Beyond Competition: What Solvers Do and What We See | Pangting Huang, Portus | ETHTaipei 2025

[Music] Hello. Uh, thanks Ryan for the introduction and thanks everyone for coming to the talk. Uh, I'll be talking about what we have done in the past year and I'm Hanking. You can feel free to call me Ben if you find it hard to pronounce that's fine. and I'm the CTO of Dingo and we are doing some researches uh about being a solver and how to run in a intent ecosystem.

So today's talk will be mainly about what our everyday attacks are and what we see in the future for intent and this is our PO app. You can scan and claim the POA app. Um yeah and we have a series of uh activities and we have a booth here. So make sure to reach out to us if you are interested in everything. Okay.

So okay so we are Porters and um actually we are our company name is Tingo and we start up from doing uh from building an exchange and then we uh start to build defies like through combo protocol link start from uh 2018 but uh in the whole process we found out the potential of uh what can uh intense solar system do in the future. So uh we started to run our server on cowop in production since uh January 2024 and we already settled over $8.9 billion executed over $2.8 million deck swaps and we are one of the first uh silver to support all the four chains on cow swap and now we have the highest settle batches in the past 30 days. It's continually grow growing but yeah we are the mo one of the most active player in the um solar system and besides the intent swap system we also um done some experiment with uh intent bridge system um but if you are interested in that you can come to the booth and we can talk about it more.

So what does silver exactly solves in this ecosystem or what do we see as we have we think that it has more potential in the future. Um I think everyone is familiar with this name which is uh liquidity fragmentation cuz when we have different kinds of services, different kinds of assets, it makes tokens is not just tokens anymore. Um all kinds of liquidity have different kinds of way to match them. that. So just like we start from building the exchange and continuing to defies uh the blockchain trading history is also changing.

Just like when we have uh Ethereum we have programmable uh contracts. Then we have different kinds of decentralized exchange which makes uh all kinds of assets start to grow and by far we have different kinds of uh liquidities to be integrated and then we have aggregators to handle all kind of complex uh executions but we can do more uh about the liquidity management models some things are we we are familiar to is like order book like AMMs I believe everybody knows it And another thing is like RFQS which is stands for uh requests for quotes services. Uh they are quote driven and often run by some um professional market makers and they provide a very competitive price but they are more restricted to be leveraged. So um we can say that normal users cannot be like assets sold directly. And what about other assets?

They can be stored in protocol. it can be stored in lido. So um there are different ways to leverage these uh values when we need to um solve other problems. So about the asset model suitability every assets have different uh volatility and debt. So they have different models to to use that and also some major assets like uh ETH itself or like WBDC or some stable coins they they are also um frequently being used by uh as a counterparty to most assets who also separate their own values in different pools of dexes.

And the last one specific use cases just like um when we are um I think the most frequent case is like LP tokens because when we store our tokens into a liquid pool and then we can get LP tokens those LP tokens do have their own value but they can be accessed in a different way. So that's something that we can say that if we have a better way to leverage them then we can partially solve the problem of uh liquidity fragmentation and the last one so many chains so many L2s um I think the L2s will just grow more and more so that that's something that every value can be placed in different places and not being solved naturally. Yeah. So all these fragmentation really blocks uh how the service can be um can be developed and how users know how to use them and how to uh how they can manage their portfolio. And here comes the intentbased system.

Aggregators do solve some problems to um simplify the whole trading process but they are centralized. And on the other hand intentbased system user just need to express their goals. They don't have to do the execution we will do for them and we are trying to be decentralized. Yeah that will be done someday but we are still working on right now. So what exactly do solvers do for you or what are we good at?

So the first part is obvious that we have good quoting powers and it comes from two different aspects. One is the diversity of liquidity integration. We need to support all kinds of liquidities and for all kinds of assets. Uh why that's obvious because everything needs to be in our toolbox to give better solution and these kinds of liquidities doesn't means that only dexes though they have different um interfaces they have different ways to use it and we also leverage some RFQS and also some you um like protocols that involves different kind of functions and we support over 80 protocols across four chains right now and another thing is the liquidity utilization ability which means the routing we can optimize the routing based on the orders we are processing. We can separate the orders into several orders to find the best way for each one and we can decide whether to execute the um this strategy or not based on the risk.

And on the other hand, we can also combine orders because for some RFQS they they require to have a minimum amount to do uh to give a better price. So in in that case we can combine the orders to uh build up the the execution and also we can have a better um performance on how do we handle the gas gas efficiency and also since we do the execution instead of the users. So we can continuously optimize until the last second till we actually settle the transaction on chain. So that's something we can do but um others might not and since we do the execution instead of user so user don't need to have their uh capital at first place let's say um as long as they when we are executing the order then uh they have the money on them then this uh execution can be done and another thing is that they don't need to handle with the gas fees we will handle it by them uh for And the last one uh I won't say is the most important one but is more important than we have imagined at first place is the risk management because the market always moves right. So if you have experience on doing trades you know that when you get a price it doesn't mean that you will get the actual amount at the uh at the end.

