Partnered With:

Interview with Uniswap founder: The rise of UniswapX and the future of decentralized markets.

Read this article in 25 Minutes

UniswapX addresses issues such as gas fees, validators, and MEV, while optimizing user experience and trading efficiency.

Interview: Bankless

Guest: Hayden Adams, Founder and CEO of Uniswap

Original article compiled by Qianwen, Chain Catcher

Let the market handle the routing problem

Host: Uniswap is cool because each token has its own exchange and pool, and this is true for V2, V3, and V4. I call this path the "Hooks-centric roadmap," which is consistent with Ethereum's Rollup-centric roadmap. Ethereum is doing this, trying to reduce complexity as much as possible and letting Rollup do all the execution work.

Uniswap's Hook is also doing something similar - in this highly visual (expressive) situation, empowering Uniswap's pool builders to build. The problem now is that despite the reduced complexity, the problem still exists, such as the existence of many different types of pools. So, can we say that UniswapX is a solution that utilizes all of this complexity?

Hayden Adams: We are discussing this issue in the context of a protocol today, which is highly decentralized. What we need to do is solve the routing problem, also known as AMM. Currently, exchange, routing, and liquidity provision are bundled together, and we try to find the best price in this process. But I think it can be slightly unbundled because we now have various strategies, different routing methods, and various aggregators.

So UniswapX is a competitive routing market that achieves higher efficiency through decentralization, allowing more people to participate in finding the best route and discovering the best liquidity pool. In the long run, users can get better prices.

Host: The coolest thing about Uniswap is that you can list any token. I think the routing problem - finding the best price for your trade among the liquidity sources in the token pool - is actually a calculation problem. My understanding is that with UniswapX, you want to leave the routing to the market instead of having the Uniswap router manage all this complexity, to ensure that traders and swappers get the best liquidity and the best quotes. Is that correct?

Hayden Adams: Yes, we are not only solving complexity issues, but also other problems such as gas fees, validators, MEV, etc. For example, by competing in auctions, it encourages people to seek the most complex strategies, provide better prices, and save gas costs.

Uniswap X uses off-chain signatures, where users do not directly send transactions to the chain, but instead sign off-chain messages expressing their intent. Then, people compete and once the price reaches a certain level, someone will be willing to submit it to the chain. These submitters (fillers) are similar to today's block builders, as well as trading companies and arbitrageurs. Other benefits of off-chain signatures include the fact that if your transaction fails, it will never be submitted. It also allows for gas fees to be extracted for users.

UniswapX is a competitive market

Hayden Adams: Traditional DEX aggregators integrate liquidity sources to cover as much liquidity as possible and manually find the best price. The solutions provided by these aggregators are singular. But rather than calling ourselves an aggregator, it's more accurate to say we are a market where everyone can compete and various solutions can compete with each other.

The current user experience is purely based on the gas price curve, not the asset price curve, to determine whether your order is included. But our solution is to combine them. By combining gas auctions (if included) and price options, we optimize your price. In contrast, the past situation was not as efficient, for example, your transaction would only be executed when the gas price dropped significantly, but during this period, the tokens you sold may also have dropped in value, so you need to pay more gas.

Off-chain orders promote user experience

Host: In the past, when I broadcasted a transaction, the gas fee I paid was hardcoded into that transaction and had no relation to the transaction I was actually conducting. These were two separate and completely unrelated variables. The result may not have been optimal, for example, when I tried to save two dollars on gas, the token price dropped.

Hayden Adams: The current off-chain order refers to signing a transaction off-chain without broadcasting it, waiting for the submitter (filler) and the receiver (taker) to execute it. They perform a comprehensive calculation between optimizing gas fees and optimizing actual exchanges, using this complexity to generate the best results. The complexity issue is left off-chain and assigned to service providers who can manage complexity off-chain.

Through this approach, we have solved the current suboptimal situation. For example, block builders are not currently providing users with the most money, but rather making the most money for themselves, and this money is often auctioned off to ETH holders through efficient MEV. We still hope to return these profits to users. Off-chain orders also make cross-chain exchanges possible. Starting on July 17th, US time, people can choose to join the beta version on the user interface, which is currently only available for a small number of tokens. But in the long run, we can truly achieve cross-chain exchanges, and you can sign orders for a better user experience.

AMM can be used as a way to create liquidity and build on-chain integration based on liquidity. However, in the long run, this solution is not entirely feasible because we need to consider factors such as MEV. We now have multiple routing solutions, such as client-side routing, as well as smarter routing that is faster and can discover more routes, but uses open-source APIs behind the scenes. I would consider the latter to be a protocol in this field, rather than a protocol like AMM. So, in fact, we already have these routing solutions, and we are updating them as a protocol. We also have on-chain smart contracts to settle orders, which are immutable and open-source (based on GPL). This is how it works: first, the front-end broadcasts the order, and then the submitter needs to discover these broadcast orders and submit them directly to the smart contract.

