# DeFi Asset Management: Risk & Execution | Shyan Hussain, Jon Libby, Brian Huang, Brendan Maciel - Channel: [Ethereum Denver](https://streameth.org/ethereum-denver) - Date: 2026-03-09 - Duration: 25:36 - Topics: ETHDenver, Crypto, Web3, Blockchain, Event, Conference, ETHDenver 2025, ETHDenver 2024, Bitcoin, Ethereum - Watch: https://streameth.org/watch/yt-AZCMGNnrqvg - YouTube: https://www.youtube.com/watch?v=AZCMGNnrqvg ## Description πŸš€ Get Ready for ETHDenver 2026! πŸš€ We're already hard at work preparing for next year's biggest Web3 event! Keep your eyes peeled for more info on ETHDenver 2026β€”it’s going to be epic! 🌟 ## Transcript our moderator, Bran Masiel. She leads BD with Upshift, institutional BD with Upshift and she's gonna introduce her panel. Please, >> please come on stage. Please come on stage, guys. >> And they're going to be focusing on risk and institution and execution >> at scale. All right, let's go. >> Thank you and welcome everybody. So, as she said, I am Bren. I lead our institutional BD efforts at Upshift and today we're going to be talking about DeFi risk and execution at scale. So we've got a rockstar panel here. We don't have too much time, so we're going to try to get through all of my questions. So I'll let each of you quickly introduce yourself, a oneliner on what you do, what's your mission, and then we can get into it. We'll start with you, John. >> Uh John Libby, uh DeFi lead at Gauntlet. Um run the curation stuff. Hi everyone, I'm Brian Hong. I'm one of the co-founders of Glider. Uh we are building effectively onchain fidelity. We're backed by A16Z, Coinbase, Uniswap, and and many others. >> What's up, guys? I'm Cheyan Hussein. I'm the founder of Block Bites Capital. We are a digital asset management company based out of uh V Virginia and uh we're here to help the price go up. >> Awesome. So a lot of us have started our careers in Tradfi before coming into this space. So I'm going to start off talking about execution and coming from that background. How do you guys view the gaps in the market that you know are fully covered in the Trady ecosystem when it comes to execution? What's really missing to bring that Tradfi scale onchain? >> Right. I think in higher frequency trading, execution is always the name of the game and with DeFi, execution still seems to be more of an afterthought. Um I think you know most of the ecosystem does optimize things like TVL uh incentives uh yieldbearing products but I think where we're where we're lacking is uh really in kinds of uh mev defense uh risk mitigation and uh information leakage which I think is very important for institutions to really on board with DeFi uh and and yeah >> yeah I can I can take a stab here. So I started my career in high frequency trading at a firm called XTX Markets. And uh one of the biggest problems that we see today is in particular with equities coming on chain is being able to trade size. And so if you want to go and trade like a million dollars Nvidia on like X stocks or something like that, the spreads will blow out massively. And that is not acceptable because a million dollars of Nvidia is really nothing when it comes to traditional finance markets. Um there are some really nice primitives now around pulling traditional finance liquidity onto uh you know DeFi rails. Um and for example uses this mechanism for their liquidity so you can trade much more deeply but uh liquidity is still the the biggest problem. >> Um it probably is liquidity actually. Uh but uh I'll take a hot take and I say predictability. I think like um which is probably the same I think it sums up whatever we're trying what we're all trying talking about. Um a lot of these institutions they come on chain if they do anything in traditional finance they they generally have a very good understanding of what they're doing. Um but but on off-chain activity is not actually uh conducive for onchain activity and vice versa. They don't know how to deal with atomic liquidity. They don't know how to deal with CCTP transfers or or uh bridging via like query based OVT models. Um and so when they look at these things like um I think like definitely the biggest problem is like um if they get enter something they don't know exactly what they're getting into because of all this onchain risk. Um and I think one of the biggest things we can solve is like um walk in like a model of like a fra mental framework of predictability and um how do we produce a better outcome? Lending is a great example. Uh August Digital Shift is on Morpho. We're on Morpho. Um if you're trying to do lending on variable rate interest uh these guys it's very difficult for these guys to create leverage or um leverage their activity to constantly um um to get whatever the returns they want on their fund. Um fixed rate lending is a great example where we can actually start offering duration based yield um with actually um better interest rates for suppliers that um we could actually show like BTC USDC borrow and offer something predictable um similar they do with prime brokerage rates um while actually offering a better solution for these guys that are kind of on chain kind of not that can then activate their USDC more easily rather than going via um a custodian or another party like that. >> Yeah, you're explaining my day job for me. So >> just to add real quick um completely agree with you