almost like middleware or part of the infrastructure stack for a lot of them. Um, but also I think there's a point to be made with regards to like just general competition. And so like Cloudfare for example just announced their CEO comes out and says look there's a net dollar stable coin for aentic web payments coming out. Like this is a thing that we're going to be pushing. Agents are going to be interacting with each other in the in the wild using stable coinbased payments. Like this is the there's no credit card. There's no, you know, three-digit code on the back of their

card. There's no expiration date. You don't have to KYC. You can just use an Aentic wallet or a Gentic, you know, account on chain. So I'm curious if you see, you know, in this example like Cloudfire, you know, stablecoin or other big tech stable coins as a threat to what Fireblocks is trying to do or do you see this as an opportunity for you guys to grow? No, I think first of all, I think that's awesome that these companies are coming in and again, you know, whether they're issuing as as we just make, you know, laughed about a stable coin earlier one or if they're issuing AI agent SDKs to do payments with stable coins. I think all of that is great and I think that, you know, we're using

Cloudflare and uh at Fireblocks. I'm sure if Cloudflare scales and needs to issue a billion wallets like what we do for a living, they'll probably go to a company either Fireblocks or someone else that that's their main business. That said, that just the fact that they're issued that stable coin and you know that machine to machine AI to AI agent to agent payment is probably going to be something we're going to see a lot of, right? Like people are now building these MCP servers and I don't know if you if you got to cover this in the show already, but like this easy way for AI to communicate with your website or with your app, whatever. A lot of people are

writing those things. These are called MCP servers, right? So there's really no great way to monetize those or to make a living if you've built a really good agent or a really good MCP server. But but now with these payment SDKs that are coming out um it this could all change like and I think this will also roll out to Fireblocks's world in a way because there'll just be a lot more wallets a lot more assets to move securely and a lot more um let's call that end points to connect on a network that is secured which again talks to what we believe and what we do right and you know believe it or not we're spending 40 million a year just on security and we have 50 people

on the payroll that are either you know our hackers or other security devs sec ops people. So so this is like uh obviously this is not a hobby moving money securely. >> So I'm I'm sure that >> I'm very happy this is happening. I'm sure we'll find oursel part of that mix eventually. >> Yeah. Yeah. Cool. No, I mean I think that is kind of like one element of what's happening is more of this commercial enterprise entry into the stable coin and payments ecosystem from a lot of large players as we've already kind of alluded to and discussed. And then you've got like the government entry as well. You've got like world liberty USD one you know I think recently there was nine major

European banks forming a eurobacked stable coin consortium. um right and thinking about ways so I think like there's demand coming from both commercial and government level you know I'm curious from the government level perspective whether it be this European bank news or just generally the way that you know Bohines left the uh you know uh digital assets committee and joined USAT so like there's just this government fuel demand happening as well aligning with secretary Besson's $3 trillion of stable coin supply by 2028.

You know, when it comes to the government creating demand for payments and for uh transactions on on blockchains, how does how does Fireblocks capture that? Like how do you make that you know regulated demand something that comes to Fireblocks? >> Wow, Andy, how do you think of all these texts? Like I've seen the write up before the show. It it contained 10% of what you just said. Like is your brain connected to Twitter directly to a prompt generator? How >> we do this a lot? I'm actually trying to make my question shorter, you know, >> that is crazy. That's a lot of knowledge in one question. Um, look, I mean, so first of all, let's talk about the bank

thing, right? Banks are obviously not a big fan of stable coins. I don't think it would shock anyone because as as a business obviously a stable coin is not good for a bank. They cannot hold your money one for one as a reserve because that prevents them from rehypoccating it and making money on loans and interest which is fair. That's their business model. Um and but they have to play this game right now. U they have to play this game because as I said it before they have to be connected to this payments rail. So right now with stable coins I think that will evolve over time. By the way, I think that you know what what is that we're using today this this dollar pegged asset I think it will change in nature maybe to a

tokenized deposit maybe to something else whatever now obviously what everyone is wondering and you and me are sort of joking about is why is no one putting an emphasis on interoperability right and then comes this consortium of banks and says hey you know what like we're nine banks we've created this eurocoin coin that's interoperable between us at least which I think is great and I think that over time what we'll see is probably more coins become interoperable our >> our way to move that is obviously as as as we provide the infrastructure to a lot of those we're we're telling them listen here's your chance because you guys are all running on the same infrastructure let's build this

interopable from inception which a lot of them do so so that's like our also in a way >> you uh push towards that that path which I hope more we'll see but what I think is sort of happening right now is you know the banks are really experimenting with stable coins in a way it will evolve probably within a year or two to something they can actually use in a business model um and then the government at some point where the banks have tested this enough it will probably reach the place where you know we all thought it will be there but like the governments themselves right and what we've seen up until now and you know

governments are not you know the US government has been holding Bitcoin since 20 I think 13 probably even before like there's people in the government they know what they're doing with digital assets um and up until now they've been using tools right they've been using companies you know like like us they've been using chain analysis I think that was very public right so I think as as this you momentum will continue from the from you know from the payment companies to the fintex to the banks later on to the government entities then I think all of all of the people who are now providing these type of services like Fireblocks like chalysis of the world will probably grow to be companies who serve governments just like Amazon by the way

>> grew to have an a a government cloud >> right um and and by the way this is why >> I'm sorry to say this just like I'm not trying to shamelessly promote Fireblocks. I'm just saying this is why I'm excited to work at Fireblocks because >> if we if the world continues on this trend, >> then you will have probably 100x more subscribers >> and Fire Blocks will probably have 100x more clients. >> Yeah. >> And which means we're like 1% of the company we could be a few years from now, which is why I get up in the morning. So >> yeah, it's exciting.

>> I hope that answer all your questions. >> Yeah. Okay. And so All right. So that is that is like un like unpacking why there actually is a need for essentially what you this network this payments network that you're building because when you look at all these stable coins or these you know types of tokenized assets that people are trying to issue we've already seen them go and go to interoperability providers like chain link wormhole etc. Okay, we want to have the interoperability within the blockchain ecosystem, but what Fireblocks can essentially provide is what I would consider the the interoperability for everything outside of that. So,

bank-toank communication on and off ramps, last mile, you know, exiting into local currencies into local uh uh local, you know, commercial banks. Maybe you could just a bit further unpack the like the core the core kind of service that Fireblocks network brings because you guys are doing about $200 billion right now in stable coin flows per month. So, I'm kind of curious like is that is that like I'm trading on an exchange and then you guys are selling that or I'm sending that from Peru to Bolivia like like you know maybe kind of explain a bit of exactly what the core services and who that target market is and then we can start to think about like and understand

who's using and what's powering this uh you know current monthly flow. >> Yeah. So powerful question. So let's look at that $200 billion a month of stable coins, right? So this year probably >> will end up processing I want to say something like 2.7 trillion uh dollars of stable coins. Um let's take an example of 200 billion. 50% of that is attributed to trading >> and about 50% of that is attributed to payments. So when I say trading, I mean, you know, exchanges run on Fireblocks, OTC desks running on Fireblocks, whatever, hedge funds. That's that's all goes to that bucket. The other 50% are

real payment companies, you know, the like world pay, like Visa, like you know, bridge, you know, it's powering stripe, like those type of companies, uh, but also the B2B companies like Open Paid and Zodia Markets and conduit. And let's take one of those companies and really talk about a use case that will help also understand the narrative of the network. Right? So let's say we have I think uh conduit is a great example, right? So conduit is this company that helps importers and exporters uh pay large amounts of of dollars really uh for containers they're importing to Brazil, for example. So they would they'll be in Brazil, they'll get, you

know, um Brazil Royale from someone, they'll convert it to a stable coin. Now they're thinking, I want to work with someone in Singapore. Um I do not know anyone in Singapore. I do not speak the local language. I do not know what's going on there. But if I go to the Fireblocks network, I see there's like two providers that work in Singapore. Here's the currencies they work at. Here's what I need to do to open an account. I click I open an account and then I connect it to my Fireblocks world and I just send that uh that new provider let's say Zodia Marcus that operate in Singapore I send them the stable coin they then convert that back to Sing dollars and send it to whoever I ask them right so I can be a business in

one country and develop my payment flows my off-ramps or my on-ramps in different countries uh without traveling there just by because of file blocks network I can be introduced to those businesses and integrate with them with one click right so that's the magic of that network and I can when I do business with them I get all my reconciliation data normalized because it's all on the same platform I get all my trading data I get whatever just from the same platform I use for my daily treasury management so that's the concept of the network >> and would you say like that that B2B kind of crossborder payment flow is is where is driving a lot of the demand growth.

>> So it it's driving 40% probably of the volume. Another uh something like 25% is payouts to gig economy employees creator that has been by the way the largest growing segment within >> Okay. Is that like Uber and like Grab and that type of stuff or >> So it's I mean I know Grab does that uh but I'm not aware of Uber but a actually you'd be surprised. It's actually a lot a lot of these um businesses like uh Scale Scale AI that you know have like 10,000 payments they need to pay. They use Bridge to do that. Bridge uses Fireblocks. Boom. You get like hundreds of thousands of payments, right? Um, so

it's a lot of those type of companies, a lot of marketplaces, a lot of actually payroll companies starting to get on this. And I think the wave I'm seeing right now, and I think you'll see it over the next three months, I would say that this last quarter or next six months would be the remittance companies. >> So you've already seen Remittly making announcements. Moneygram uh that has done uh like wallet swift, cross mint. I think the remittance companies are getting on this train and you'll probably see a lot of them coming in and that's like a trillion dollar market. So yeah, it's gonna be interesting.