We now have exchanges and liquidity providers, while Uniswap allows fillers to take the stage. They already existed, we just give them more attention and restrictions. You can imagine them as people who use the most advanced trading strategies today, engage in large-scale arbitrage, block builders, and MEV extractors, and what we need to do is to impose restrictions on them, let them compete with each other, and return value to the exchanges.

Host: Previously, people who traded on Uniswap will now be directed to UniswapX, so in the future, trading will be done on UniswapX, right?

Hayden Adams: More accurately, over time, there may be more trading volume flowing to UniswapX, and users can choose to use UniswapX. But if there are better pricing methods and better resources, people will definitely choose them.

Uniswap is committed to building a better ecosystem

Hayden Adams: The significance of Uniswap lies in driving the development of the Uniswap ecosystem, enabling it to surpass the strict definition of AMM. However, I would like to say that the AMM routing method is still a good way to go.

Returning to the issue of Hooks, a concern that many people may have is that they have created a Hook, but people may worry about its security and so on. So how can you ensure that people will discover your liquidity?

The solution is that if someone creates a new custom pool and a new custom Hook, they do not need the Uniswap Labs team to review their Hook and integrate it into our frontend. There are simply too many Hooks for us to do that. Instead, what they need to do is find a filler who is willing to integrate their Hook. Once they find one, the liquidity they represent will immediately return to the Uniswap user interface.

UniswapX How Does It Work?

Hayden Adams: Starting from the front-end, the user experience for the front-end exchange remains basically the same: select the token and make the exchange. The first biggest difference is that you will find that your exchange does not require gas. However, I want to point out that each token still has an initial approved permission contract, based on which you can make a signature. Every once in a while, you need to approve a new token that you have not traded before, which requires gas and cannot be avoided unless the token itself includes a gasless signature or you use a smart contract wallet.

Then, it's time to make the exchange without paying gas fees. You sign the signature and the order enters processing. The difference is that when you sign a transaction, you can immediately enter Etherscan and start waiting. This is just an off-chain signature, so the order has not been executed yet. It is not even in the memory pool, but in the pre-memory pool.

Starting from the front end, the order will be broadcasted to the submitter's network. Basically, this process is like a Dutch auction, where you set the starting price higher than your expected price and then gradually lower it. Once someone is interested, it creates competition, which theoretically produces the best results.

If the submitter submits a transaction, it means they will pay gas, and then they will withdraw your tokens from your wallet. Because you have already approved this process before, and there is a smart contract that enforces the rule that they can only withdraw tokens from your wallet if they send the specified tokens. So it's like a timestamp component.

Some people may be concerned about transaction delays, but personally I think auctions must be very fast. So we offer optional components that allow you to add an RFQ. This basically means that you use an RFQ pricing system, so when you parameterize the auction, you don't have to guess, but ask others for the best price they can give you. If you want to motivate people to give you the best offer, you must give them some rewards. For example, in a smart contract, if you use optional RFQ parameterization, the RFQ winner will have a slight priority in the first few blocks.

For example, if I signed an order, and someone can provide a better price, they can fill in their quote, so there is still an element of public auction. This way, you can get very efficient and fast price discovery. If using RFQ, it may only take one or two blocks, otherwise it may be five to ten blocks. This component is optional, and if the user feels that they can wait for a minute or two, then they can choose not to use it, which is very flexible. Just saying, if you are willing to wait, then you may need to bear some price risks, such as significant fluctuations during this period.

Uniswap 如何改变 MEV 攻击？

Uniswap How Changes MEV Attacks?

Hayden Adams: What we want to do is to allow value to flow through the MEV market to the block builders and validators. When someone makes a transaction, it is actually publicly broadcasted. Just like now. When a user makes a transaction on Uniswap, people compete with each other, but their competition is not to provide the best price for the user, but to extract the maximum value from the user's exchange on Uniswap. This is the current MEV market, this is the current block building network.

Can we change the way we encode transactions? For example, instead of having a group of smart people competing to extract value from the exchange, why not have them compete with each other to extract as much value as possible? Some value will always go to Ethereum miners, validators, and block builders, but how can we maximize the percentage of MEV that returns to the exchange?

They compete with each other and take away all the money. Now, we hope they continue to compete with each other and return most of the value to the traders. So, it's about how you code the transaction, and auction is one of the optional ways.

