Tokenization, AI Agents, and the Future of DeFi | Aryan

Transcript

I think that a lot of Troutifi uh funds and companies and institutions can benefit with their own business model from moving things that ought to be in a decentralized clearing kind of efficient environment to it. And I think that we'll see re real P&L improvements for those institutions as they start adopting those technologies especially when it comes to something like tokenized treasuries. >> Hey all, Blockworks co-founder Michael Aelo here. Quick break to talk about something we've just launched. Block Works investor relations. As the market shifts toward institutional capital, investors want more transparency, more standardization, and a higher level of professionalism. But the traditional IR model is slow, manual, and not built for

how crypto works. If you're building on chain, your data is already live. Your business is already transparent. The challenge is turning that into a clear, credible story for investors. That's exactly what we're solving with Blockworks IR. It's a single platform that brings together real-time analytics, branded investor portals, and hands-on advisory support so you can communicate what matters. If you're an onchain business looking to level up your investor strategy, check out Blockworks Investor relations at blockworks.com/investor- relations. All right, back to the episode.

>> Nothing said on Zerox Research is a recommendation to buy or sell securities or tokens. This podcast is forformational purposes only and any views expressed by anyone on the show are solely our opinions, not financial advice. Pocachio, Ryan, and our guests may hold positions in the company's funds or projects discussed. What's up, guys? Welcome back to another episode of ZX Research. Today, I'm joined by Arian, the head of research at CMT Digital and also longtime friend of the pod. He's already been on, I think, twice now. Arian, how you doing?

>> I think twice as well. Yeah, doing well. Thank you for having me, Bob. Good to see you. >> Good to see you as well. Um, I have uh some questions noted down primarily revolving around your 2026 thesis. I think it's always kind of scary when you get asked about your thesis because there's some things in there, Arian, that haven't panned out exactly how you thought it would be. But before we get started, um, were you at DAS? Did you end up going to DAS? >> I did. I, uh, enjoyed DS. Thank you for hosting at Dodo. It was a very successful event and uh it was great to see everyone in both the crypto native and also more of the trady finance side of the uh world coming together and coming to consensus on things. So thanks

for hosting. >> Yeah, I appreciate you going. Did you feel um a significant Did you come to D last year? I wasn't there. >> I did. Yeah. >> Was it different? Yeah, there's there's certainly more excitement in the room. Um I I think people could feel the actual progress being made in real time and and obviously I think that was evident in the caliber of you know guests from outside of the crypto native ecosystem joining and so that was hugely encouraging and and I'm looking forward to seeing kind of how DAS changes in years to come and becomes more ingrained in that overlap between traditional finance and the DeFi native side of the

uh ecosystem. Yeah, I think um I like I said, I wasn't there, but I saw that I think year on year the amount of conversations on tokenization grows up more like um increases more and more. And I think this time it was like almost only about tokenization except for like one HIP free um chat that Shondaanda gave. Shout out to Shondaanda. Um that's actually one of the big topics that I wanted to chat with you about. I know you guys are investors in a few tokenization and I don't know how much you guys can chat about it. Like I said, I don't understand compliance and I don't want anybody to yell at me, but um what are are you happy with how far we've come in terms of tokenization

um specifically in how it is involved within decentralized finance? Like do you think we have all of the infrastructure required to support all of the tokenization that is going to come on chain? And the second part of this question is what chain do you think wins out in terms of the tokenization race? >> Yeah, I mean I think we are from a market structure perspective um getting there. You know the the tokenized treasuries have really proven to be the first mover and and I think that's beyond a reasonable doubt at this point.

Um, you know, personally next I'm looking for things like equities moving on chain because that's what we saw with Galaxy and that kind of makes sense over something like mutual funds or or ETFs necessarily. I and and that sort of collateral ubiquity that is unblocked as a function of that is what we're looking for because yes I mean if we if we look at like the Ave incident with USDT um there have been DeFi hasn't necessarily been made for this sort of use case and so we're looking at a technology that was created for permissionless use within DeFi and now we need to make it work with risk management ment in the traditional

world, legal enforcability of liquidation across these jurisdictions. In the real world, a smart contract can trigger a liquidation, but it can't enforce the recovery of this tokenized loan which is backed by these treasuries or real world assets or whatever it may be. And I think that that gap is the broader gap for institutional adoption, not the tech. Um the tech I think really does what it was designed to do and that was, you know, to facilitate these functions permissionlessly. Do you think um Morpho has a better chance at doing what you said, which is like legal um going legally after people or like favoring