>> That's like Western Union is using stable coins instead of whatever they're currently using at scale in the future type of thing. >> Yeah. >> Yeah. That's Yeah, that's huge. Um Okay, cool. Uh one question about that though. when I am sending a stable coin from, you know, let's call it Brazil to Miami. My cousins are in Miami. I'm in Brazil, let's say this. Um, and I'm a first-time user and I'm sending a stablecoin to like a ox address or like a name in an account. Um, I don't really I like like I don't understand that this is an unreversible or nonreversible

transaction. like maybe it's my first time. I'm curious like Circle's been hinting at exploring reversible USDC transactions. I'm curious if you think like chargebacks or you know reconciliations or pullbacks from payments for blockchain or for in these kind of like fintechy apps is going to make a big, you know, comeback or a big entrance into the scene. To me, it seems logical, but I could also see it being like, "No, we don't need that. We can just get it right the first time." And, you know, because when when I go to Venmo somebody, uh, there's no like pullback option. Sure, on PayPal, if you do a payment for services or goods, you

could do a chargeback process. Venmo, maybe you could, but it it it's feels permanent in today's world already. So, maybe consumers are used to that. I don't know. Curious your perspective. Well, I I think you know I think here's here's like an uncomfortable truth, right? So, the reason that fiat works and most payments are done on fiat still and not on digital assets is because there's trust in that closed loop ecosystem. And banks, you know, whether you like banks or not, you like Swift or not, it doesn't matter. They have a way to communicate and say, "Hey, I'm going to send you money. Here's the here's the one who's sending you this. do you want to receive it on this, you know, to this uh recipient? They do their own checks,

whatever. Eventually, they say yes. You feel like it's taking a century maybe sometimes, but you know, it's safe. There's even like a way to reverse that if you need to, right? Um and when you do commerce, which is not the same as P2P, right? As you say, Venmo, you send it, it's done. You send to your cousin, it's done, right? Talk to your cousin. But if you pay a store, you do expect as a consumer like this is why we've created consumer protection. So I really think that actually bringing those, you know, schemes and uh let's call that even behavior protocols into a blockchain makes perfect sense because we want I mean you can you can say that banks are horrible, fine. But we I'm not

saying you're saying it, but like people can say that. But we want banks to run uh on stable coins as well or some sort of digital asset so we'll be able to move money 24/7. If we want that, we need to give this industry the tools to implement that and still have the precautions they need to operate. And and this time because they're using programmatic money and programmatic maybe forms of identity then maybe that whole process where a bank talks to one another and ask questions about the identity of the sender or the receiver maybe this time it can take like three minutes instead of two days because we've given them the tools. So I think part of this is giving them the tools.

So I'm actually all for it. I don't get invited to crypto parties but you know I can't hear you. >> Oh.

>> Oh, yeah. >> So, oops. Sorry. I just had a was edit that out. So, so it's almost like how uh Ethereum came on and took the idea of Bitcoin which and kind of made it programmable. We're basically looking at the payments landscape of stable coins and payments as they exist in like any any blockchain today and almost being like okay here's the tools that we have presenting this to you know the parties interested and then maybe thinking about like a programmable refund capability or a pro a programmable verification of a of a sender and a recipient uh rather than like doing like a chargebacks. So it's almost it's like creating that same level of interoperability that you would expect from Swift but just doing it in like blockchain infrastructure um to

create this kind of like interoperable functionality. >> Yeah, this is like you know Visa and Mastercard have their own rule book of how you can get you know a refund right? Uh Swift has their own rule book. Pix by the way has one as well as well as pay as well as everyone. this is how payment systems need to work because sometimes people make mistakes. It's fine. The fact we're introducing this to the blockchain doesn't mean that we're destroying blockchains. We can still have distributed ways to to to make decisions that prevent centralization uh if we want to, right? But it doesn't mean that we don't need to provide the right tools to build a reliable system.

>> Yeah, >> that's my that's my point. So, we're deeply in this neo bank meta is what I'm going to call it today, which is I'm a crypto company. I create a phone app. I allow people to earn yield on stables. I allow people to send stables and do form of lending in the app. Give cash back on a card. Have onetenth of the margins of of a traditional bank or if you look at it the other way, 90% better. uh 247 settlement. Stable coins with instant uh transactions and free payments. And we're putting this on an easy ass app.

You're just going to open it every day and you're not going to use your bank account anymore. And you can use the card to spend. And you just got to be topping up your Bitcoin or your ETH and then borrow against it. Get that money to spend. You don't have to sell your crypto and all of a sudden you're just completely outside of having a full credit card. You're pretty much there. >> This is a business that could explode. What's your what's your read on this? >> So, first of all, that sounds like um you know, wirex, cocoa, dollar app, cheaper cash, sinko, like a lot of other a lot of apps that are trying to make in a way revolute, right? Yeah.

>> That is that is trying to do this. the you know the the the reality is and this is interesting there was the first wave of fintex that obviously were built on top of banks but they gave you a better interface to that what we all experienced as you know like a horrible experience right and then these fintexs are now sort of being disrupted by this as you're calling it the neo wave of fintex right because these fintex like u you know have been around for like 10 to 15 years so if they've built things the certain way with licensing on banks as they were required. Now they need to make this infrastructure jump into the next generation. Some are going to make it, some are not. So some fintex will

evolve probably into these neointex uh you know I think revolute has done that jump of creating their own like um exchange eventually right and a lot of that those things Robin Hood you can say with tokenizing stocks they definitely made that jump and and some probably won't right. Yeah. Uh so that would be an interesting play to see what's happening there. Yeah, that is. Yeah, I mean I think it's going to be hard for some of them to adjust to the real- time nature of blockchain and just the technical shift. But the advantage is, you know, I think of like Etherfy or Plasma with their plasma one thing that they've recently announced or even, you know, some of the work that, you know, you you could see on Stripe's Tempo Chain or um you know, these types of things. And sure, some

you know, like Robin Hood or or Revolute might take a while to figure it all out, but they have like, you know, a quarter billion customers ready to use anything at any given moment on that day. Whereas like you know plasma etherfy are looking over their shoulder at like all right we got like you know 250 rich crypto dgens who are going to spend some money because they're rich. We do we even need to target that that that base is is another interesting question actually. >> Yeah I well I have to ask you a question. Do you do you think just based on what you said, do you think that when Tempo comes out like 6 months after, do you think that um companies who are not Stripe who are not Stripe's customers,

what do you think would be the usage? Would other payment companies adopt this as well? That's what I'm like wondering recently. >> Yeah. or what? Or like even if it's better, would they be would their incentives be aligned in such a way that it would make it it would harm them to even use it, right? And so I think like if they're seeing Stripe as a direct competitor to what they're doing, which they likely would, then I would think that they wouldn't use it. And then I think that they would just probably fork it or adopt it or or adopt a similar version of it. It's very, it's hard to see Stripe coming out with a bang on this and then other payment companies just being like, we're going to concede because Stripe is the best blockchain ever. Tempo is the best L1. We're just

going to use Tempo. I just feel like, you know, insert payment company here would just be like, okay, well, if they can do it, we we, you know, we'll just do it. So, >> yeah. No, it's interesting because in a way, you know, Stripe isn't Tempo, Tempo isn't Stripe, but because it felt like, you know, there were obviously a big part of the announcement. It felt like it is, but maybe Tempo will end up being like this really amazing chain. And yes, Stripe is an investor, but maybe that's it. Maybe, by the way, we'll we'll see more payment companies investing in Tempo, and that could be interesting.

Like, maybe they'll create that. Anyway, I'm sorry. I >> John Pok's tweet from Mike Dudas about this. Um he said um it's kind of like a a funny like meme tweet but it was like um something about Bridge and Privy basically like um the like oh here Stripe buying bridge and privy for $1.5 billion total is going to look like Instagram level theft in in a few years. Do you buy it or cop or drop that? What do you think?

>> I I agree 100%. Honestly, it's actually a really good comparison. >> I mean, it's uh it's definely a Yeah, it's definitely a a steal. Just a matter of like what kind of return could you expect on that? But I suppose if Tempo becomes the a one of the largest payments chains, then acquiring bridge and privy for $ 1.5 billion is going to look like pennies compared to their total tempo token allocation or equity. I mean it's not just just that like if bridge is making and I don't remember um if if they ever actually showed the numbers other than in certain events but you know let's assume it's making u $500

million half a billion dollars um in Ebida or then if if that move eventually allows them to be the first mover in stable coins AI agentic which by the way go to every AI marketplace there's a stripe module there already, >> right? >> If they are able to capture that, >> just think about what happened in AI recently in the last 18 months and what's the revenue of of um of of uh open AAI, what's the revenue of Entropic, right? If you think about that, if you think about if those on top of those platforms, there will be even more value created which Stripe can

actually connect to as a processor and put their chain on and use stable coins to do that because this whole AI world will run on stable coins 100%. >> Right. For sure. Then >> then yeah, maybe that could be like a another stripe was actually created just by acquiring those two companies, right? like a a cryptobased stripe was. >> Anyway, >> I I want to drill down a bit on this X42 thing. Um, we didn't have it necessarily prepped, but just going to read you something. Basically, back in the 1990s, web architects reserve HTTP status code 402 for quote unquote payment required, but traditional payment systems like credit cards prove too slow and expensive for

microtransactions and you know, insert blockchains. So, Coinbase comes out in May saying X42 resurrects this dormant standard. It lets websites flag premium content during HTTP requests. when you or your AI agent hits paywalled content. The browser automatically handles a crypto micro payment and unlocks access instantly. No redirects, no login. It's built for chains on Ethereum. X42 enables transactions as small as a fraction of a cent transferring the web into a kind of pay peruse marketplace. And you know, then there's also kind of the ERC 80004, a universal standard for agents to discover services, identities, and transact. Each agent gets a blockchain passport quoteunquote for you

know reputational checks and this is should extend agent to agent P2P uh kind of usage. So that is what is happening and it sounds like this is probably the preface for cloud fair's announcement like stable coins later that was came about three months later. Um, I just wanted to to just flag that, make sure that the audience gets to understand that. Um, and just give you a chance to kind of sound off on this. I I know we touched on it a bit, but maybe just give us the the fireworks or or the fireblocks uh roadmap for aentic payments and how you guys are thinking about integrating them.

