A short introduction about me. So, I left Google around 2018. I have always been passionate about cryptocurrency and have been creating on-chain products ever since. On-chain natives, DeFi natives, DeFi trading solutions, Digex score, if any of you remember. For the past few years, my team and I have been developing a product called Lighthouse—essentially portfolio management for on-chain for on-chain asset managers.
I am also a partner at Nemo Trading, which curates on-chain storage curates on-chain storage . So today I invite you with me into the depths of blockchain data. My team and I have been studying on-chain data for years, and despite the industry being 10 years old the industry being 10 years old , we still believe that the main potential of cryptocurrencies lies in the transition from raw data to valuable interpretable information. And I would like to start with a story. This is an on-chain story, and this is my favorite type of story.
Our main character is the address on the left. Let's assume it's a person, although it could also be an agent who knows also be an agent who knows . So, this hero, this depositor, executed a smart smart contract transaction in the specified block. And for those who are not used to block times, we are looking at November 4, 2025. We know that this user has deposited about $4 million in USDC into the protocol.
Of course, as you can guess, immediately after the transaction, it goes to Etherscan, right? Which I think right? Which I think most of you are familiar with. He goes to Etherscan, and on the page we see that it's probably an asset manager, right asset manager, right ? He manages over $4 million.
He looks at Etherscan and sees that for his 4 million USDC he received 3.5 million ERC20 tokens. There is no name, no price, no icon. Financially, it's not very clear. So he does what is logical.
He goes into his wallet, looks at the interface, and tries to get more context. Now these 3.5 million anonymous ERC20s are being interpreted. The name, icon, and price appear. We don't know exactly where the price comes from, but let's assume it works.
And this is getting a little more interesting, is n't it? Financially, our depositor, our hero understands: okay, I have 3.5 million tokens in this vault. There is an icon, and I know it is worth about $4 million . But our users want to know more.
So now it goes into a specialized DeFi specialized DeFi application, it could be Zapper, DeBank, or the interface of the protocol itself. And here's where it gets a little more interesting. He no longer sees the token that was there before. The token is interpreted as a position. Now this is a financially meaningful abstraction, built on the data we've seen so far.
This is a position. Now we know that these are assets in the Morpho repository. We know it generates revenue, and we know that USDC was a deposit. So, our user, having switched from Etherscan to a specialized DeFi application, better understands what is happening from a financial point of view. Although this is literally the best we have at the moment.
So, if you are an experienced manager with multi-million dollar assets, that is all you will get. And is that enough? If you are familiar with DeFi and understand that these protocols are interconnected, then to understand how this position works, you can try to build a knowledge graph between objects and their financial connections. So, we see that our main character had a wallet. It holds nameless ERC-20 tokens representing a stake in the Morpho repository, distributed by a curator—in this case, MIV Capital, which supplies assets to certain credit markets.
In this particular example, one of the markets was secured by XUSD from Stream Finance, which was already mentioned on stage a few speeches ago. In fact, we see that the risks lie in the connections between objects directly in the blockchain. Unfortunately, as mentioned earlier, on the same day the deposit was made, Stream Finance announced a protocol crash and approximately half of the $200 million in TVL disappeared. This affected our depositor, the main character of our story, and he lost about $200,000 in 7 days. What happened was this: xUSD lost its peg xUSD lost its peg , the price fell, and the vault had to distribute the losses among all depositors.
We see that the user see that the user who made two deposits ended up losing about $200,000 in 7 days. The most interesting thing is that the user made a deposit after the Stream Finance collapse, not once, but twice. So what does this story show what does this story show ? It really shows that the data we get online is we get online is not really usable, and data from many projects and interfaces is still of low quality if you are managing large volumes or trying to professionally manage assets. So we think that the really necessary step to move from the current closed state of the industry to a place where finance— place where finance— that is, finance on the web—can become the software for all finance is to better understand and better manage data.
And it's quite surprising, at least to me, that 10 years later we're still trying to figure out how much data we're creating, and yet how little of it is truly meaningful and valuable. And specifically, if we're talking about what kind of data do we need, how do we bridge that gap from where we are today to where we want to be? What type of data do we need? And I would like to talk specifically about three types of data. So, the first one is data that helps us answer the question: what is it?
That is, data that specifically relates to the subject's taxonomy. Data that allows us to understand what this entity is. But we also need better data to understand how it works. How does this protocol work? How does this asset work?
This is data that should reflect the relationship between these entities. This is an ontology that we believe is one of the most interesting challenges in this field. And ultimately, we also need data to make decisions. We need to be able to combine this better data into products that allow us to more effectively interact with and use finance online . So, this is the third category.
Speaking of taxonomy and better data showing what a particular object is. I'll try to give a simple example. We have two storage facilities here We have two storage facilities here . One is the Steakhouse Prime USDC repository. Another is Steakhouse Prime USDC repository.
Another is Nimo Prime USDC storage. If you look at the raw blockchain data, these two repositories look almost identical. They both meet the same ERC standard. They both use USDC as a deposit asset. And we would call them both storages.
But if we dig deeper and try to understand, classify, and create a taxonomy of what each of these objects is, we see that in fact the nature of these two products is fundamentally different. From a financial perspective, from a risk perspective, and pretty much anything risk perspective, and pretty much anything else you need to understand when you're trying to invest and access a product. You will see product. You will see that the strategy used is different. So, Stakehouse is engaged in mortgage lending.
Another company essentially uses multiple strategies around options and basis trading. A manager has fundamentally different limits, he is restricted in a different way. The oracle and the method of calculating the price of an asset are completely different. And finally, liquidity. So, how can I access, how can I log out?