regulation because the curator model allows them to kind of compartmentalize who has to deal with what versus like Go Yeah, I do like that curator model because it does kind of extend or it makes very clear what the role of that curator or operator is and and it kind of extends their role into the real world as well. Um, and it it showcases who should be creators and who should be these sorry curators or operators. Um, because ultimately they're responsible for the collateral decisions. And so it's also their responsibility or naturally it should be their responsibility to extend that responsibility to offchain legal frameworks. Otherwise it's just a tuning of the onchain parameters for them and

is that really worthy of incentivizing or uh is that worthy of rewards or revenue? >> Yeah. Um, and then one thing you mentioned was tokenized um, treasuries leading the way and then you said equities. You'd rather see equities. Why not ETFs I guess is the first question there. And then in terms of equities, are you looking at like only public very liquid equities or do you think it would be interesting? Um, we've tried preipo purp I think as much as people would want it to work out. Um, yeah. What are your general thoughts on the equity side of the

>> on purp. Um I think we start to see things becoming per markets as it makes sense for them to become per markets. Um I think we've seen treasuries win, you know, in every dimension through regulatory clarity, sovereign backing, secondary markets, the precedent itself. Um and and that's still, you know, being created. We're we're still in the early innings of like the DTCC canton partnership. I'm sure DTCC will partner with more. Um but it's the important thing is it's post theoretical and I think equities are also post theoretical and we're seeing demand for equities um you know superstates opening bell SEC

registered public security issuance and we saw that as well with the uh partnership that they just announced and so if we if we take it back to why it wouldn't be ETFs I think it's about access and the usability of those collaterals which obviously everyone has equities or the people who are looking for putting those uh assets to work in uh DeFi would hold equities. Um I think also the the the structure and the market infrastructure is really aligned there. Another angle that I see is private credit. Um, you know, if we look at Sid Powell's thesis from your episode, if it's about yields, private

credit does present an interesting opportunity more so than treasuries. Um, and it's I think it's important to have a landscape that has varying levels of risk for varying levels of risk tolerance and reward. Um and the the challenge with something like private credit is yes it's non cryptoba yield and I think that's really important but it can't be easily composable in the same way as treasuries can be because it requires standardization and these are ultimately you know uh structures that are heterogeneous in in nature not necessarily composable and I think that's the delay that we need to work through is how do you make all of these varying uh risk turn uh instruments

composable with each other in DeFi. >> I um we're recently trying more and more I agree with the general thought that we need more non-cryptoba yield, but we're trying to get there. I think the most recent one I saw is reinsurance is like becoming more and more popular. Um, how much does the crypto risk/curator landscape need to change for this to for not because I from admittedly I am not very involved on the risk side either at Blockworks or anywhere else but from my understanding

um treasuries are relatively easy um because they're very liquid um versus like private credit is getting it's very big but not as liquid. Um and it's like you said heterogynous. One of the worst words in my >> the worst >> um uh how much does the risk side need to change within crypto? Um I don't know how much risk you deal with on your end like how much you look into some of the curator models but we've had a few issues obviously over the past month and we've had a few over the past I don't know decade. Um, >> right. I think it's part of that broader question of, you know, where does how close are we to real non-crypto beta yield that's extendable and scalable.

And I think we've spoken about the infrastructure layer around vaults and and modular lending and programmable collateral kind of being functional and and and scaling in real time. and we we talk about more first curator model and maple's model but there are these open questions that I think all contribute to that risk framing that you're bringing up um collateral quality is one under stress you know ETH has huge liquidity and so liquidation liquidity is huge the these RWAs increasingly uh come from these more disparate ecosystems and and we don't know how they're going to perform under stress we don't know the liquid liquidity depth of some of these

assets and so they're not really being extended to DeFi ecosystems yet. We also don't really have RWA specific liquidation mechanics and part a huge part of that is that legal enforcability across jurisdictions that we were talking about. Um and so it's not really the smart contract that's or the or the blockchain infrastructure that's contributing to the risk. It's that um real world uh processes that that need to be composable with this infrastructure. Um another one is Oracle reliability. I think you pointed that out in an earlier episode. Um there's we don't really have standardization for this. Um and I think that then feeds into the ability to underwrite these vaults from either a individual Lloyds of London style

insurance product or a broader you know how do we get broad insurance for these products for for stable coins generally in the same way that you have FDIC for dollars. Um, so I think directionally we should solve for that collateral quality first, you know, leaning into those institutional grade assets like tokenized treasuries and and the underlying. But as we expand to these more uh fringe RWAs or or collateral structures, we should really pay mind to how we're uh making them composable with each other.