>> So, first of all, I have to respond about that, you know, um, HTP 42 status codes thing that, uh, that that Coinbase released. I think I think it was a really sad moment for me because I was you know these moments when you think how the hell didn't I think of that >> like that was like that's really good >> it was obvious >> I was like looking at this and I like I should have thought about that those status codes a lot of people anyway that's good so you know I don't know if it will catch or not but I was thinking to myself that is something I can work with um and by the Uh Brian Armstrong is uh is in Israel next week. I'll if I'll

see him, I'll definitely tag him on this one. >> But uh I thought to myself that could work. The Cloudflare thing then I was like, "Okay, that's interesting." And then I started seeing more people sort of riffing on that. So I think that will like >> that's good. That's a good start. The reality is um you know that that's still not a solved problem. Just the fact that they, you know, thought about this, which is really cool, still means that everyone needs to implement something to support it. Yes, they're using the HCTP protocol, which means like the network clarity in a way is solved. How do you move the message? You still need to interpret the message, which has been the hard part in you know distributing uh um you know uh these type of protocols. The way we're thinking the

other side of your question was how we're thinking about agentic payments. Well, the the way we're sort of we we believe eventually, you know, everyone would need to run a wallet. Our goal in a way is is, you know, you can think about us like I would say like the Intel of of security, which is like we have this microchip. How do we make it smaller, smaller, smaller, smaller and make it better, better, better? So, how can we as Fireblocks make the wallets that we create and we've created more than a billion of them, you know, for our customers. How do we make them smaller? So they could be agentic AI size let's call them but also more efficient so they could run even at a faster pace than what they're running

right now but still maintain the security and in a way the enterprise uh protection you expect when you use Fireblocks because if you know if Microsoft you know this is a thought experiment will allow every Xbox developer or Xbox box to have like this micro wallet in the box right So they still expect some sort of enterprise governance and control over that but they still want everyone to have a wallet and those like you know that's like millions tens hundreds of millions of people. So, so that's how we're thinking about how we're making our technology faster, better for that. And we have released like our own MCP server, obviously SDKs to handle like AI. Uh, but I feel like as an industry

of crypto and AI, two buzzwords you should not be saying in the same sentence, but we're we're still really at the early days of exploring how the interfaces even look like. So, I think a year from now, I'll be smarter. today. I can tell you I think everyone's guessing and we're trying to guess as well. >> Yeah. I I gotta ask you, Goldie, before we go, um, you know, rumors on the street is that Fireblocks has a big IPO coming after posting just absolutely crushing revenue numbers quarter over quarter. It's going to be a huge year for the for the guys over there. Got a big smile on your face.

>> Yeah. No, I wish. I mean, I'm first of all, I I vote for whoever is is, you know, uh has those rumors and and please talk to me. I have equity. Uh but uh but but but I think, you know, I think this this wave that's happening right now, I think we're still trying to to build Fireblocks. You know, as I told you, this business could be like 100x bigger. Uh we're still building. Am I Are we ready today to go public? Probably not. Uh, will we ever be public? I don't know. That's up the the market probably eventually.

>> I think so. I think so, buddy. >> But but yeah, it's a good time to go public. I I I wish every company going public right now. Best of luck. >> Yeah. Goldie, man, I appreciate your time. Thanks for the segment. It's been uh it's been a good one. And uh I I guess just last thing just to close you out. Um, I didn't I didn't really get to mention much about any of the work that you're doing with Polygon, but they obviously hooked us up on the uh appearance today. Maybe you can just give us a little tease as to what you guys are doing together. >> Yeah, so first of all, you know, Polygon is one of those chains that have we've been partnering with very early on. The the reason behind this is that they're not just thinking about, oh, okay, how

can we support the latest fad, whatever that is. They're actually they've been building payment related protocols for the last I think four years. And what we've been working with Polygon on is mainly how are we bringing more payments businesses online? What does it take behind the scenes at the protocol level to do it at the liquidity level? And they just been a great partners for you know actually bringing uh you know bringing more businesses on chain probably than a lot of other blockchains. So super helpful in that. Thank you by the way Polygon if you're listening. Yeah, thanks Goldie, man. I appreciate your time. Uh, very insightful and and I enjoyed I feel like uh I learned quite a bit of just about generally how this payments, you know, kind of landscape is actually panning

out. It's it's unsexy. It's really not that appealing to people, but it's really the most important thing that's happening in crypto right now. So, it's funny how these paradoxes of uh industry maturation happen and everyone's focused on whatever like the latest freaking plasma memecoin is or whatever the yield farm is or the NFT hyperlquid drop and it's like no no we're actually doing crossber remittances at scale right now like it's happening. So >> yeah, we're actually ch we are actually changing the world you know and you didn't think you'll hear that in crypto.

So yeah, thank you for having me Andy. This was great. All >> righty man. Thank you. Man, Rand is just an incredible incredibly well spoken guy. Um, you know, he does a great job and yeah, we're gonna roll out of the uh Money Moves Fast episode today. Really enjoyed that. Uh, we're going to flip back. Shout out to Polygon. Shout out to the team there leading payments. Guys, we have a special guest coming up next. We've got John Camarado of ZK Verify. They have their mainet today. So stoked on that. Um and yeah, should be a good good segment. I'm going to check out what that what the ZK Verify

um chart is looking like. I think that mainet's already been out, but yeah, ready to get John up whenever he's ready. Um and yeah, let's have a look. Here we go. Boom. Big launch for the boys. Trading at 173 full FDV. Looks like we got the ZK the VFY. Okay. [Music] Interesting interesting token uh ticker choice. Um and other exchanges. So yeah, we're gonna get John up.

Uh hello. Hello. Welcome on. >> Thank you Andy. um breaking up a little bit for me, but but thank you for having me on. How are you? >> Oh, it is breaking up. I think it should be all right. Um can you hear me? >> Uh I can hear you kind of like really slow. >> Johnny boy. Okay. Is it still pretty bad?

>> I think an hour good. >> Is it still bad or good? >> There we go. John, haven't you made that day? Congratulations, man. How you feeling? Give us some reflections. Talk to us about how you're feeling on this fine Tuesday morning. >> I feel really, really, really good. Um, I'd kind of, you know, been getting a little bit nervous going into today, uh, yesterday and the day before, but I think we did a pretty good job. We're we're live. Um, we're obviously, you know, trading. you were just talking about the the the numbers. Uh we've done a I won't talk about the price, but we've done a crazy amount of volume, which is super exciting. We've got DEX listings. We've got some pretty big centralized exchange listings. We have probably most importantly, um well, two

more important things. We have a ton of really great validator partners securing our network for us that we're really excited. We'll have announcements coming out over the next few days. And then the most important thing is we do have a bunch of proofs hitting mainet. Um, we've been tracking everything on testn net. So far we had hit over 5 million proofs verified. Earlier uh in the week and you know now we're seeing them on mainet and it's super super cool and super super exciting to to be here uh talking about it now. >> Well, I'm loving the hat man.

Congratulations man. Um how's the team feeling? What's the last 48 hours been like for you guys? >> Honestly um team team is feeling great. Uh it's been really really cool to see it. There's a bunch of videos on social media that I would encourage people to check out um under ZKV protocol. Um so the team is super amped. Uh I think they did a a toast uh when we launched at at 8 a.m. Eastern time. Uh I think it was like 2 p.m. Milan time, which is where they're all at. Um, so yeah, they're they're all amped and the last 48 hours honestly were not really bad at all, which I think goes to tell you like kind of how prepared I think we all

were. Um, everything was more or less handled. So I I was, you know, I I went I I was talking to some guys yesterday internally and I was like, "Guys, this is like kind of like nothing left to do. We just have to wait for people to flip a switch tomorrow. It's kind of amazing." So, it's been a really really cool experience for the last 48 hours or so. >> Awesome. Awesome, man. Yeah. Yeah, I mean I mean we were talking just uh maybe a couple weeks back uh into into late summer there or midsummer and you were doing a lot of these kind of like vibe coding experiments with ZK which I think it really got people excited about ZK because oftentimes we kind of get overwhelmed by the technical uh

terminology and explanations of all of what ZK is and can do. But then when somebody shows you the practical nature of the technology in front of your face in a demo, it's pretty it's pretty uh it's pretty easy to kind of be like, "Oh, this is, you know, this is clearly the uh like like the application." Um, you know, talk to us maybe a bit about some of these practical uh, you know, apps that you were able to build out to an MVP form and kind of, you know, what ZK Verify really enables for these types of new applications.

Yeah, absolutely. Um, so we we were running and currently still are running this kind of incentivized test net program called Proof Points. And if you enter, you know, our website for that, you'll see kind of a lot of different types of games that are all ZK enabled. Um, so the the real idea that we were trying to kind of get to here was we wanted people to actively participate in these games and and see how zero knowledge was employed. Um, so we have, you know, some puzzle games like Wordle that are um, you know, uh, what you kind of are proving in the Wordle game is that the server hasn't cheated and

hasn't like changed the word out from under your feet. So every time you guess and you get the um responses like whether it's you know you have the right letter in the right place or you have the right letter but in the wrong place um the it guarantees that the server is still you know evaluating your guesses against the same uh word without actually revealing the word to you. So you get cryp cryptographic proof that that's actually the case. Um I took a little bit of a different approach because I am really big into gambling and I built some gambling games. So I built a slot machine which was super easy uh super straightforward. It wasn't easy but I did it all you know with with the help of AI um where I'm actually guaranteeing to the user that uh the

reels that actually land on you know whatever the symbols are are um generated using you know a fair fairly generated random number. Um, and sort of the impetus for this is every time I travel, you know, to the north northeast, like to Boston or wherever, I typically play slots on my phone, which is a really stupid move. Um, I don't recommend that to anybody, but I typically um, you know, it feels like I always win at first and then slowly like I lose all my winnings and then half my bankroll. Uh, and I'm like, is the is the server cheating me here? they know they know they got me, you know, they know that they know that I won. They know that I have money to play with.