The current situation is that as long as I can get the lowest slippage tolerance, I can trade. Then the validator can trade first, reach the lowest slippage tolerance, and take away the money. Through auction, the price decays over time. During this process, as long as someone thinks that including this transaction is profitable, they will submit an auction. Before the tolerance bottom out, the order has been executed and profitable. This way, you won't be able to front-run trades on the trading target like before. This system ensures that someone completes the order when the first profitable opportunity arises, which is itself a MEV protection mechanism.

For example, if you have a bunch of transactions off-chain, a submitter can discover all of them and complete them simultaneously, meaning they can submit orders early in the entire cycle. In a price auction, the earlier you auction, the higher the price and the less value leakage.

Cross-chain Transactions

An order is represented as an off-chain signature, which means you don't need to sign the order specifying that you want to exchange A tokens for B tokens. Instead, you sign the order specifying that you want to exchange A tokens on Ethereum for B tokens on Optimistic. From the perspective of the trader, the user experience is exactly the same, and in fact, the user experience can feel very fast. The submitter will handle the complexity and delay of using cross-chain bridges and figure out settlement, among other things.

Other benefits include the ability to enable native asset exchange. When people move funds between chains and conduct cross-chain transactions, it is usually done by bridging assets. Then, you have bridged tokens of that asset on another chain, and the funds passively exist in the cross-chain bridge. We have also seen many cross-chain bridge hacking incidents.

When using UniswapX, users first indicate their exchange intentions and then participate in a Dutch auction. Once a submitter has a willing price, they initiate the transaction on the input chain. They transfer the user's input to a custody contract and then send these tokens to the output chain. They must prove that they have performed this operation before requesting the user to input tokens on the input chain. This proof can be optimistic, which means that they assume they have done so without proof, making cross-chain exchanges very fast. Alternatively, if someone questions whether the submitter has actually transferred the tokens, there will be a seven-day asset lock-in period during which the submitter must provide proof.

Chain-signed orders not only solve the complexity problem of pools, but also the complexity problem of bridging. The complexity is solved by the same service provider and the same submitter. They can manage time, delay, and complexity. All of this seems to be determined by the free market and the submitter, and the process is also supervised by smart contracts to ensure that the submitter brings what the exchanger needs.

I believe that in the future, most assets will exist on their origin chain, or on their safest chain, or on their most typical asset chain, rather than bridged. That is, if a submitter performs a cross-chain exchange, they will receive tokens on the native chain of the token. As a result, the use of bridging seems to be really minimized in this model, and rather than saying that cross-chain bridges are bridges for assets, it is more accurate to say that in this model, cross-chain bridges are only used to transmit the final information. You don't even need that data packet unless the submitter is lying.

This can be called the minimum viable bridging, and users will only bear bridging risks when trading across bridges. Once the exchanger receives the output token and the submitter receives the input token, neither party will have bridging risks again, unless these tokens happen to be packaged bridging assets, but they don't need to be.

Therefore, we have minimized the degree to which people need to bridge, while also abstracting them. For example, this system can support any possible bridging. You can think of it as a bridge aggregator, where submitters can use any bridge, but each transaction has a specific cross-chain bridge, which is the "settlement oracle". It can be any bridging device or any other system, and you can also use multi-signature systems, governance systems, or unilateral systems, and you can trust the submitter.

Utilizing Complexity

Hayden Adams: I believe that this should be a critical moment, indicating the forefront development of the Uniswap ecosystem. People expect to use decentralized markets, but there are still areas that need improvement, such as inefficient processes and insufficient liquidity, which affect user experience.

In the long run, decentralized markets will actually be stronger and more powerful, with better user experiences and more efficient market structures. Using a market-based approach to handle certain things, rather than operating like a single company.

With this philosophy in mind, and utilizing the new skills we have learned in blockchain and cryptocurrency, I believe we can now begin the first chapter of our story. We aim to establish the ultimate centralized order flow network and hope to collaborate with other teams to address these issues, such as those who are also researching decentralized order flow. Overall, I believe this is a very important moment for the entire industry.

From the perspective of trading volume and user numbers, more than 30% of all gas on Ethereum may come from Uniswap transactions, and I can guarantee that this proportion will be even higher when considering MEV. I believe that our work in this field is very important, as it can help drive the development of the industry. Additionally, we have users who we can learn from, optimize, and improve the experience for those who are truly engaging in transactions.

On July 17th, Uniswapx launched its mainnet test version, which is only an early version and is only open to some tokens and transactions. During the trial, you will see new features such as no gas fees and signature instead of transactions. This is a gradual release process. There are still many issues to consider, such as parameterization, and it will take some time to start integrating cross-chain, etc.