>> Yeah. Um, I'm glad you mentioned the Oracle thing because I think Oracle is Oracle design um is one thing that not a lot of people think about um on their day-to-day but we saw how much more attention it can get well specifically with HIPP like every single um HIPP3 provider has their own Oracle partner that has designed a specific oracle. A lot of oracle design work goes in in these lending markets. I think a lot more than people realize, especially on Morflow from my understanding. Um, one of those issues might be the data side from what I understand if something is

offchain. Um, and like you said, it might just become more and more difficult to liquidate things. You know, the easiest fix which we've started we've tended to do is just to peg things to like one or to whatever it is. And I imagine we're going to go for that for a bit more um >> before we can figure out exactly how to bring um everything on chain or the pricing of >> we we see that happening. I just worry about the reflexivity that that introduces into the system because people do loop these assets and having that sort of peg being the last frontier of defense against some sort of catastrophic unwind I just don't think is very prudent.

Um, I actually wanted to talk a bit about looping with specifically this. Um, if we're bringing I guess we're just going to say it depends on the legal and the jurisdiction side of the curator. Like that's what it is. Because if you do end up looping it free crypto people loop things 10x like they don't care. They'll do it as much as they can. Um, I guess it just comes down to the curator handling it. Right. >> Right. And I think it's about representation of that curator strategy and the risk tolerance that's associated with that. And maybe that's a break right now. Maybe we don't see a accurate representation of the underlying risk in

some of these strategies that are packaged off and and sold. >> Yeah. Um Okay. I want to flip to I think you're the last man standing on the Ethereum side or you were for some time. I don't know. Crystal, >> what do you mean by that? Like I'm a defender of Ethereum. >> Maybe not defender, but I guess like because I asked and we carefully avoided this question which was um who do you think wins the um tokenization slash bringing securities onchain um race?

A lot of people have historically said Ethereum or Solana. I'm recently hearing more and more people. I have to give a shout out to James Kristoff here because he said this like five months ago and he was like it's Canton. Um what do you think? What's your take? And then we can do your Ethereum take. >> Well, I think the Ethereum take is part of this take, right? Because Ethereum is becoming a settlement layer and I I wrote a paper about this or an opad about this. Um, and I think while that's good for Ethereum's original vision and the technology of it becoming the settlement layer for the world, I think it's also part of the problem in that we do have this L2 liquidity fragmentation issue uh that is increasingly pushing Ethereum into the background of just

coordinating. Um, and I think that that reality risks Ethereum becoming the infrastructure that everyone uses but doesn't necessarily notice or or or value. Um now if we look at a volume perspective, Ethereum certainly still leads on stable coin volume and you know RWA interest. Um but the computation and the userfacing activity has really moved to these rollups and so while the liquidity can still be coordinated by Ethereum it may use may lose that narrative but ultimately the fee revenue which I think is so important for Ethereum's future. Um, and yes, I think you're very correct in saying that the right sort of composition isn't

necessarily Salana, but someone like Canton who is working with the Goldman Sachs, Citadels's, DTCC's, BNPs of the world as backers. And so when we look at why do these firms choose someone like Canton, I think it's really about the design flexibility or or handholding to an extent that's offered by that foundation to these institutions that are building. um if they can really tailor the environment to meet the privacy requirements or the institutional compliance requirements that these that these public entities need for chains and and someone like Ethereum that public transparent chain can't structurally satisfy then the then the institutions can't meaningfully get

involved. Now do I believe that Canton's model is the best for open finance or for you know a scalable ecosystem on its own? I think there are some things that are a barrier to that in that you can't see very easily how much activity exists on these subnets on it on Canton. You can't see overall volumes. You can't see counterparties or liquidity movements publicly. Um so I I I do think that Ethereum likely wins on settlement but the value acrruel may not flow back to ETH holders the way that we had historically expected and and it's because of that invisibility associated with the coordination of the liquidity. Um just a

disclaimer on this point we are invested in temple which is canon you know decks. So and everything that I say is not investment advice. We are >> this is for generalformational purposes only. I'll >> absolutely >> um very uh so I will first of all admit that I was one of the initial people who saw Canton and dismissed it. Um, primarily I think because I saw some of the ways that they were doing the validator system and in the early days it wasn't as serious as it is now because now it's like oh we're onboarding like Visa and Mastercard and they're like super validators and before it was like and this is in no way disrespect to um some of the companies that they worked with initially but it

more so felt like oh this is just like relatively large chain that isn't really putting in a lot of effort um and it's just like kind of using brand names to put itself forward. Um, my perspective has shifted drastically since then. I think partially because the price kept going up and then partially because they kept working on the chain clearly and then bringing on more and more people. Um, on the Ethereum point, I kind of agree on the sense that I think Ethereum probably becomes a settlement layer, but I also agree on the sense that I don't think it's bullish for Ethereum. I think Ethereum just kind of becomes um I don't know Unix or like Linux like something very very base. It doesn't make any money.