about the predictability aspect and I think that is a major point for institutions to start to come in and start using D5 protocols a lot more but I think there's also a lot of doom and gloom around the space in general and if we kind of zoom out and we look at the fact that we are essentially trading billions of dollars of liquidity through through products that are retail grade. I don't think that that is so much of a hindrance as it is an opportunity for people in the space to continue to build and continue to bring in new liquidity into the market. >> Yeah, that brings me perfectly into my next question. So, at Upshift where I'm at, we started with the idea that we were going to most likely be serving all retail depositors. That's kind of what our vision was. And then we looked on chain and realized that 70% of our depositors of our LPs were actually institutions. Uh institutions in crypto is obviously a broad term but a lot of it was executions deploying into these curated strategies, but that's kind of changed how we look at risk and how we decide, you know, what vaults we're going to list, what kind of strategies we're offering. So, I'm curious for you guys in each of your different areas of the market, how do you change your risk profile or how you view risk depending on if you're serving retail or institutional clients? >> Well, I think the magic of permissionless systems is you don't have to draw a line, right? So, um we're building a retail investment platform, but if you want to come in and log in with your Anchorage, Bitcoinbased Prime account, you can. And so we don't necessarily have to draw that line. And I think that's also part of the permissionless permissionlessness of uh of building on chain, right? Like previously you had to be an accredited investor to access these things. And now anybody for the most part can access these things. >> Um for >> Oh, sorry. You're good. Yeah. I feel like I think we're building for retail institutions. Sometimes it's similar, sometimes it's different. Like um glad I think acknowledged that with KYC accounts. Um I think like you first have to think about the risk they're currently exposed to. Um and uh and it starts rather than looking at the actual like maybe tokenized asset itself. Um it starts by actually looking at like what is how how does this activity work. Um so like for example bridging um auto compounding yield and uh various products like that I think is where you kind of start and then you you kind of bridge out there then the risk profile. And so first design products to sign that solve that solution um is how you really manage uh all these different kinds of like um onchain risks and then from starting that framework um I think that's like where you kind of build from. We have our own infra as well era that actually does all this for users on various various of our products. Um and then from there when you think about risk profile um it's really hard to differentiate uh when you look at something like a morpho or an era or a glider or an upshift what are these risk profiles you're doing and it's really about um designing very good education about what you're actually exposed to um and like um understanding like creating very strong terminology um for to help define these risk profiles um for example we made it up in a slack channel one day we called our vault prime core and frontier um to design risk profiles ironically a v4 now uses is that exact terminology uh which is hilarious and um and uh so you kind of start by defining clear terms that everyone can start adopting and using. We see that now being added on Morpho and from there then you kind of just like um then you start actually um figure out clear ways to understand like okay we solved all the basic execution risk you're thinking about about this activity and then you kind of think about like from there it's it's really about um using the like the terminology to actually build that layer of trust. A lot of people are not going to look into that the details of what this terminology means, but if they have something they they they understand as a core um as a core concept. It helps define from there where you how you can scale. >> So I can answer this question from the perspective of someone that is building and managing portfolios for clients in the digital asset space. And so DeFi for me has always uh it's always been uh concerned with how fast can we achieve uh you know whatever uh whatever we are doing uh with the D5 product. It's always been about velocity, but now that we are kind of entering a new paradigm for cryptocurrencies in general, I believe uh real wealth creation is based off of durability and having stronger infrastructure and you know while while swaps and all of that it's great for narratives but uh portfolio management is more about how we are controlling draw downs and how we are planning on allocating liquid liquidity and of course strategic rebalancing and you know there is a constant rotation uh in in crypto. People are always chasing the higher yield uh and always looking for the shiny new object. But I think once we start to get the infrastructure a little bit more a little bit stronger than what it already is, the opportunity will come on its own. >> I want to add one thing to what John said. I'll take it a step further. So, um, yes, I think it's great to try to start