How long does it take for me to receive the assets? It's just completely different. So, again, we clearly see that the data on the web is not enough, and we need to dive deeper and call a spade a spade . And again , the problem is not that we can't describe it. The problem is that we have to do this work every time we talk about these things.
Every time an analyst, agent, or asset manager studies an asset, if we don't have a common taxonomy that everyone agrees on, the work has to be duplicated and repeated over and over again . And I think it's hard to overstate the importance of taxonomy, and it becomes obvious when we try to draw parallels with traditional finance, right? So, imagine an analyst in Hong Kong trying to understand a company and creating a report and then sending it to his partner here in Lisbon. They may value the company differently, but they don't need to explain to each other every time what an asset is and what a liability is. Not everything needs to be redone over and over again.
In traditional finance, there is a very strong and reliable taxonomy. Of course, traditional finance operates in a disparate manner, each running on different databases, and they don't have the sophisticated, unified level of infrastructure level of infrastructure that network finance has. However, as I mentioned, what we really lack in terms of taxonomy is that the network taxonomy is very weak and poorly understood. And sometimes even we ourselves, as crypto enthusiasts, waste time arguing about what a repository is, what an asset is, what the nature of a native L1 token is. And yes, I think ultimately it's not very conducive to effective collaboration, moving quickly, and getting things done.
Let's assume now that we have a good taxonomy. So, let's say we spend time studying on-chain studying on-chain contracts and assets and develop a good taxonomy that we can agree on. Is this enough? Is this enough to use assets on the network? Is this enough to take on-chain take on-chain finance to a global level?
Well global level? Well , not really. If we look at the example from earlier, our protagonist, we see that the real risk lay not in the asset itself, but ultimately in the relationship between the assets and another smart contract that was running that safe. So, as we see here, the attempt to reflect the principles of operation, that is, the relationships between these objects, is no less important. So, what would an ontology for on-chain finance look like?
We need to take the position of a market participant, for example, an asset manager. And we can take the safe from our previous example. So, I need to understand: let's say I know this is a safe, I know it corresponds to a certain type of financial product, and I need to understand what key relationships are important to me. Okay, if I'm going to deposit funds into a safe, I need to understand what control over it looks like? Who can operate this safe?
Who can change the way a safe works? Liquidity. How long will I have to wait to get my assets back? Input data for pricing. So, how is the cost of a safe calculated?
How is net asset value (NAV) determined? Transparency of commissions. What do I pay for access to the safe? Are there any hidden fees that I also have to pay? And finally pay?
And finally , what type of collateral do I have risks for? So if we start mapping these relationships, we get a graph, essentially a knowledge graph about how this particular object interacts with other facets that I'm really interested in. And this finally becomes useful for understanding and evaluating a specific investment. We talked about taxonomy. We also talked about ontology, meaning how can we classify things?
Ultimately, all of this is pretty pointless if we can't collect this data and make it usable for specific tasks specific tasks . If you look at existing products in the crypto space, I would say that we already have some pretty powerful projects that do a particularly good job of taxonomy of things. So, for example, I would say that CoinGecko, Token Terminal, RWA.xyz—they're putting a lot of effort into really understanding what's what. And you would be surprised how many complexities there are in this work.
So, for example, what is an asset? Is the moved version of an asset part of the same asset? Does the issuer have to be the same for different transferred assets for them to be called a single asset? All of these questions need answers, and the less agreement we have, the harder it is for us to cooperate. Speaking of ontology, I think there are projects that have really thought through how we can map the relationships between entities, between project protocols and assets.
It is also worth adding Dune, DefiLlama worth adding Dune, DefiLlama , as well as Debank and Zerion. But I find the project I would like to highlight very interesting— I would like to highlight very interesting— The Herd, The Echo. They are a truly agent- truly agent- oriented way of trying to understand how different contracts relate to each other. And that's, I would say other. And that's, I would say , the current state of the products that we have online from a data perspective.
But again, my point is that this is great and it's a good start, but it's still not enough to take us to the next level of on- next level of on- chain finance as a real foundation for all global finance. And I'm trying to argue that there's now a new generation of products that can be built on this data, that can span the entire knowledge graph. That is, they can include a good taxonomy, a good ontology, and make them suitable for a specific usage scenario. So, this is actually the business of our entire lives. This is what we have been doing for 2 years.
We tried to collect all this data, organize it in a unified way, and ultimately create an application that asset managers and their agents can use in one place to manage assets across the network. I think what we saw was that the lack of attempts to unify data and create an effective level of collaboration level of collaboration is not only important for people who have to agree on what and where to invest, but also that this level of reliable data is necessary for agents to be able to do their jobs. An agent cannot spend tokens every time it needs to interact needs to interact , conduct research and validation, and create an ontology and knowledge graph every time it works with a smart contract. This is impossible. So we tried to bring everything together into one unified application for on-chain for on-chain asset managers.
We currently work with about $2.4 billion in client assets, which we track through our protocol and our app. These are mainly treasuries, trading teams and liquid funds. We see that in a time when it is becoming increasingly easy to get raw data from the blockchain—you can just ask the Cloud to do it, and it will do it beautifully. We believe it is really important to create a logical layer that everyone, including your counterparties, agrees on, that works well across the stack, where transactions clearly map to protocols and protocols map to assets, and this strongly typed data layer will power the next generation of on-chain applications.
That's all, thank you very much. If you have any further questions about DeFi or similar things, I'd be happy to chat, please get in touch.
Automatic transcript — names and jargon may be misspelled.