It's just kind of like you know how um Apple covers a like they pay Linux a lot because the foundation is very important. It kind of just becomes something like that probably at some point. >> Yeah, I mean that I could see that happening. Linux fixed systems, man. I uh we might see a resurgence of Linux as Windows and Microsoft continues to shoot themselves in the foot. But um I I I do think Ethereum's current roadmap in the next year or two is really important in determining whether or not the value is going to flow back to the ecosystem because if they are successful in selling proof verification broadly and people want proof verification to extend the underlying security then yes that

makes sense to me and that could that could be a value acrewing mechanism. It's just historically that's not what L2s have shown to be building towards or demanding. Um they seem to be fine with the trade-off associated with security in not being a true L2 and that's kind of what led to this divergence in in road map and and dispute from Vitalic. So yeah, I think it really depends on what the L2s value uh or are willing to pay for in the future. on that. I think that um first of all, I think like most people would agree on this is that blockchains and the way that they earn revenue is not a good way

to earn revenue like charging directly to the user. Um I think all of them realize this and it makes more sense to charge applications rent or some form of like subscription fee. Um do you think I guess Ethereum tried to do this for quite a bit of time but it just never monetized it enough. Do you think there's going to be changes in the ways that other chains approach this as well? Or do you think like Ethereum is going to somehow fix the way that it charges L2s or do you think that this is in any way a reasonable way to monetize the chain or do you think we're just going to go with the transaction based version going forward?

>> I have a bit of a unconventional view here in that I think L1's are ultimately just selling blockace and that that blockace should be differentiated. They should each be selling a different type of product. And so an application building on top of an L1 or an L2 building on top of an L1 should choose the L1 whose design principles align best with the application that they're trying to build and therefore naturally they would be more willing to pay for the products that that L1 is offering. I think part of the issue has been every L1 has kind of taken the same approach and hasn't really experimented around what the core function of their clock space that they're offering is. And it's it's stayed in like this very simple I

would love to see more consensus mechanisms where their the fee model is based off of something other than maybe base security. >> Yeah, I I argued with Daniel Smith about this because he argued that Hyper Core wasn't really a blockchain for some time and I think it is a blockchain. Like obviously there's a blockchain in the back. No one's arguing about this, but >> the EVM or the core? >> No, no, the the the main decks. I think that just because something is separated into like it's not called transaction fees, but it's called maker taker fees.

I think that's still gaps like you're still paying gas because you you have to pay a fee to submit a transaction. >> I would agree with you, Bob. >> Thank you. And I think that is probably what a lot of the other chains should have done as well. Like I think Solana should have probably done that as well. Um but for Ethereum, I don't know how they're going to make it work, but I feel like that's the differentiation at that point. I do think there's some differentiation to be fair. Like Ethereum is I think the most differentiated because it's clearly not going for speed. Everybody else goes for speed. Um and it's just going for like, oh, we're really slow, but you can trust it.

>> Yeah. >> As much and the head of the chain is going to be secure. you're not going to have a real attack. It's not going to hold. I I I think that's completely fair. I mean, if we think back to 2021, Falcoin, Aweave, they were selling something different. I mean, it just so happens that they were selling something that there wasn't a huge amount of demand for at the time. But that doesn't mean that there's not a whole universe of things that you can sell that make sense in a distributed capacity. You know, Double Zero is quite interesting in that sense. um bandwidth where we're where investors in optimum and and they're doing the P2P side. I think there are so many different things. It's just what can you come to consensus on in in the case of hyper core you're coming to consensus on that make a take

a fee. Is there value in that? I think absolutely it's it's it's taking that function of making and taking and making it broadly scalable on this first principles of what should be the value of making and what should be the value of taking or what should be the cost of having your transaction execute or your trade execute in this case. Yeah, I feel like most of the time the differentiation has been like oh we have a we're using a different language and oftentimes it's like important to an extent but like some of the times it's also I don't know everybody was launching like oh you can do like the big one was move obviously and then arbitum stylist came out and was like okay well you can just use any library you want for whatever you want um >> yeah I mean we see so many different I

mean canton's in like a variation of right um it's what language is best for the function that you're trying to provide or what language is best for the logic of the applications that are trying to build on top of your environment. >> Yeah, we talked about I I imagine most of the time it's Rust. Like I don't know what I don't know enough about Haskell and I feel like there's definitely going to be one person upset with me about like why I don't know how important it is that they have. But um I have >> EVM is moving away as well. Ethereum is moving away from the EVM. So >> they're going to risk, right?

>> Yeah. A risk five style approach, which makes sense with the ZK uh verification road map. >> I still need to look over >> B. Sorry, Risk B. Please don't come after me, guys. >> Arian's Latin, so it's very difficult for a B Roman. Um I still haven't looked over this ROM up. I need to do it. But I don't want to talk about Ethereum as much anymore. Are you want to ask you about agents? I don't know how interested you are with them. So, we can avoid it if you're not, but I imagine there's some interest on your side about AI agents in crypto.