to explain the risk levels of these vaults, but fact of the matter is, and I know Upshift and Gauntlet won't like to hear this, but fact of the matter is no one really cares. And so, when we offer things to our users in terms of lending, like the name Gauntlet, Upshift, Morpho, A, these things don't mean anything to users. And so, we actually just don't even mention it, right? They want to know what the yield is, and that's what it is, right? Ultimately, users are looking for a better user experience. the Vault infrastructure sits at the infrastructure layer. It's not important. It's not a brand name that actually is meaningful to users. Um, the other thing that we think about too is when you go to Morpho and you see the 37 volts that they have, like how do you pick one, right? Like it's like I don't I know these curators by name. I've been in the industry for forever, right? And you just see the yields. Like it's honestly not the correct user experience that we're offering to people. It's it's a buffet of options that uh you need to take an opinionated stance on on how to offer it. >> Yeah. Go ahead. >> Yeah. I think um I think it's better, but I think you're actually I I you're you're you're only talking about retail. You're talking about one distribution line and you're talking about u one platform they're going to >> with the institutions are here. Um distribution lines are here, fintexs are here. >> Um the reality is they're going to be exposing their users to what they want and it matters to them. It's not a lot of the entrance of new new capital is not going to come into a protocol. It's going to come into a new distribution line and they're going to identify and create that for their user base naturally whether it's via retail app or whether it's via that this is already being decided on all these new distribution lines. Um that's like got to be very clear. Um because that's missing part of the point. >> I agree. But I I would say again like when you're thinking about mainstream audience of things, the brand name of a of a morpho or any of these vault creators is incredibly commoditized across the space. It's it's hard to differentiate and actually make a meaningful difference in that. >> Sen, did you want to add anything? You look like your head's nodding. >> Oh, no. I'm just I'm just enjoying I'm just enjoying the >> I love when we go back and forth. It makes it a lot more fun here. Yeah. Uh I mean I think you guys are saying a lot of you know the same side of a different coin or a different side of the same coin rather where we tend to sit in our echo chambers especially when we view like these are the risks I'm tackling this is how we're going to approach it cuz these are my customers and you know I'm listening to funds all day long telling me one thing and this is what they want. But then I'm also trying to balance how do I maintain the risk parameters that make sense for my users because again when it is a permissionless protocol you are tailoring it to one subset of users however anyone can come your way right so you kind of have to have all of these things in the back of your mind and it's important to have a well-rounded picture >> for sure I also think sorry I also think there's a big kind of fear that you know institutions are going to come in and kind of change everything up in DeFi I don't think that's the case. I think DeFi is going to force institutions to change the way they allocate for their clients. And capital, you know, doesn't doesn't move because uh because there's innovation. Uh capital generally moves based off of reliability and based off of structure. So, you know, as as he mentioned earlier, the predictability factor, the structure factor and all of that will definitely help uh institutional money to flow into DeFi more freely. Yeah. So, I think all of this is a good segue into we kind of already touched on this, but you know, Morpho keeps coming up, right? That's the hottest place right now. Um, and so Morpho introduced these isolated markets and that created a whole host of complexities. You were kind of touching on this, Brian. Gauntlet, you're a curator and you're an allocator, Scion. So, you know, when we're looking at now what we're building at Upshift as well, it's these curator managed vaults, right? So you're basically offloading the risk to someone and you're essentially trusting in them that they understand the risk for all of these morpho markets that they're going to allocate and rebalance accordingly. But how do you then actually look at those curators and assess the risk that you're giving them, right? Like how are you pricing that risk in, if you will? Um, and we'll start with Gauntlet. I know you guys are major curators. We definitely put a lot of our funds in gauntlet markets. So would love to hear how you take that responsibility on and then from you guys kind of how do you assess your curator risk? >> Yeah, I think the future of a lot of DeFi will be curator based. Um when you're thinking about curators, uh I think in some way people think it's fully commoditized. Um but it's definitely not. Um the reality is who loses money and who doesn't. And there's a lot of curators, especially the last like two instances we've had on Morpho between usual and stream that have in fact lost money. Um, and as you start thinking seeing how like the we're thinking about like when