So this is something that uh solvers needs to work on to mitigate the risk from uh market fluctuations including the uh asset price and also the um gas price at the same time. So let me trade for you. Okay, let's get a very simple example like how to buy Pendle just a just example. uh when user want to buy uh a certain token, they oftenly can you can do some research and find the best price on the market or you can go to go straight to the most familiar decks you you use and then you take out your ETH and go to DEX and buy Pendle. That's a very normal uh process.

But how do we do that? If you use uh intense system, you choose your own asset that you want to use and then we will do the price discovery for you. And when we found the price, we can say that uh when you want to buy by Pendle, you can swap your USDC to ETH and then ETH to Pendle. That's something you can do, but we will do it for you because we have lots of different kinds of liquidities to use like uh you can swap uh your uh USDC to ETH uh on BBop which is a close partner of us also and use this uh RFQ to perform the swap and then swap your ETH to Pendle on balancer and this ends up with a better price than you just go straight to some dexes. So this is an actual uh settlement happens on Ethereum which is handled by us.

It's a buying order and but this case is really simple because this there's only one order, right? So when it comes to very uh a large amount of orders to be processed at the same time, the whole routing can be very complex and this is something that we do every day and hope that makes you think that we are better on this. Okay. And the next one uh after the quoting power then we comes to the risk part. Um cuz I believe that if you use a DEX you know what is slippage it means that when you get the price it doesn't means that you get the amount at the final execution and those mark uh those um slippage is caused by the market movements and it may also cause some failed transaction if it if the the amount exceeds your uh the threshold you have settled.

So, so these are the points that solver cares more than aggregators because we are the one to do the execution. So, when we um propose a quote based on the order then we guarantee to process the whole amount uh when when the settle is when the transaction is actually settled on uh on the blockchain. So silverware take carees of the risk um so takes care of the risk and uh cover users loss that caused by the market movements and another thing is that as since we do the execution we have the ability to um like interact with some MV services so no more transaction failures will be caused by the slippage on the users perspective so um if the transaction cannot be processed then users order will just return back to the order pool and then wait for the next round. It provides better user experience and another thing more important is it provides a more better system robustness is if the swapping system is integrated with uh by a product then we can guarantee that your like maybe a content system can actually process successfully instead of just reverting a transaction. Okay.

So how is the risk model if it is so important? Um so we try to like simplify the question uh since we have the slippage and also the failing transaction need to be handled and we found out that slippage is really highly correlated to reverse although they have different reasons sometimes but it's a strong positive. So we can try to solve the reverse transactions and try to monitor that uh instead of doing everything and one of reasons obvious would be the price change who caused the um slippage and another thing is the quote expiration and the quote expiration is mainly being caused to um some RFQ interaction to be failed cuz uh RFQ often have their own restrictions that you need to settle your quotes within a certain of time. Another reason are some smelly tokens. That's something we need to deal with case by case.

It's annoying, but there's nothing we can do. Just do it ourselves. Yeah. So, um after we know the the reasons, so we are trying to um decide some factors that can be modeled to solve the issue. And one of it is that we I think we can agree that when the network is busier, it makes more possible to the price change.

So the price gas price is something that we can monitor. Uh higher gas price means a busier network. And another thing is that with more interactions when we are trying to do a better routing it might have more interaction to different pools but it also increase the risk that longer path and more hops will cause uh a higher risk of failure. Another thing is the order amount. If we process a lot of orders at the same time, maybe we can have uh better performance on gas efficiency, but you also increase the risk and block settlement latency that will be something that I I just mentioned about um the RFQ restrictions and also if we settle in a longer longer period of time then the market market state might be different from the beginning.

And the last one if we cannot think of any more that can that will be the root cause we can um we can consider on some results like revert monitor we can monitor other how our solvers doing right now. If a lot of store is still is already reverting their transaction that then there might be some systemmetric problems right now and token based reverse tokens base token open reverse in in a bunch uh together. So um this will be something that we can consider if we need to do uh aggressive quoting or just preser uh more preserved quoting at the moment. And after um being able to model these factors into our strategy, we can have a better rever comparing to the market right now. So let's over take care of these of for you.

You don't have to concern about uh what kind of risk need to be managed. We can do all of these for you including the asset selection, price discovery, the market fluctuation and gas fees and also the last part. Remember I say there are so many L2s, so many chains all these can be quite different on different uh chains. So these settlements can be um decided by us. We'll handle it for you.