They know that I'm going to keep playing. Um, so I really wanted to kind of create a provably fair version of slots and and that was sort of the the the backbone for that uh particular game. But then I also took it one step further and and built out a blackjack game where the the proof uh the zero knowledge proof is guaranteeing that you're playing with a full and like fully randomized stack of cards. Um, which I think is actually really cool and again a good way of demonstrating that I can prove to you that you're dealing with a full deck of cards that's fully randomized without actually showing you the full deck in the order, which would, you know, obviously ruin the game for the for the house at least.

>> Yeah, exactly. So, so your ZK uh endeavors are centered around uh, you know, one could call it anger at the house for always winning. Uh, knowing that there's some rigging going on. Um but you know if we extrapolate this you know small kind of like microcos and even though gambling and betting is a massive market um there is a real practical need for things like chess for online chess for online poker tournaments for online um even things like as prediction markets have been growing and and growing like you you want to have some concrete knowledge that when you're able that when you're putting your money on the line for something that that that the system is

is operating in a fair capacity. And I and I think that that like level of transparency around the g the verifiability of the fairness, anybody can verify the fairness of a given outcome given a set of rules is is essentially like the big breakthrough here. And then you know maybe we could talk about how this operates more at scale with like stable coins or like with you know L2s or with um you know uh payments and tokenization and kind of like larger uh trends or even like AI right like the the entire AI space maybe you you know I'm curious you had this kind of microcosm of experimentation that has proven well and you know you've mentioned that there's now proofs being

uh you know there's a demand for proof verification on the network already. You know, where is that demand coming from and maybe you could talk to us a bit about how you see the kind of application of this technology in some of those sectors that I mentioned? >> Yeah, absolutely. Um, I mean, you we have we've obviously already talked about the the sort of gambling use case and like the the sort of key is like when there's money on the line, whether it's in a peer-to-peer or like from a prediction market standpoint, peerto-peer or in a like situation where you're dealing with the house, you obviously want to know that you're getting treated fairly and that you're playing on a level playing field. And when you're playing against the house, at least you're guaranteed to not be on a level playing field. So, you want to

at least know like what you're up against, right? And um my argument in this case has been like if I can, you know, guarantee fairness and I can keep people playing these games longer, they have more money that they're churning over and over and over again, right? They're going to continuously lose money over time, but maybe you can, you know, charge a little bit less because maybe you have to prove to fewer people that like auditors or for for example that you're actually, you know, administering a fair game. And if you can keep people around longer, you can sort of keep people buying more drinks or concert tickets or whatever the case is talking from the perspective of casino. Um, and I think you can increase revenue that way. So that's sort of the gambling use

case. I think in AI, you know, there's a tremendous sort of situation that I think we're going to be confronted with soon where you and I know each other, Andy, and if you gave me a work product and you said, "Here, John, like take my word for it." like I did this work product the right way. Um given our relationship, I can sort of trust you, right? But the whole idea of zero knowledge proofs is removing the need for trust between two entities. So if you're dealing with a fully agentic world where you have, you know, two different bots that are trying to communicate with one another and do business with one another, whatever the case is, how does one bot guarantee that the other bot is doing their job in the way that they were supposed to be doing

it? And this is where zero knowledge proof sort of come in. I can guarantee that like this bot is actually operating the way that it was programmed to to do because I'm able to verify a proof in a very you know fast and efficient manner and this is where ZK verify hopefully starts to see like sort of significant volumes. If you wanted to go into payments, like there's a whole use case around private payments, um, where I may want to pay a bunch of employees that work for me, but and I want to do this on chain with stable coins or whatever the case is, but I don't want to reveal the addresses of my employees. Well, this is now the use case where this use case here is, you know, private payments

where I can um, use zero knowledge proofs to essentially shield the amount of tokens that were paid to a particular person. and what their address is so I'm not exposing them on chain. Obviously, we need to do this in a compliant way. So, there needs to be some sort of compliance aspect to it. But the zero knowledge pro proof component is really hiding, you know, those details so that people aren't exposed on chain. Um there's certainly you know a tremendous number of of use cases in in trading all sort of related to you know keeping maybe your strategy private where I can guarantee or I can prove that you know I have ownership over some funds and uh the funds belong to me when I want to withdraw them but not reveal the

strategy that I employed to actually get from you know certain amount of funds in certain wallet with a certain token distribution to another um which obviously allows you know proprietary trading to happen on chain um where maybe you know you or I can't copy what's being traded. So these are all sort of use cases for ZKPs zero knowledge proofs. This is why we think we're so bullish on ZK verify because all these proofs need to be verified somewhere. We think they all need to be verified onchain if there's going to be onchain activity happening and um we're set up in the right way we believe to be able to handle that volume which we hope explodes over the next you know number of years.

>> Yeah. And so that the the proofs are being verified on the ZKverify chain itself >> uh which if I if I understand correctly is effectively a substratebased chain that is in L1 with its own set of validators. Effectively the the reasoning here is the same reasoning why like the entire modular kind of concept exists is you just have a specific chain specific use case. It's going to help Ethereum scale. Ethereum obviously can't verify all these ZK proofs that scale for a cheap fast cost right now. There's too many transactions going on. It does not have the load capacity if you will to to kind of manage that. And so basically this the ZK verify L1 is able to then kind of be a hub for just proof

verification. And so that specialization allows you to have this kind of edge in the market. Uh maybe you can talk to us a bit about the chain design. You know, I think I I mostly got it there, but talk to us about why you chose to to build this chain in the way that you did and why you are a strong kind of believer in this specialized or purpose-built kind of chain thesis. >> Yeah. Um, so I think what you said earlier, you kind of hit the nail on the head. It is a substrate based chain. So it's based on the same kind of technology that Polka Dot is is based on. And that's important for the particular reason that um we're in an industry, we're talking about an

industry with zero knowledge proofs that is pretty old actually. I think first zero knowledge proofs were um thought up of in the 70s or 80s if I recall correctly. Um but the the technology is still at a point where it's like really really rapidly advancing very very quickly. Um, and so, uh, what we've done, what we've seen, we we've built verifiers, um, into this specialized blockchain that are, uh, can essentially verify, I think, eight or so different types of proofs as of right now. We can expand to any number of verifiers that that we'd like to. Um but particularly with a couple of the ZKVMs like Risk

Zero for example, which just did their um token launch recently, uh they've been iterating through versions of their verifier and their prover code um really rapidly, like almost bi-weekly, I think. And the point that I'm trying to make is we chose Substrate because it's very very easy to deploy changes onchain without having to do hard forks which require like a tremendous amount of coordination across a pretty vast array of stakeholders including exchanges uh validators anybody running a node essentially has to update their software. And you know I can't even count the number of of stakeholders that we've had to deal with to set this chain up to start. um but deploying changes are going to be very very

straightforward for us. So given the sort of uh all the innovation that's happening happening in the industry right now, we think from that perspective, we're very well positioned because we can rapidly upgrade the chain and accommodate you know technology needs of our users very very rapidly. Um to talk to your specialization point, we've stripped out almost everything from the blockchain um except for verification staking and governance um functions. So instead of having to deal with you know like an EVM has to deal with just generic um computation, anybody can deploy any kind of smart contract on chain that they want and they have to compete with everybody else for resources. We've

said, "Okay, we we just want to do these this one thing, and it's really three things, but the one big thing is proof verification. We want to do this one thing really, really, really well and dedicate all of our block space and all of our computational power toward it." Um that uh coupled with the fact that we've you know natively built these verifiers uh in the node infrastructure makes it hyperefficient for us to verify these proofs and do it at scale and very very cheaply. If you contrast this to Ethereum, you have to have certain pre-ompiles for some of these newer proof types. some EVMs that are deployed on you know uh that haven't you know added certain pre-ompiles can't even

verify some of the proof types that we verify on ZK verify um or that we can verify on ZK verify. So I think we're really really well positioned and are putting ourselves in the right position to be able to you know capture a lot of this demand as proof volume increases to accommodate all these use cases that we talked about before. Yeah, I mean I think that that is where the space is trending, right? If you look at the payments chains that have been announced recently, the plasmas, the tempos, etc. It's stable cocoin, it's payments chains. If you look at ZK, you've got verification chains. You're going to have pretty much like RWA or just like permission bank chains. You're going to have gambling kind of casino kind of memecoin fund chains. you're going to

have streaming chains that are just meant for streaming and they'll they'll be trying to dominate that. And so I just think like as the market goes into this more uh sophisticated fashion, you know, those focused on this general purpose outlook are going to struggle, those with a specialized approach will win in the more immediate future and then they will also be h they will be forced to evolve over time as well. Um, you know, it's funny. It's kind of a loop here where you kind of start with a specific use case, then you realize that you got to kind of build out the whole thing, then you go back. So, I I I could see multiple loops here happening for ZK verify as you guys evolve and try to think about, okay, we're doing, you know, this much proof verification, if we get, you know, if if the ZK market is

going to 10x over the next five years in terms of just total proofs being used on chain, well, here's how much, you know, revenue we could, you know, theoretically forecast, you know. Okay. Well, h how do we kind of grow that even bigger? Is there is there another service that we need to offer on our chain or is there another vertical? And so you kind of maybe could start where you are now and then end up evolving into multiple product verticals which I'm sure you guys are thinking about. Um but I think generally like the trend is clear for ZK and it's a really just about capturing part of that pie and then figuring out a way to redistribute that in a sustainable fashion towards getting the token to do well. I mean, more or less. I mean, like that's what the market wants right now. I think the

market's obsessed with buybacks right now. I think it's also obsessed with like over fixating on it. So, I don't think that you need to like go too crazy on that front, but like the general trend is clear. And so, yeah, maybe you can just talk to us about how you're thinking about like the TAM here, like you know, how big could ZK get? Like what is what is the wedge for you guys? >> Yeah. So, I mean the TAM just to maybe put numbers to like kind of what I'm going to talk about here. I think at peak right before launch, we were verifying about 350,000 proofs a day on our test net. And you know, people come out and say, "Oh, well, it's test net." And oh, well, they're just fluffing the numbers. But