you're picking a curator, um, that's kind of our job. We do very strong diligence. We actively look at these things and we always make sure how do we keep funds safe at all cost. um you want carers that are supplier focused, have no interest on the borrow side and um in the sense of really trying to focus on bringing in borrowers rather than like they focus on protecting their suppliers and borrowers have to comply to that model. Um and when you think about curators, it's kind of like who loses money, how do they perform on on rates um on a you know a 24 three month nine month average on whether it's blue chip or high yield exposure and then um what is their track record on scale u and that's really what matters with the curator. Um, a lot of curators try to think people think it's easy. You'll be very surprised how many curators uh sit make a change once a week, every two weeks to the money you're putting in their vault rather than automating and changing these strategies. I've heard terrifying stories u of uh vault infras where it's um the curators aren't even curating. They're like handing to the vault infra like you can make three changes a week uh and that's like it or a month and it's it's insane. Um we are like always so when you think about curator you're think about like you're you're you're trusting a brand of of these different track records um this will this will differentiate more and more over time um and it'll become less commoditized actually I think in a long long enough time frame um we look at black rockck different than we'd look at like a typical fund and then from there it's like I think to build on that is like um you know it's it's what the curator can offer from there it's uh you know um layers of strategies that are bigger than morpho that are bigger than restaking or whatever the curation meta is. It's careful selection and trust and then having a full stack of resourcing to offer to whatever client or consumer you're looking for to meet whatever need they want. And that's how you kind of build vertical stacks that like target the distribution lines. >> Yeah, I think everything John said is is entirely what we look at when we decide what we want to integrate into the platform to offer users, right? it does go to brand new. Um but um when we think about the broader audience of like regular people who have no idea what crypto is um unfortunately these names even like a don't mean anything. Um the one thing though that we are sort of maybe counter to to vault and how we think about things is that we've finally built Rails so that you can directly own your assets, right? And if you really want to directly own your assets, you shouldn't be putting them into either an ETF or a vault because likely they are extracting some form of fee. they are some form of middleman. Now, if they are actively doing something really cool like which Gauntlet does, right, then it makes sense to pay them or have them take a fee from what they're actively doing. But to his point, like there are vault creators that don't do anything. And fact of the matter is you should directly own your assets because then you can lend them, stake them, participate in governance, claim air drops, like all of the stuff that they might be getting that you're not because you are in a vault that they're doing one rebalance a week. And that's like what what are you paying them for, right? It's like a wealth manager. And so direct ownership is really important, but um you know there are people who want other people to manage their money and it that's how the industry works. Um and so choosing your vault creator specifically for the reasons that he mentioned is is important. Yeah, I think uh you know just uh just based off of what John and Brian have said, you know, uh smart smart contracts are usually the headline risk, but again it is uh more we look at more of structural fragility in D5 protocols before where we are allocating for for our clients and um yeah. >> Yeah, I think we have a lot of interesting points across the board. At the end of the day, Brian, you're kind of the curator of curators, right? So, you said earlier that you're not looking at brand names, but at the end of the day, you are the one that's looking at all those names. I mean, I think one thing we've taken very seriously at Upshift is having everything beyond chain. So, you can click on that vault, you can see all of the exposure for that vault. We do have cross venue capabilities. So sometimes I think it's going to be really interesting for a market maker that's market making on Hyperlid for example. That's something you know maybe I'm not the best you know I'm not going to go make a lot of money shorting per I'm going to lose a lot of money. So I'm going to delegate my assets to someone else who knows how to do that. But I want to make sure that I can see that's what's happening. Right? So there are pros and cons. And if they're just running a vanilla carry trade then just like go by Athena, right? Um, >> amen. >> Yeah. Or don't. Um, but awesome. Now, I want to finish up on everyone's favorite topic, which is AI and automation. So, when we're talking about all of these risks, obviously, not like what's the risk in AI, but how do we kind of offload some of this risk into AI? Or rather, how are we automating things like rebalancing and actually executing your strategies