Okay. So what will be the next step after we have a I hope it's robust system for um the every solvers to do the trace for users we can start to think about how can we um start off to give a unified experience across different chains which means the crosschain swaps and this is uh a phrase that I like a much uh I I like very much from the open intense framework which is intense for everyone and with everyone which means that everyone should be able to uh express their intents of what they want to do and then other people that are being able to solve this problem which is solver know how to uh understand understand the the thing that user wants. So it comes up with standardization uh and something else I will mention later and for the intent standardization the intent stack is basically about the intent flow which is how orders are being created and handled and other thing is the settlement mechanism is the order being processed into solutions and actually happens on chain. So ERC 7620 uh 83 uh defines the interface about intent creation, execution and settlements which is a very good start and already being used in some um intense bridges protocols. Another thing is the open intense framework pushed this even further.

It proposed a moduleized software implementation and also the settlement smart contract which uh which are the most two of the most important part to settle all these things onto a um intent ecosystem. And besides the intent standardization, there's another thing about liquidity because uh solver we always need to deal with all kinds of liquidity every day and this will be a really redundant work between every solvers including like every liquidity have their own strategies not strategy design on indexing encoding and how to simulate every transaction if they are if they will be successful. uccessfully process and what will be the state after processing this request and there's another project named tao there are uh and liquidity indexer if you are interesting in these kinds of questions about liquidity standardization be be sure to you can take a look at this project it's quite neat and yeah so after the standardization remember I just said let's talk about later. It will be the collaboration because um with everyone for everyone means that we need to decentralize everything and standardization is a very important part. But the next one is how do we actually have these much different roles different solvers to do these works.

So we need to collaborate in instead of just competing with each other. Um the competing um the sover competition is quite straightforward to select the solution and the consensus on intense system is basically about the solution selections cuz we have all kinds of orders and we need to choose a answer to it. But the current selection method is mostly like one server takes all. So every solver needs to be the toughest one instead of you can still give some solutions partial solutions and then being combined together and those competition is often driven by surplus maximization and first comes first serve. So um in this case the challenge of current model will be that everyone need to like collect all the answers together instead of colle uh cooperating with other people and this cause it lack of differentiation and sometimes some solvers they have their own liquidity source that cannot be accessed by others.

Then in this case if those uh if some some of the order cannot be processed by these solvers then these solvers will uh these orders will be put to uh put aside and cause their execution to be delayed. This is also a problem that we need to solve in some cases and combination of these two we can see that scalpity issues comes up comes up that solvers need to uh propose a very powerful single strategy instead of cooperating. So it's something we need to deal with in the in the future. Oh, so I believe that everyone sitting here believes in decentralized just like Dexus decentralized where the liquid lives. Solver will be the one to decentralize how the prices are formed which are also credential uh critical to uh for the value where the values come from and centralization initiates the scalability improvements which makes the decentralization even possible.

And also there's another case that maybe we can just like uh when I want to trade I can go to aggregators and do the trading but I can also go to dexes instead of go to aggregators and this standardization makes it possible to create a server tot service which means that services or users can go straight to the solver if they want and this is something we're also working on And if you are interested, be sure to come to the booth and let's talk. And the last one um it will be uh um something that we need to think about is a new model of solution selections instead of just one solution each batch. right now. Uh the final goal is to incentivize the solvers uh specialization since that we we need to have different diversity of solvers solving different problems. If we continuing expand our uh scenarios from a single chain to multiple chains to cross crosschains.

So after we done everything above we are finally being able to solve the fragmentation of liquidity then it will be a important layer to support every protocol as the value uh value layer. Okay, thank you for coming and hope everyone enjoyed the talk, enjoy the event and we are continuing to explore the future of crypto and wish everyone have a good time and the last one we're hiring. So, uh if you are interested in a topic I just mentioned or you already have some thoughts, be sure to come to our booth. We have a series of activities to make you familiar with this um how everything works and yeah that's that'll be it. Thank [Applause] you.

Okay. Thank you bounting. Um do we have any questions from the audience? Okay. Okay.

Hi Ben, just curious uh if you if you solve intents from Daxis, how do you guys prevent from me? This is my first question. And the second one is do you guys do hedge until your liquidity being being settled on the settlement network? Yeah, good question. Um for the first one, how do we prevent from MEV?

There's I assume there are two parts can mitigate the effect of ME. One is that we use ME service to settle our transactions in on Ethereum. And when we do these uh transaction, we do them in batches. So if you if anyone wants to do the ME um like attacks on our uh solutions, they need to deal with multiple orders instead of just one. So that will be make make it more harder.

But the main solution I think uh at the moment will be using an MV service on Ethereum. Yeah. And the second one, do we do hash? We don't do hedge right now but because that's something that we need to preserve the uh the value back to the user. So we're not doing currently.

Okay. Um thank you Bing for bringing such an insightful presentation to us. Thank you. For the next talk we have [Music]