>> Yeah, there is some interest. I mean, I think there was a huge period of hype associated with AI and crypto and I think that might have done more damage than good. um for the intersection of these two technologies. I think ultimately the framework at looking at AI and crypto should be what can AI benefit from being in a distributed environment and compute doesn't really make sense to me in that sense. I mean I I mean I love the work that Noose is doing on you know DRO and and and getting these uh decentralized compute frameworks to to be as viable but it's not the immediate option that's

obvious as resulting in like a net increase in utility or efficiency or something else. I think crypto and and blockchain specifically do give agents the opportunity to transact autonomously, you know, 24/7 across these counterparties that they've never met. Um, I think that is a huge strength of having not a human in that loop. Um, and traditional payment rails, you know, typically have those compliance checkpoints and banking relationships and hours of operation that are pretty hard for agent to agent commerce to actually uh navigate. Um but I think that three things are still holding back that adoption of uh crypto and

blockchain as a coordination mechanism for these agents. I think gas abstraction is huge. Um agents are still having to manage gas to a degree and it's a bit of an implementation failure. Um they also don't have you know deterministic execution or they don't have reliability on that execution. um unpredictable finality on the execution breaks the agent logic because it can't necessarily wait for that undeterministic function and then there's also no identity you know there's no standard way of having reputation or credentials or accountability and I think that's why we see things like X42 coming about because it fits with this legacy framework of

what has been done before um so look I our roots as a firm are within trading and executing thousands of decisions per second without human intervention. Um, and agent economies are just kind of the next evolution of that same logic applied to cryptonative rails. And so I think how do we extend the functionality of agents without introducing those security considerations or um security risks when it comes to agent execution? I think that's uh really lying within gas abstraction deterministic execution and agent identity.

>> So okay I think that makes sense. I have two questions on this ignoring what we need to build. So we need gas abstraction deterministic execution and then identity all well gas abstraction is not very difficult. It can be solved. Derministic execution I don't know enough about to even give my opinion and then identity I know is very difficult to solve and it's going to only get more and more difficult. >> I think deterministic execution is pretty much solved um or or can be solved through things like preconfirmations um and and the work being done there. I think the absolute hardest is that identity piece. Absolutely. Because there's no standard way of approaching it and they really do need to be on the

same page for it to work. >> Yeah. Um, on the identity bit, I've seen in my time in crypto, which is admittedly not very long, I've seen multiple teams try and build out some form of an identity platform. And oftentimes it's been actually not agent focused, but it's been people focused. So, it feels like a lot of the time they were trying to solve for this issue of, oh, is this um like five accounts or is this just one account? But how does it focus on the like how does a proper identity system help on the agent side? Is it just that I know that like agent X is Audience and

therefore it can be trusted or how does this what is the vision here? >> Well, there are so many different approaches to it, right? And I think that's part of the problem uh of of standardization. Some take it off of piggybacking off of um an existing identity using something like ZKTLS and and kind of using that as an anchor to reliability or trust. Ultimately, if we step back and look at the problem of like what are agents trying to do? Well, they're trying to transact autonomously. And so, what do we need to verify? We need to verify their intentions in their action. And so maybe identity looks different in in this in the agent realm and it's not the identity of the agent and who it maps

back to or how much money they have but it maps to their actions and whether the uh representation of what they want to do actually matches with what they are trying to do. Um, so maybe proving that they are honest and trying to achieve some function without revealing that function and that that function is not a malicious function is enough. >> Yeah, I think that makes sense. It's kind of like law enforcement for to make sure they're doing the right thing. I have a um James uh set up open claude a few weeks ago. He he actually asked his wife to set it up for him because his wife's an engineer. She sets it up and then James is tells the gives it like

some money and says, "Hey, can you buy like X amount of the token?" And it ends up just sending all of the money to like another one of his wallets for some reason. All of the money in the wallet and he like asks it, "Why did you do this?" And it's like, "Oh, I'm so sorry. I wasn't meant to do this." And I guess that's exactly what needs to be solved, right? Like you interact with a specific contract about a specific amount of thing. Um, and a lot of that I feel like is going to be very very difficult to solve because I'm thinking about it now and even in when I'm using not agents doing stuff autonomously, but if I'm just interacting with AI like single interface just doesn't do what I want it to do and it doesn't have any context to get confused about. It's not interacting

with anything else except for like a text message system that they set up. Like I imagine it's just writing in markdown at that point. Um, so it's going to take some time. Um, do you think that it has to be in crypto? Because I don't think it had to be in crypto and I still don't think it has to be in crypto. It can I can't imagine Visa's like, "Oh, yeah, we'll just let the crypto guys out." >> I I don't think it necessarily has to be in crypto, but I think there are benefits to having this system structured in blockchain environments because of the autonomous nature of the environment. Um, I think there can be more if then statements in a crypto environment for these agents that allows the blockchain to actually coordinate them better than a non-crypto

environment would. Yes, you can absolutely have APIs and set up to some sort of payment gateway and and have that process, but would that payment gateway then be able to continue to coordinate that agent's behavior or or you know continue to create that feedback loop of agent action verification payment and then you know updated state on which that agent can then make another decision. I think that's the benefit of having crypto systems. you have this updating state that is always feeding back to the agent and giving it more information which then helps it decide its next action.