outside of the vault meta? Um, so why don't we go the opposite direction? We can start with you. >> So I don't think AI is going to be replacing people such as myself, such asset managers like myself. I think it replaces it replace it it allows us to make decisions faster. I would say if anything, it it allows us to be more efficient. And uh you know everyone kind of jumps towards AI trading bots or autonomous agents. Uh but as an allocator it is more always always about the risk. Um as long as you're protecting the downside and you are correlating drift detection uh then you know uh with along with uh liquidity fragility modeling and that as an asset allocator for me that is more important than uh any kind of changes that AI would be bringing into uh into the space. >> Yeah, I think we've all seen many trading agent bots on on X and and various places. Um, you know, fact of the matter is there's no incentive alignment to ever share a trading bot that actually generates money with anybody else, right? Like that's a trading firm. I I've worked at a trading firm and trading firms spend massive amounts of resources protecting their alpha and how they execute trades. Um, so I would be very weary about giving money to an agent that is trading on your behalf because whoever created that agent, why aren't they using it? Why aren't they putting their own money into it? And why would they share it with you? it doesn't make any sense. Um, so just sort of debunking that whole thing. When we think about like AI when it comes to managing wealth and and growing retail wealth, um, we think there's nice ways to proactively nudge users towards things. Meaning like, hey, like um, last week people with your risk profile bought Ethereum and Bitcoin, like does that make sense to add to your allocation based on your risk preferences? Um, but things that are like LLMs are also really faulty because, and I can go on this about this forever, but when you use an LLM, you have to know what you want. And if you know what you want, you're not going to use the LLM. So, for example, like you're not going to type into a bot, TW me Bitcoin over the next seven days. You're going to go and click a button to to do it. Um, and so LLM themselves are not great for trading. They're great for research and like assistance, but in terms of execution, you know, there's so many problems with going from natural language to actual call data on chain accurately that then there's a layer of trust that also has to go into it that, you know, we internally don't think there's any product market fit in that. And many people have tried and they're nice gimmicks, but um yeah, we're not going to be doing that. >> Uh let me walk you through a really fun example of how we leverage automation um in our curation. um yield bearing stable year ago. Um everyone was hot on uh every what K was allocating to grew more hundreds of millions of dollars. Um it was a very exciting time uh last summer uh or 2024 I don't remember uh or spring 2025. Um but when this T bill based stable coin that everyone is like how it can be risky to T bill um was exposing themselves to on morphs we knew for a fact this T bill was not safe. Um but we also saw a lot of rapid growth from curators allocating 50 60 $100 million to this market with very little liquidity on it to scale it. Um we were thinking all right we are risk curators we are risk managers and our job is to get good yield while also prioritizing suppliers at all cost. How we thought about it is um we looked at the overall size of this market like 100 million plus available liquidity. We also tracked we looked at this yield bearing stable and we tracked a million parameters of it and we said how do we automate this exit and design an allocation in a way where we can get yield while also making sure we can react in a second's notice if anything happens. Um this usual uh this company uh made an announcement and um loved the announcement that they they're uh they're uh it's not working out. Um there's they're not honoring redemptions. It's a it's a big problem. We tracked the DEX liquidity, the redemption queue, everything related to the stable coin. And we also tracked the morphul liquidity in that market. And as soon as a percentage of liquid liquidity on the DEX pool dropped, a certain amount of the redemption queue um orders were not filled and a certain percentage of liquidity in the morph market was um past a certain parameter we were set up. We were automatically out of that market. It took us one hour. Every other curator the day after woke up, freaked out. We were sitting on our butts. We woke up in the morning like, "Oh, that happened. Cool." We already knew it was fully automated and um every other car now claims they were now exposed to this um had $80 million stock, a $300 million stock, $70 million stock and they all would have been gone if that company did not come in and start liquidating the markets themselves. Um we automate everything as that we can. We make it we take our suppliers ve our supplier capital very very carefully. We want to keep users safe at all cost. >> All right, thank you guys so much. We are past time, but I'm sure if you have any questions, you can meet all of the panelists out over there. So, thank you guys so much for joining me. Thank you guys for listening, and hope you enjoy the rest of Youth Denver.