the reality is, um, we weren't fluffing the numbers at all. Uh, I mean, you could argue that the incentivized test net is fluffing the numbers, but I think most of our volume was actually coming through partner proofs and that's 300,000 a day. I don't know what that, you know, amounts to over the course of a year. Almost 100 million maybe, right? If I do math the the right way. Um, and that's like pretty substantial volume given where the like industry is at in terms of adopting ZK tech overall, which it really really hasn't done to a large extent. Um, so I think like you know we're on a really good path right now. I I think internally I've seen some numbers uh externally that have

envisioned like 90 billion proofs by 2030. Um, I think if we, you know, were able to capture even a small percentage of that market, we'd be doing really well. Um, but I think that that number is probably underestimating the potential for ZKPs. Uh, we're kind of internally talking about a number around a trillion proofs. Um, or even more. Um, if you think about all the places where you can insert zero knowledge into like your normal workflow where when I log onto a website, I might want to hide my IP address or I might want to prove to somebody that I own a particular number of tokens on one chain and, you know, have that um have that translated to another chain. There's so many different places where you can insert these

things. And so we're big believers that the volume is just going to erupt. Um if you look at sort of like some of the other trends like I kind of look at zero knowledge development as like being broken into maybe three or four components. The first sort of like ZKVMs I think there's 20 or 30 out there. Two couple big ones but 20 or 30 out there. These are the ways that you write these zero knowledge programs that actually generate a proof. Then you go down one step and there's these networks that we've seen a couple recently launched tokens i.e. boundless and succinct. um that are starting to pop up. There's maybe five or ten of these. Um I think at last last time that I looked um

there's these proof aggregation networks. There's maybe five, six of those that are big names. And then there's three or four verification dedicated networks. And so if you like look across this, you're seeing a tremendous amount of investment being put into this technology. um you're seeing these big token launches which you know uh have done tremendous amount of uh they've added a tre tremendous amount of value to the ZK space overall. I think it's really hard to kind of ignore and and think that this thing isn't going to be really really big and you know we're at the bottom of that stack so we're sort of thinking about how do we you know move up the stack and try to capture more volume. Um, but it's really hard to to to think that there's not going to be billions and billions if

not trillions of proofs long term. Um, and you know, all those proofs are going to need verification. If they verify on ZK verify, which hopefully they do, that means, you know, fees paid in BFY token, which we just launched today. Um, so that's, you know, I think pretty bullish for for the revenue generated by our network. Um, and you know, if you pair that with you're talking about buybacks, we're considering uh, you know, putting a governance proposal out to maybe burn some of the fees that we're generating from these tokens. We could be in a situation where we're actually on our network at least seeing deflationary tokconomics instead of inflationary tokconomics. So, yep, >> a lot of cool stuff to come.

>> Very cool. Very cool. Um, and I appreciate the kind of wide breath explanation here of just how big this could really get because I think people think um, I had Ava from Electric Capital on earlier this week. you know, the guy kind of invented Google Maps, like was like one of the co-creators before like, you know, it was like a whole thing, you know, you can check his pin tweet and he says that humans aren't wired to think in exponentials. Like we we don't understand exponential curves from a pure brain perspective. And so like when you think about ZK, it's going to be so obvious in hindsight when right now it's like we are literally doing the test pump for the test pump. like like

we haven't even got to the test pump yet of what's about to happen and it's hard to even wrap your head around that because it literally means from where we are today like 100x growth in like a short period of time in terms of usage of this technology. Now, obviously, we could still be in the kind of like early phases of the S-curve, which is like the beginning of the exponential growth for a while, which is why it with people's heads because like you could stay in like the flat zone for years and then all of a sudden in a year or two, you go like 10x, 20x. It's really just and I, you know, I think that that's pretty much >> it's not guaranteed to happen to ZK, but it's pretty damn close. uh in terms of you know possibility

>> I I think what we're hoping for is that this mainet launch and token launch um is sort of the catalyst for maybe the test pump that you're talking about right like we're showing people that hey maybe you were thinking about ver like generating a proof maybe it was really hard for you to do that so we're thinking around like how do we maybe consolidate some of that development stack into a much simpler product on on one side of things, but then on the other side of things, oh, it's expensive to verify these proofs. Like, I have to charge my customers maybe a dollar or two extra for doing what they're they're needing to do just to cover the gas cost of verifying these proofs. Well, now that cost has collapsed from maybe a

dollar on Ethereum, maybe 10 cents on some of these L2s to a fraction of a penny with CK Verify. And so we're hoping to be sort of the the spark or the catalyst for that, you know, trajectory to start to really really take off on the on the exponential curve that you're mentioning. >> Yeah. John, last thing for you. You know, I mean, this is just the the beginning. It's kind of like a new day one type of vibe where, you know, it's been a it's been a long road to get here from Horizon side and then now rolling out ZK verify. I know that was a whole process that has, you know, been an undertaking and I think between you and I, you would have wanted to see this go live more like Q2, not, you know, the day before Q4, but that's timelines for you in crypto and in

development. What's next? What are you guys focused on to end Q4 strong? What is the main priority? And then what is the first half of of 2026 look like for Zika Verifa? >> Yeah. Um, I was actually okay with the timeline. Uh I think I I think we actually may maybe nailed the timing perfectly um with this last day of Q3 launch. Um seems like a lot of things are going in the right direction. We had you know a couple of these ZK projects launch earlier in the month. I think that was really good for you know continuing to hopefully grow the space a little bit. Um so not to correct you but I'll correct you anyway. Um but yeah, I mean rest of Q4 is really going to be around growing, you know, ZK verify,

growing the number of proofs that are verified on chain. Um really making sure that everything's kind of good to go, good to continue to scale. Like the big thing for us is scaling right now. We need to figure out how to, you know, accommodate the trillion proofs that I was talking about. um because we we do expect to see maybe not super short term but short term expect to see a lot more proofs being generated and needing verification. So I think a lot of our roadmap is really centered around that. Um we are working on an API based product where want people to send proofs much more easily to the network without having to set up a crypto wallet. Um kind of more like a SAS product where we make it really easy for people to

integrate ZK into their stack. And part of the idea with that product is maybe we want to vertically integrate some of these things that I was talking about, right? Like there's so many different technologies that we're asking people to use when they enter this ZK landscape that they already don't really know very much about. It's like okay then choose from 20 different uh ZKVMs, five different prover networks, five different aggregation networks, and five different It's really confusing, right? on top of ZK already being hard to grasp from like a lay person's point of view.

So, we're thinking about how do we make that whole stack much easier to use. Um, I don't want to spoil too much there because obviously we're on a live stream and everyone's listening and, you know, I don't want to give it away, but um, you know, there's we're putting some thought into putting some thought into that. It's really about growth from here. Um, >> I I keep telling everybody internally like everyone's like, "Oh yeah, tomorrow's the day. We're done." And I'm like, "No, it's just the beginning." like we got to keep accelerating from here. >> It is just the beginning. >> Yeah. So anyway, >> that's thank you for your time.

Congratulations on the big launch today. Thanks for joining the show. Um friend of the show. Welcome on whenever and uh appreciate it, man. >> Thank you. Shout out to the ZK Verify and Horizon Labs team. Great job, everyone. >> Yeah, man. Definitely. All right, John. We'll see you soon, buddy. >> Thanks, Andy. Great to be with you. >> All right, buddy. Peace out. >> All righty, guys. Bang, bang, bang. ZK verified mainet is live. I feel like it's mainet season. I mean, it's literally just Ray season. It's freaking launch season. It's everything season right now, guys. So, strap in. Let's get some music playing. Leo in the back.

Let's get the screen up. We're going to We're going to scroll the timeline a bit. We're going to check out what the latest is. I I I got to talk about Andre Cronier's stuff. I got to talk about um the latest with the um uh yield basis stuff that's just bonkers. So yeah, guys, we're going to we're going to pull that up and just kind of rip through that a bit. Uh let's get some tunes playing in the back. Uh get the vibes going and yeah, we'll get this thing started here again. Give me a quick sec, y'all. Um yeah, I mean guys, bullish on ZK verify. I'm glad that I'm glad that they were able to get that thing out.

We've been helping them for a while. So stoked to see that. Stoked to see that. This Andre Cronier thing was where is where we're going to start. This is just bonkers. Absolutely bonkers. Absolutely bonkers. All righty. So, I'm going to pull it up on my screen and then we're going to rip through it. Um, one sec y'all. It's coming up. Let's get some vibes tunes going.

All righty. All right, let's get this up and we're going to scroll the feed. Talk about the timeline. Um, all righty. Let's go. Let's pull her up. And um, yeah, we're going to scroll the feed a bit and have a look at this Andre Cronier announcement. That was absolutely crazy. Here we go. Let's pull it up, Leah, and jump in there.

Yeah, guys. So, Andre Cronia came out and said, "Look, I'm launching a I'm launching Flying Tulip and I've raised $200 million." Um, and yeah, like I've raised a $200 million seed round at $1 billion token valuation. The trippyst thing is that if you go on the the website or like the Twitter account, it literally says it literally says no token. Like you're telling me there's no token on this, but in this you literally say 200 billion 200 million at a hundred billion or at $1 billion token valuation. So what I would like to understand and to read to you guys is

this from Dan. Dan's a friend of the show and he's done a really cool job of being in the know and being, you know, able to talk on these subjects. And so it is a $200 million seed round at a billion dollar FTV, but it's not exactly. And here's why. Basically, Flying Tulip is building a full stack DeFi exchange, but the way that this works is that if you invest, you actually put your stable coins, ETH, or other assets into the protocol treasury. There's no lockups or anything, but basically, you get a a option to sell and get your principle back whenever. So basically, if the token goes to zero, you can get your money back. if it goes 100x and you try to get your money back

before you you you miss out. So basically what's happening is there is literally uh a a option to like get out at any rate anytime um and get your capital back. And so Dan argues that the mechanics are deflationary from day one, right? Um, and yeah, I mean I think that I think that this is an interesting phenomenon. This is probably the most re insane part of all of this. Take a second just to read that. Um, basically the team gets no token allocation. They only make money if they actually create

a product that makes money. Like this is the revenue meta in a nutshell right now. And Dan is spot on. like this is just going to keep happening and this is scary. Like if I'm a protocol with no real product market fit revenue, I'm terrified right now. Look, I think the game is still going to be going for a while. The VCs will still be able to make money on products that don't have real PMF, but I'm telling you guys, this type of stuff is we it's clear where the trend is going. And like why don't I just give you the next example of where the trend is going. Yield basis comes on and goes uh absolutely smashes the crowd sale.