>> Yeah, I think that makes sense. I have a less sexy answer for this one although yours is a lot better. I think >> it might do well purely because of the same reason Hyperlid did okay with equity purification. is just it's going to take so long for the big guys to move on. And it's like if you're a team of >> Yeah. There's no like technical regulation. You don't have to do like a lot of super like go through a bunch of steps. Make sure your boss approves and then make sure your boss's boss approves and you can just build it out. Um how likely is it that tempo just ends up taking the whole pie? What do you think? Hey, I mean I think the Tempo team is incredibly strong and and

you know it goes back to these L1s just selling differentiated block space. If they can make their environment extensible or extendable to all of these different uh ecosystems that agents want to transact with, then yeah, I think it's in a good position. Um, Ethereum's just got the liquidity uh advantage right now. And I think that liquidity advantage does matter when it comes to agent execution because it's very black and white when it comes to which is the better execution environment. But if they if the tempo team can build that uh connectivity to all of these other ecosystems and also build in the updating state associated with accounts,

then yeah, I think it's very very well positioned. I do I do want to give a shout out to the Tempo team. I think they're very strong. >> Very good. I also think >> I mean they're just they really I think Tempo um is probably scary to a lot of new stable coin chains but not just to them but also to I think everybody who was trying to compete in the AI agent space. Um although like competition there hasn't been as serious as it was and serious here being that like a lot of people were shipping things. I don't know how serious like Olas was if you remember back in the day. Um like they were doing >> Go ahead.

>> I think that was very necessary kind of experimentation that was going in the right direction before um narratives kind of took hold and uh virtuals kind of dominated. >> Yeah. Um >> very crypto. >> Very very crypto. I have um kind of tangentially related. We we chatted about this a bit before. Um, how have the So, you work both on the research side at CMT Digital, but you also do work in investment as well because you guys in investment. Um, how has AI uh specifically? I would say like I assume you're a Claude co-work kind of guy. I'm

hoping you are. Um, how much has that shift been? because I feel like if you were an early adopter 2 years ago, um I would still count that as early. Um things have recently gotten very very intense. How has your research process? How has the team's investment process shifted? How much less time are you spending on doing something versus like more time you're spending on something else? >> Well, it's a lot of automation. Uh we have been building a AI system internally. Um our uh team is fortunate in that we've got two very skilled uh AI and data guys who who really you know work on these systems a lot of a lot of the time. Um and it's just been very helpful for us in automating

processes that were previously quite labor intensive. Um, and I think that will continue to proliferate in the sense that it allows us to increase our velocity of deals, like our velocity of capturing opportunities, our velocity of being able to do diligence them and make decisions in a quicker time frame. In in terms of research, I think it's helpful in the data pipelines and and kind of surfacing the right sort of data without necessarily having to comb through it.

my my uh schedule used to be a couple years ago that I'd come in and log on and check out DeFi Lama, check out the block explorers and see kind of what was going on. I don't need to do that anymore. And I think it's that continuous um tooling that allows for monitoring of onchain data, passing protocol documentation, cross referencing our prior research, which is then hosted in a centralized location, which then gives the agents more context on a continuously updating basis. Just allows us to interact with them and give them the tools that they need to be able to help us more quickly. And I think it all goes back down to that velocity of deals

and our velocity of decision making. >> I think that makes sense. So it's a lot of the times it's a support thing. Um and it's just like cut down I assume eight hours of work to like 20 minutes. >> Yeah, exactly. One of the I don't know if that's an issue for you or if it's an issue for the other listen for listeners as well but um one of the issues that I have is and I think this happens a lot when you do you know that meme where it's like computer are you alive and the computer says yes and then I feel like a lot of AI is that and half the population is like using it as you guys are which is automating a lot of tasks that used to take eight hours into 20 minutes and the other half is like relying on it to make

a lot decisions, but it's giving it the AI its own biases such that the result that it gets is actually not a good result. Um, it's kind of like you decide on like a valuation you want at the DCF and then you work backwards from it. Just doesn't make any sense, >> right? We give it a um context. So that's part of the research portal. So So we'll we have this portal where all of our feces, all of the research on various different ecosystems are hosted. And that's kind of the knowledge that the agent can tap into to know how to think about certain things. And so instead of bringing its own bias, it can tap into ours.