They have 87 million and then I tweeted it out. They hit 150 million but it actually hit it actually hit $194 million in USDC deposited for the sale. Are you kidding me? Like this is like the trend is so clear guys and we have Michael on tomorrow so strap in. um that's going to be a banger. But like this is so clear as to where this market is going for capital formation like that the the market is going towards more fair experimental ICO crowds style capital formation and there's no way around it. it like it is

going to happen and it's happening and so thankfully for the regulatory environment that that we're in is better um you know it's great but it's happening and so I think it will continue to happen so yeah I mean oh here we go >> it looks like I can layer launch >> and you have no idea why moved $50 million and you have no idea why There is no way to verify its logs or prove it used.

>> There is no way to verify what just happened. >> If somebody used a nuclear weapon, it's verifiable. AI when used may not be verifiable. Wow. >> When it's not verifiable, then you reduce your trust. >> We prove how we work. AI does not. >> Yes. I love that movie. >> Yes. >> If we can't prove it, should we trust it? >> You realize you're renting your brain. It all remains a mystery. Every trade, every choice, every thought simply unprovable until now.

Introducing IGEN AI and IGEN compute. See the prompt, verify the logs, check the model, prove the result. >> Build really powerful, expressive crypto apps, verifiable agents. Make smart contracts smarter with AI. >> Stop wondering if your AI is right. Because there's never been a better time to trust AI. With IGEN AI and IGEN compute, AI finally gets its verification layer. >> Wow. I'm saying that crypto launch videos are hitting Silicon Valley level.

I layer knows this obviously. I'm sure they've looked at Roy Lee's clue stuff. Maybe not the best example, but that was a banger. Wow. Fire. I AI and I compute. Wow. So essentially all of this is is around um model like this opensource AI and if you guys have been watching our AI super cycle show with Near we just posted a banger with E-mod uh Must from uh intelligent internet accessibility AI founder. It's on our YouTube. It's flying in views. People are loving it.

Crazy video. And I cloud's attacking the same thing. is basically just like verifiable LLM inferences, deterministic APIs, untampered prompts, I call them organic, some could call them virgins, IEN compute uh with TEEES. So it looks like IEN is just fully fully AI pivoted now like they like their full direction is like our technology is best match for AI and like here's our future. So they've got IEN cloud, IE AI and IEN compute. Um, wow. That's pretty sick. That was an amazing video. I'm gonna play back the top part there or just the end here. This thing was insane.

>> Smart contracts smarter with AI. >> Stop wondering if your AI is right because there's never been a better time to trust AI. With IEG AI and IEN compute, AI finally gets its verification layer. >> Wow. One more time. Prove the result. Build really powerful, expressive crypto apps with verifiable agents. Make smart contracts smarter with AI. >> Stop wondering if your AI is right because there's never been a better time to trust AI with IGEN AI and IGEN compute. AI finally gets its verification layer.

>> Banger. I think we got last minute that Shirro's going to join the stream. So, we're going to get him on and uh yeah, that'll be crazy. Oh my gosh, people are trolling Andre so hard. Um wow, that's hilarious. That is hilarious. Wow. Okay, look. Check check check this out. Bridge launched open in issuance a new product aimed to help every country company launch it own stable coin.

Oh my god, guys. I want to talk about Laura Shin. um she has announced a pump fund coin and she got absolutely slaughtered for it and I don't think it's fair. So, I would like to just kind of think about how people are treating Laura and maybe give us a bit of a I think people in media in crypto get such a bad rap and people hate on us so much and Thread Guy had a whole thing about

this with Luca Nets when Luca was on. People in crypto just destroy people in crypto media. Like it is not easy being in media in crypto and managing it all. Like it is a total hard job and people absolutely just go to the end of their wits end to make them freaking like just angry. Like they make they just get angry at people in media.

And so look, Laura is just testing some stuff out guys. Like I'm not saying that I'm going to launch a pump fund coin. I'd be lying if I said I didn't think about it though. Like I would be lying if I said I have not thought about launching a pump funcoin. New type. What is Dan's Twitter? Peep the video from Kyera. What's up New Type? Good to see you bro. Um yeah, what is Dan's Twitter? Who's Dan? Um I don't know who Dan is, bro. Um and peep the video from Kyra. All right, I can find that video. Um, I'm not sure what what that is, though.

Trade to win. Okay, here we go. Yeah, I have my ice in a couple of yachts. Sell some Bitcoin. Ah, maybe just a little. >> Yo, this is so fire already. >> Let's see what's going on. [Music] Round one, fight.

Bro, this was crazy. [Music] >> Come on. Come on. >> Oh my god. [Music] [Music] Oh my god.

>> You good? >> Who is this guy? >> Oh my god. This is crazy, man. >> Wake up. Next trade's live. Round two. Fight. [Music] >> All in. Oh my god. This is ridiculous.

>> All in. [Music] >> Full house. Straight flush. Congratulations. [Music] >> No. [Music] >> He's going to run away with all the girls at the end of the night to the freaking villa. >> Oh my god. That might have been one of the best videos I've ever seen for a product announcement. That might have been one of the best videos I've ever seen for a product announcement. Wow.

Oh my goodness. Wow. I've never seen Wow. Holy. That was insane. Um, wow. Um, that was one of the craziest video product announcements I've ever seen in my life. And that was insane. Uh, wow. That app looks insane. It looked like he was trading in a virtual screen and then he was in like some AR VR headset on his computer, but he was playing poker from his That was insane.

Um, wow. All right. Well, we're going to get Shirro up on a special last minute appearance. Gotta love it. um he is calling in from token and uh I'm stoked that he is able to join us because I just did a segment reacting to the launch video and I also thought the launch video was awesome. So, we're definitely going to clip that. And I watched the whole video and it was sweet. Um and so we're just going to get him up um in a sec here and talk about their news today. I'm just getting things set up. We'll have the team get it back in the back. And Shirram is up, man. So, Shirram, welcome to the show and uh thanks for calling in from from

Singapore. I know it's late over there. >> Hey Andy, really excited to be here. Yeah. >> Yeah. Good man. Good. Good. Well, I just watched the video. So, I am I am in the know. Um so I en cloud is now has announced IEN compute IEN AI two key parts to this broader verifiable credible commitmentbased AI system that you know you've seemed to pioneered somewhere in the brain up there over the last five years this it's just slowly starting to come out maybe you could unpack this a bit what is the you know what was the kind of lore behind these this kind of dual wielded product announce announcement and you know why

is this important? How did this kind of come about? Let's just start with the story first. >> Yeah, absolutely. I think maybe I start with our rebrand to IGN cloud just to remind the audience. So um we did this re you know announcement of hey you know we have this uh umbrella brand called Ien cloud. The way you think about it is like when you think of a uh crypto, right? You have an you think of crypto's evolution as you have Bitcoin, right? Verifiable money, right? But runs on a $5 hardware like you can run it on a Raspberry Pi. Ethereum, you know, this verifiable computer on which all the DeFi verifiable finance took off. Uh we think of it as it can run on a $500 computer. And then you look at Salana,

the L2s, you know, Mag things like that. They basically bumped it up to a $10,000 server, right? Like uh, you know, uh, a Raspberry Pi to like a laptop to a server. But I think I when I got into crypto, we were, you know, I used to say, oh, you can build an Uber onchain. You can build like a recommendation engine, you know, with a social network on chain. you can build uh a new open AI onchain and you know it just turns out those are not really possible with a $10,000 server you need much much more compute you need cloudscale compute so that's why we re what is enough to scale you know crypto

and verifiable compute to this kind of scale so that that's kind of like the main question for which I cloud is the answer and so when you think about the cloud right what is the base layer of the cloud, right? So we talk about base layer in crypto with alons and altos. The base layer of a cloud is a data center, right? So Amazon or Google, they run a big data center like you know hundreds of megawatts and uh that's the base layer of the uh cloud. The base layer of IGEN cloud you know but that is centralized non-verifiable the base layer of IGEN cloud is IGEN layer IGEN layer allows anybody to stake and operate and like keep these computations ongoing and that's the base layer of cloud it's decentralized by a fireball

as opposed to centralized opaque so it a cloud like Amazon or Google is not just made of data center it's not like just like a bunch of a rack of servers and then that's the cloud it's made of these data centers and services built on top primarily like storage, compute and AI are the three main services that like an AWS will have right S3 for storage, EC2 for compute, bedrock for you know AI. So that's it that's what I cloud is. So we have the base layer on layer. We have DA which is the data layer of IGEN and we have IGEN compute and IGEN AI were the two big

missing pieces that we just announced. But that's how we think of the whole picture. That's why we rebranded to cloud. So each of these new things are don't seem like oh why is a restaking protocol doing AI or a restaking protocol is doing you know some compute and all this stuff. It's a crypto cloud. So that's what we want people to remember us for. And so in a crypto in a cloud AI is the most important service. In a crypto cloud obviously that's the most important service. So that's what's happening. I can AI and compute just launched today.