>> Okay, good. So you um Okay, I have a mandate to mention Hyperlid once an episode and I think this time around we're going to talk a bit about HIPP. Um specifically, I want to know how successful you think it's been. um and what you think has been wrong. Um and what you are looking forward to. I don't know if you have a lot of opinions on HIPP4 as well. I feel like that's relatively new, so we might maybe get to that on your next appearance if you'll ever come on again, but um yeah, general takes on HIPP3 would be very much appreciated.

>> If you have me again, I'd love to come on again and talk about HIPP4. Um, look, my HIPP3 I think was actually a really uh beautifully designed mechanism and and I did write about it and I didn't really have a huge amount of concerns. My biggest concern was around I mean maybe a small principal agent problem that can come from the original HIP3 market creator pivoting to a different market when the depositors didn't actually align on that to begin with. But that is such a small, you know, consideration. Um, the other concern was really around whether this would generate durable volume or just this spiky speculative noise. And I think we're seeing kind of both. And and the data is telling a bit of a nuance story.

Um, volume has surged from like less than a billion to somewhere between I think like 15 and 25 30 billion on a weekly basis rolling it. And and that does point to like a real adoption signal. Um but I think you were the one who mentioned tension around revenue not being scalable um or durable. Part of that I do believe is like a design decision by them to have massive volume expansion but with no real revenue expansion is as a trade for them short term. It's probably right. you know, they're building this liquidity formation habit in the ecosystem where people look to HR3 for forming liquidity

as opposed to other ecosystems, but it's not yet self-evidently sustainable. Um, another interesting point is that lighter observation. You know, we haven't really seen huge bumps from HIPP3 trading on lighter which suggests that that volume isn't spreading to adjacent platforms as clearly or as cleanly as the bull case had assumed at the beginning. Um, so I think that's something that's worth monitoring. I do think it's also very early and the initial HIPP3 ecosystems old markets were going to be experimental to a degree. Um but I mean going back permissionless market creation we think is great. Skin in the game via hype staking we think is great use of hype and a great value of

hype. Um again takes that core first principle of maker and taker as being like a state change and taking that to broader market creation or broader markets. I mean it it's feasible to me why that results in the expansion of hyperlquid ecosystem. Um it's just a question of whether the cadence of market slot auctions is creating the right builder incentives right now or whether the governance on slashing clarity is clear enough right now for comfort for these market creators. I think all of those things are still open questions but where I land now is that it's a qualified success. the the volume thesis

was validated. The revenue model needs a little bit more time for uh determination on whether it's sustainable and and these governance gray areas remain. But um I'm not ready to call it fully proven yet. >> I agree on the the the thing about the revenue is I think they did the right thing because they were very clearly the re leader in um uh crypto trading on chain like very clearly at that point. perpetuals. They were the leader. They were not the leader in equity burps. Um and I think they had to go out with this growth mode which is 90% reduced fees.

Um I need to check I actually should have just checked this beforehand but I need to check because the gold and the mass um micro strategy volumes on hyperlquid are not um they're not reduced fees. So I I need to check whether like compared to the other guys those are as high or like higher because that could give kind of an insight on like whether it's achieving that whatever moat such that people want to pay higher fees they don't mind just because there's liquidity. Um I was concerned about fragmentation and I think a lot of people were concerned about fragmentation but it seems that trade XYZ is just winning it out. So

fragmentation isn't really an issue. If like one guy or like one team controls 80% 85% of volume, it's not really an issue. There's other >> more I think it's like 91% of open interest. >> Yeah. So like if you have some guy controlling not some guy a very good team controlling 91% of open interest, it's not really that big of an issue people because people just >> Yeah. Is it an issue? I think they're just showing that they're creating the markets that people want to trade and it it should be a survival of the fittest type thing. if you you either create markets that people want to trade and have open interest in or uh you don't.

>> Yeah, it is. They also did just pull off the S&P 500 thing which I think is very very impressive. Um I think one of the biggest like probably the biggest equities on chain news I would say maybe this year. Um I'm sure there's bigger ones that people point me out to. I'm interested to see if that um consideration in giving the data feed to XYZ and Hyperlquid is going to be an exclusive thing because right now yes they're the only one but will that persist? Will there be other ecosystems that also receive that same sort of price feed? It wasn't clear to me that that was exclusivity baked into their relationship. Um and so I do expect to

see them partner with other ecosystems as well. >> Yeah, I think that makes sense. Maybe it is just exclusivity on hyperlquid which is I mean there's other markets offering S&P 500 style products like markets by kinetic has US 500 which was actually doing quite well beforehand but >> yeah but they all have their own pricing mechanisms and their own data kind of feeds that they're constructing. So I think there's nothing as trustworthy as the official feed for sure. Um I just doubt that it's going to remain exclusive in perpetuity.