>> Yeah. I mean I think it's very clear that the market has there there have been players in the restaking landscape quote unquote that have evolved and have you know kind of use the core idea to then now grow into new parts and those have kind of just like stuck in it. And there's a clear discrepancy in where like value is is and attention is going to like Etheri is a great example where they're now doing this whole neo bank thing this entire credit card system you know and that is all built using IGEN layers technology with their uh W E product or ETH product state ETH product but they've now made ways of you know extrapolating that to actually create like like an EOB in IE layers case now I

cloud is like this foundational cloud layer as you've described uh with with these kind of services stacked up on top. It's kind of like a cake uh if you will. There's different layers to the cake. So, I compute and I AI have just added a new layer to the cake. Maybe talk us through each one individually and why they're compatible with each other um and kind of what what this brings to your product suite. >> Yeah, absolutely. So, uh I can compute when I want to write a program. I just want to write an arbitrary software like I want to write a Java program or like some complex database all these things and then I package it into a docker container and normally you ship it to a

regular cloud to go and run it instead you write a program instead of writing this program normally when you're doing onchain development you have to write it in a smart contract and you're constrained by the constraints of the smart contract instead you can write in any language of of your choice and then say oh I need like an instance which is this big right so normally when you're running on a chain you're constrained by the nodes that's available on that chain right so there's a homogeneous set of nodes say no I need this excel instance that needs to run this it needs 128 GB of memory whatever your constraints are you specify that you just send it to cloud compute what it does is it takes your like container executes it offchain

and brings the results back onchain So it's a container with smart contract grade security. Okay. So that's our kind of like end vision. And you ask like okay you know you're running the container somewhere. Is it like a centralized cloud? What is this thing? And so the answer is the way we get so one of the things with a smart contract is it's unstoppable right it's run forever. So as long as there is gas fee being paid but smart contract lives on forever. How do you get a similar property with uh like IGN compute? Anybody can stake and run the igen compute node. Okay, so I download and then I'm staking and then I'm running the igen compute node. And if I'm running the igen compute node, I also

have to write the snapshot of the state you know periodically into INDA. So you just, you know, store it in DDA. IDA is this decentralized network, you know, which we built. Ida is the fastest DA network in crypto. So you can there's a lot of bandwidth you can accommodate I think maybe now like 15 solenas inside I can da right and we're scaling this more so we we will be at thousands of salon someday so you take the um uh you take uh uh this container and then like the node runs it writes a snapshot to da and then it can go and trigger actions on any chain. So even though IGN layer contracts are built on top of Ethereum,

the utility of the IGEN cloud service we want to be prevalent in all of crypto right so whether you're on base or arbitrum or world or eventually you know avalanche or salana or mag you know monad or wherever right so you we want it to be inversal it's going to take stages to get to all of that but so that's the computer executes offchain and then like you know brings the results on chain so the obvious question is whoever is executing it. What if they are malicious or like they don't do their job right? So first they get slashed so they lose their stake and then second anybody else can come and stake and continue the computation. So it's not like oh they are bad and the whole system dies. So anybody else can come and stake and continue the

computation and but to continue the computation they need the data because you know where do I start right say like oh go continue that agent's computation you don't have that agent's computation so download the data from ID and then continue on so this is the core functionality of compute uh and you know when we said like the this is a mechanism to maintain the uh property that the system runs on forever so as long as there is money then somebody will have the incentive to show up and then keep continuing the system on. So then you have the other property that you need which is safety or correctness.

How do you assure that the program is executed safely? And here you have many different options right. So you could have a zk proof right which is the most rigid version but it's also very expensive. So pro creating a zk proof may be like 10,000 to 100,000 times more expensive than just running a computation. So imagine you're pay playing a massively multiplayer game, right? It may cost like hundreds of thousands of dollars a year to just run the computation. To do a proof, it may cost you hundreds of millions of dollars. So nobody's going to look do a ZK proof for those things. But you may want to do a cryptoeconomic proof where you say, oh, you know, there's staking and slashing. And the earlier thing I

said is taking and slashing for livveness, but there could also be staking and slashing for safety. Okay? You know, you didn't run the computation correctly. If it's verifiable, you'll lose your stake. Or there could be like I'm running inside an encrypted hardware encrypted enclave, you know, a TE and then that's how I get verifiability. The vision of IGEN compute is to support all of these. These are just modalities, right? So it's a stack that runs offchain compute and onchain is going to bring the results verifiably. You can choose the grade of verifiability. Obviously, there's a cost to the different grades and you know, you can choose what you want. There's also like some constraints of programmability, right? If you want to write something where you want to use cryptoeconomic slashing, then it should be verifiable,

observable that you ran the computation correctly. You need some level of determinism. You run the program, you run the program again, you should get the same answer. So there are some constraints whereas if you just want to slam it inside a TE, you don't need those constraints. You may get additional privacy and other things. So you get some trade-offs and like compute wants to support all of these things. On day one it's built with uh TE grade safety. So that's the day one launch. I can compute run computation offchain bring the results onchain and you have all these different ways in which to do it. But the abstraction that you can think of think of you're going to write a smart contract but you can write it in any language. It's running offchain but it's kind of like bringing the results

back on chain. I think this is a big upgrade to crypto. Just this one thing. Um many things that were simply impossible before become very easy to do in this method. Um like one example is gaming, right? You know, I had this slide where I say like, you know, show Vitalix like uh he had this World of Warcraft spell that he had and then it got nerfed and he was pretty pissed and uh uh he said, "Okay, I'm going to start Ethereum, right? So, semi- urban legend, but it's useful.

>> And you know, a lot of crypto gaming was kind of like built around this thing. Oh, you know, you should put your assets on chain, so then you really own it, right? And all of these things. But I think, you know, if you just put an asset on chain, but the game itself is running offchain by a centralized opaque party. Even if you own like that spell as an NFT, I can always say I'm yeah, you know, those NFTs are not going to be respected inside the game. So having game execution be run nonverifiably, but game assets being verifiable or onchain is not doesn't really give you anything.

I mean, it's it's okay, you can do secondary trades, you know, that kind of thing, but even inside of most games, you have like, you know, the ability to trade assets. So, okay. So that's the one one example. The other one, you know, the other phase I think of crypto gaming was, hey, I'm going to just build the game onchain, right? Like Dark Forest or something where like the entire game logic's on chain, but that is very constrained like barely, you know, a few games can be rich enough to build that and have a good enough user experience. But what we're seeing is I think the third third phase of crypto gaming where you're going to have assets onchain and you can run extremely complex massive multiplayer like online games because it's just running on the

cloud right so you just throw your logic there and then it executes this massively multiplayer game and it is forced to follow the rules that are already encoded which means if you have that NFT which shows that spell in World of Warcraft it is going to be respected because the code that's running automatically reads it and respects it. So finally Vitalic can stop you know right >> and so I start to see how um this this this idea of a verifiable off-chain environment which reminds me of kind of like the co-processor uh kind of concept right with certain modalities that have for specific use cases. So for highly

financialized use cases of very private information, there's a certain modality that is able to be able to adapt to that. And then something like a more fast-paced gaming environment where maybe there's a requirement for the privacy of your next move is important, but for all other intents and purposes, they just want the speed and the verifiability consistently. There's another kind of modality here that I compute can provide. And then if you just think about that spectrum being both sides of the spectrum than anything in between and so that is happening that is I can compute then we've got this kind of corollary situation here where we're also living in this blackbox closed source AI environment this unure

this uncertainty as to where data is going when you input it as prompts this uncertainty as to where what models you're using and who what they're being trained on and so I can start to kind of at least infer as to where the IEN AI comes in in terms of okay if we know that a lot of the things that are going to happen in the world whether they're onchain uh or not are likely going to have some sort of agentic influence then the same concept that we just applied to the verifiability of offchain of events and verifying them onchain and being able to use that onchain well we need to do the same thing for AI and then we need to do the same thing for bringing that those AI and their capabilities and their models and open sourcing them and

bringing those onchain as well. So I can see a parallel here starting to be drawn but maybe you could illuminate this a bit better kind of draw that parallel for IGEN AI as a nice kind of uh counterforce if you will to compute. >> Yeah. So when when you think about uh you know as AI systems you know move into more central position in our society and our life right you know uh Eric Schmidt the former Google CEO uh was mentioning in a podcast that uh imagine your kid is talking to an AI agent and you know uh on a daily basis what you do is you have already you know

vetted this agent and seems like a reasonable one okay it has reasonable guard rails. I'm going to give it to my kid and then like uh let my kid talk to it. Over time, what can happen and what is sure to happen is that the agent that you're dealing with is not the same as the agent that's going to be two weeks from now, from one month from now to two years from now. So there's a drift, right? continuous drift of the model and like you know it's it's keep keeping on getting changed know GPD went from like a normal mode to a psychopant mode like you know in the middle and then like people got plenty pissed at it and like you know some people are losing their minds.

>> Yeah because it's starting to be deceitful and deceptive and you don't you know these are starting to happen with the sophistication of these models. >> Yeah. Exactly right. Just like social media was optimized for you know virality of content like you know uh Sam Alman is going to optimize chat GPT for retention and like you know >> and like you know who is going to ride along with your delusion is >> so those agents are going to be more >> sorry yeah we had emod on emod from AI I know you know him well and I'll just I don't want to interrupt but he said there's a revenue to evil curve curve that all the open AI, you know, anthropic and tech giants are on. And

it's a it's kind of like, and this is my example, not his, but you know, like the um the hot crazy scale, which is kind of, you know, it is what it is, but like AI is kind of on a revenue evil scale and like there's a there's an entire y and x-axis of a graph happening here. And I think this is kind of what you you're somewhat alluding to. >> Absolutely. Absolutely. So now like the question is if if there was a revenue evil curve, right? So how do we actually go from I mean to quote Chris Dixon, how do we go from don't be evil to can't be evil, right? Like that's the question that you know that's the whole that's I think what crypto is trying to answer for all of us. And so how do you make AI

can't go evil? So there are various levels but one level is I guarantee to you that this model executed and is giving you your inference results. It's this what do what do I mean by this model? I specify the entire weights of the whole whatever LLM to say oh this is 011357 whatever this entire matrix of weights then I pre-commit to it and I say I can prove to you whenever you send me a query it went through this particular LLM and you got a result. So and as you can see to prove to you this claim firstly requires a determinism that if you give me the same query I should be

able to give you the same answer otherwise how do I prove that it's the same thing that's the right thing that is actually doing this and turns out determinism is not an easy problem in AI like you know normal experience suggest every time you ask an AI different question it gives you a different answer but this is due to a random seat that goes with it every time you ask a question, it's it generates a random seed and then does it. But if you control for this random seed and say every time I give you the same seed, you will still not get the same answer. It's because there's all kinds of things inside the execution of this the model and the model back end you know the cache and uh the tensor core libraries all these things have to be controlled very tightly to get deterministic

results and that's what we've done is with AI you get exact determinism so you run the model with this seed get exactly the same answer bit by bit right and so that is a really cool property because you can use that to then build because that's deterministic inference. You can use that to build verifiable inference because if I did something and then you know I put up my money and then I ran it wrongly, you can slash me and so on. So I think that is that is the coart of AAI is deterministic and therefore verifiable inference.

This I think you know what is the difference between compute and this there are two differences. Compute usually runs on CPU, right? And AI is running on GPUs. That's one thing. The second is the determinism problem that I just laid out that AI is has this deterministic thing. The third one is when you deploy a compute instance, you're getting an entire node operator just for yourself. So I I have like I've allocated a you know compute node just for you for doing your app. But you know if you're using AI many many apps can share the same AI engine. So I have an AI engine which you know can inference engine where you can ask questions and you'll get answers and if there are 10,000 apps you know each app using

inference rarely they can all share common resources through AAI. So that's the core core thing in AAI you can do an execution and then you know an inference call and then the inference result comes back verifiably on chain. Yeah. >> Oh, same primitive. >> And so, Shurom, um, is this going to exist in an IGEN AI chat interface? Because, uh, we've been working with Near for the last month or two. Been kind of experimenting with their with one of their private chat betas, which if you haven't seen that, hit up Ilia.

He'll he'll he'll definitely show you. And it it verifiably encrypts your data. Uh, and you know, their big thing is privacy, but it's kind of around a similar vein. what is the actual like product that I'll be able to use? Is is it going to be like an IEN chat GBT type wrapper thing with these primitives? How how is that going to happen? >> Okay, great great great question. Uh we can easily build uh you know we have internally uh fun uh wrapper but the main usage that we envision is an replacement for an open AI API. So instead of using the chat GPD API, which is what developers use in the pack, you can just slam in the IGEN AI API and then you get verifiable queries. You can

slam it into like a very standard chat interface and and use it. But we don't think that's the right use case because our focus is on verifiable inference, right? As opposed to private inference, you know, which is another use case that we'll support over time. But for verifiable inference, I think in crypto, people haven't hadn't quite found out what is a good use case for why would somebody want verifiability because there's always a cost to having verifiability, right? A chat interface, it's verifiable.

Yeah, you know, I know if I'm getting good results, I don't need like a chat interface to have verifiable inference. But imagine I slam a trading agent with inference. it's it's an agent that's going to do trades on your behalf and that needs an intelligence and an inference and if that inference is corrupted or like something goes wrong because you're now like making it run in an automated execution loop where go like hey think carefully about which you know token you want to buy and sell and all that and then like it goes and just does it by itself that better be verifiable because otherwise you you're going to get rugged right so it's important for us to understand that verifiability is a has a

premium right it's not just oh you know chat GPD is not clear at least for privacy it could be a very good use case right I want to ask a question where like nobody knows one >> yes >> but here because verifiable okay so what other things are so so inference is primarily an API like you can you can swap out an open AI API to an AI API just like that and you bring the results on chain. But we think the superpower of the IGEN stack shows up when you combine these pieces. I can da I can compute AI all these things. And think of an AI agent, right? What's an agent? An agent is like a computation, you know, a unit

of computation running and then it calls inference and it calls other tools and then it has its own memory, right? So you need IGEN compute, IGEN AI, IGEND DA, all of these pieces to work together to actually create powerful verifiable agents. So that's a category where you know just any one of these pieces by itself is not enough. You need the union of all these pieces to actually make it really interesting. And again the same question which is I think in my view I think more than the fact that we are doing verifiable inference or compute I think what we've done is to decode when is verifiability when is there a premium

for verifiability and then like doubling down on those types of use cases. A trading agent is a good example. It's obvious okay you know if it's a trading agent >> you don't want to like just put money onto something. In fact I think there's a lot of Sorry. >> No, I was just saying like you're not going to put your hands up and be like, "Good luck. Thanks. Trade my money with unknown parameters obviously, right?" Yeah. >> Exactly. And you know this is not just an inside crypto thing. M >> so you know we think oh you know maybe there's a last cycle you know AI agent meta and then like you know it just fell off because things were not yet ready right and I think this is you can think of this as like the next round where you know you can bring an Eliza agent you

can bring other agents and then make them verify actually have lot of TVL and value kind of go through it but even more broadly uh agentic finance is a big thing that's going to take over all kinds of other places right and I think existing frameworks without verifiability don't satisfy all the properties like for example if Grock offered a fully uh automated trading agent they are taking significant liability right imagine you legal yeah imagine you let your Grock go trade like stuff for your for you and it and Grock ends up buying a bunch of Dogecoin

for you and Doge goes down in price, you may be an afflicted party in this uh Elon, you know, game. And you know, it's a real it's a real issue because >> and so this this is I think one of the reasons why Chad GPD and Grock have not yet gone like all automated trading execution and so on because this is immediate liability center, right? Yeah. It could be GameStop, it could be Dogecoin, it could be a D that treasury like who knows how you're going to get rugged. So, and so that's whereas what we are saying is the the guarantee that the system gives the user is not that

the agent is good or bad or better or anything. It just tells you this agent code trades your money and gives you like that specific information exactly like Ethereum, right? Ethereum doesn't tell you whether the IG layer smart contract is good or bad. It tells you your money is with this piece of code. Right? >> And so that's exactly the same guarantee we want to give with verifiable trading agents. You think of this like a we think of like formally is a runtime like a verifiable runtime for these agents.

>> And uh so that's something that we're super excited about because it's not just a utility for inside crypto. It is a massive for a Robin Hood for like you know for a Grock for any of these things to say >> hey it's not me it's those things that are running and those you choose your code and like it's just going to run exactly as >> so that's one uh really interesting thing in the in putting these pieces together but our bet is putting these pieces together is going to let you do like all kinds of things which which are transformative to the foundations of crypto. So, I think that's that's what I'm excited about.

>> Good, good, good, good. Sham, just uh I know it's late over there and I I want to let you get some rest. Just curious like we've got these two products announced like where are they at in their development right now and then like what is the road to the end of the year look like for IEN compute and IEN AI? Maybe just give us a little bit of insight as to yeah what the absolutely >> development looks like. I think you know uh well the first thing I'll start with I'm enormously proud of our engineering team for like shipping this mainet alpha at such record time um the IGEN layer white paper to the IGEN layer like fully shipped was three years right IEN cloud white paper to like cloud v1

two months it's insane the our development velocity has gone up 10x to actually be able to do this. Having said that, I cloud's you know compute and AI is mainet alpha. So you know it is a very limited set of features right. So our compute each one of these I can compute AI has lots and lots of features that will get added in terms of security in terms of availability in terms of SLA all of these things are restricted initially. So for example to get access to IGEN compute you need to get whitelisted. So it's not general availability it is you know but you can fill a form and then like you'll get

provisioned. Um so that's you know the stage of compute it's running on TE uh and you know TE attestations are verifiable but not onchain. And so like there is all these like asterisks that go with the v1 of this system. And so the reason we felt comfortable in launching this is unlike you know a change to the IGEN layer contracts which is going to drive $20 billion of value in you know some application builder wants to actually use you know compute organ AI today when their value is very low and they grow with us right you know today when the value is low it's fine

like you know just go use it and then like get users and get iron out the developer experience this is one of our big shift from security first to developer first or you user experience first because we want to nail the experience of if I'm a developer I'm building a new product this is a new deployment pattern where you're writing a piece of code onchain just the token logic is returned onchain and then the remaining logic is returned in these containers and and stuff so new pattern so we need uh new kinds of uh developer experience for this but that's where the stage is it's mainly to get our developer experience tested get new

people thinking about this new application patterns so same thing with AI it is it is built on deterministic inference but the slashing for IG AI is not yet live so you know again that'll be a phased release we'll see some of these go live by the end of Q4 and again one of the things we want to do is as real people deploy real use cases on on top of these two things, we'll know which features need to be prioritized and so we want to take that feedback to actually go and nail those first. >> Beautiful. And I would expect the mainet alpha rolls all the way into the end of the year and that you know into Q1 next year we should be able to see some of those feature set upgrades really live in production.

>> Absolutely. >> Beautiful. Shirro, thank you for your time. Enjoy the week in Singapore. Say hi to Robbie if you see him. And uh yeah, man. Thanks for thanks for coming on and congratulations on on this next step. Look forward to seeing the evolution and circling back in a couple months and chatting again. >> Thank you so much Andy. Always a pleasure to be on the roll up. >> All right, buddy. Speak soon. Bye. >> Bye. All righty guys, that was a last minute appearance. I'm glad that we made it happen though. Um, pretty pretty pretty good stuff. I mean Shiram Ram is just a deep thinker uh that you know kind of has a a really really way with his words and thinking about uh applications runtimes AI compute the

kind of entire pipeline of infrastructure for AI and crypto. So you know if I'm looking at the AI crypto category right now you know it's hard to say that I layer is not leading um and I think I just want to put that out there. you know, there are some leading teams in the space and I think it's going to get more and more competitive, but I think I enigen cloud is going to be one of the leaders if not already. And so, just keep that in mind, guys. Kind of went from like this restaking ethaline thing to now Shiron's talking about we're going to deploy all of our stuff on any chain and I'm like think thinking to myself, I'm like that sounds pretty bullish. I think that sounds like a good idea. Um, so I'm just glad to hear that these guys are shipping fast. Um, I

think that was the push back from the market for them was like, "Oh, they don't have anything shipped. I da is still in production. It's doesn't work and this and that and there's no ads." So, I think like they're shipping now. We, you know, I think that the market's going to realize that. Now, it's just about the, you know, the demand side. Like they got the supply side, now it's about the demand side. So, guys, awesome stream today. See you guys tomorrow and peace out. Thanks for your time.