Yeah, it wouldn't make a lot of sense for the S&P to just only get one client. Um, okay, before we end things off, I have uh I forgot what I used to ask what I asked you last time. Um, I'm going to go with what is your vision for the next 6 to 12 months? What are you most excited about specifically? Um, and then what are you disappointed in since the start of the year that you want to see an improvement in? And then maybe we can plug your Twitter, although I'm not sure if you have access to that yet or for still on a different Twitter. Was it your Twitter or your Telegram that keep kept getting >> Dude, my Twitter has been hacked for about three years. Um, I I set up a pass

key in the process of it being hacked and so I was able to lock the hacker out, but it's no longer my email associated with my Twitter account. And I've tried, if anyone is watching this and knows Twitter, please help me because I've tried absolutely everything. Um, like I'm most excited about this collateral ubiquity p piece and the continued kind of intersection of Trady and DeFi. I do think that DeFi provides meaningful opportunities for, you know, efficiency gains for the Tradfy ecosystem. I think that a lot of Troutify uh funds and companies and institutions can benefit

with their own business model from moving things that ought to be in a decentralized clearing kind of efficient environment to it. And I think that we'll see re real P&L uh improvements for those institutions as they start adopting those technologies especially when it comes to something like tokenized treasuries. Um, you know, I was really excited about the, uh, Invesco Superstate deal that landed, you know, three days ago. Um, I think continued partnerships like that is what really excites me. Um, what I wish had happened, I wish Maades had had a huge resurgence.

I I >> What did I tell you at the time? I said, why? Like, I bought a M. I said, why? But it's okay. It happens >> because for the for the culture and I think what I wish is that DeFi leans into more of its uh uh offerings from a first principles nature. We it's been a while since we've seen something that only DeFi can uniquely offer for Tradfi. And as we're seeing this liquidity move on chain, I would love to see builders think about, you know, what can we use the infrastructure to provide that traders can't replicate? And I do think there's a lot of functions within that that uh are up for grabs. Identity is one, but you know, how do you structure

it? >> Yeah, I I agree with you. The last one I agree the most with. I think identity is one less fringe and probably a little less exciting. I think it's a very big next few years for lending. Um, and I think everybody agrees on this like there's no lending and lending adjacent products. I think Pendle or like any funding rate um any rate uh swaps um and then yield I actually think yield aggregators might do okay like if they can figure to do stuff. I mean from a credit perspective or from a lending perspective the flywheel associated with the data collection underwriting and then lending is so strong it just hasn't been utilized. It hasn't been the case

where the lending activity is feeding a data model that is being used for underwriting and creating that circular flow. I think that circular flywheel can be really powerful and ultimately results in better underwriting for everyone involved, less risk for the lenders and better rates for the borrowers. >> Yeah, I think one of the issues we've had is actually people who are dealing with the markets aren't we get a lot of people from like a lot of money has been made by trading products, right? So we get a lot of people who used to be in high frequency trading or people who like me is also one of the big examples where like there's a lot of money to be

made. So back in the day a lot of high frequency trading came into me but we haven't had lending has is like good business from my understanding but it doesn't make a ton of money. We haven't had like anybody who is I don't know what the archetype for people who are really good at lending is but I imagine it's just people who know how to do risk and underwriting. um we haven't really had those people come into crypto. Um we've had some very good people and the block works we have our own advisory side and we're working with some of these teams but I feel like we haven't had do you know what I mean like a surge of like people saying okay like I'm making this much at my institution now but I could make triple this over like two years or like over

>> I think that's happening. I think builders are building it. I think it's a a huge portion of that is forming those relationships with the historical lenders and getting them comfortable with these onchain systems. I think that's all work that is currently being done uh by companies that are starting up and um I think we expect to see some of that lending activity being offered within DeFi environments through specifically bolt products in the notsodistant future. >> Yeah. Well, Arian, this was a pleasure. Um, I appreciate pleasure >> is all mine.

>> Thank you for having me. Well, always a pleasure to see you. >> Always a pleasure to see you. And if you want to chat with Arian, you can join the ZerxPer chat where he is there probably more than he wants to be. So, I feel like a lot of people are there more than they want to be. But if you want to chat with Arian or if you want to ask me questions about the pod when this goes live, which I think is first Friday of April, um, come join that. Arian, again, thank you for coming on. I will like to have you when HIPP4 goes live so that we can talk about that. Um, yeah, perfect. >> I'll get researching.

>> Very good. And to the listeners, thank you for listening.