Let's talk about the iconic DeFi protocol Uniswap and its unique long-term value

Source: Shifting

Shifting is an in-depth podcast focusing on cryptocurrency, blockchain, and Web3, aiming to provide valuable perspectives and information to listeners while welcoming feedback and communication from the audience and community. You can currently listen to this podcast on Xiaoyuzhou App, Apple Podcast, and Spotify.

Below is the transcript of this episode, slightly modified for readability compared to the audio content. Today's guest: Chen Jia.

Xuanhao: Hello everyone, welcome to Shifting. I am Wang Xuanhao. This podcast revolves around cryptocurrency and Web3, hoping to provide more perspectives and information for everyone, while also welcoming feedback from listeners and the community. The content of the podcast and the guests' sharing represent personal views and do not constitute any financial investment advice.

Today, I plan to talk about Uniswap (which may also be referred to as Uni). The reason is that if we look back at the explosion of the crypto world since 2020, DeFi should be a landmark event. And if we review DeFi, Uniswap should be a landmark protocol. From the user's perspective, Uni should be the entry point to the entire DeFi world, taking on the most important trading functions, and this product itself consists of only a few hundred lines of code.

The Uni team, according to public information, consists of only about 20 to 30 people, yet at its peak, UNI's market capitalization exceeded 30 billion USD. So if we average it out, the productivity per person should be many times that of traditional companies. In comparison, Nasdaq currently has a market capitalization of nearly 30 billion USD, but they have over 4,800 employees. So it seems that crypto is indeed a very different new field.

There are also some interesting stories about the founder. For example, founder Hayden Adams was originally a mechanical engineer at Siemens, but after being laid off, he self-taught Solidity and created Uniswap, this star protocol. Although Uni has passed the early period of heated discussion today, as a landmark protocol, it has its unique long-term value, which I believe can still bring us a lot of inspiration. So today we still plan to talk about Uniswap, and the friend chatting with us today is Chen Jia. Please introduce yourself to everyone, Chen Jia.

Chen Jia: Hello everyone, my name is Chen Jia, and I am a developer with many years of blockchain development experience, involved in various supply chains and protocol development. I am very happy to share about Uniswap with everyone.

Xuanhao: I previously listened to an interview between Musk and Lex Fridman, where they discussed what money and currency are. Musk's response was roughly that money is actually digital information stored in large machines using batch processing, and the Federal Reserve can modify these numbers. I think this answer is very much in line with Musk's style, hitting the essence or first principle. So from a similar perspective, what do you think Uniswap is?

Chen Jia: Uniswap is a decentralized exchange based on smart contracts. Its operation is entirely based on immutable code, and transactions between users are executed automatically by smart contracts based on the trading requests submitted by users.

The transaction price on Uniswap is determined by the ratio of the assets held by the smart contract. Each time a user conducts a transaction, it affects the quantity of the assets held, thereby influencing the price. The asset prices within the Uniswap system will remain synchronized with the asset prices on other centralized exchanges because arbitrageurs will engage in arbitrage. At the same time, Uniswap can also list new tokens on its own, solving the problem of difficulty in listing tokens on traditional centralized exchanges.

Xuanhao: Since not all listeners may be familiar with Uniswap, could you briefly introduce some of Uniswap's other functions?

Chen Jia: It has several core functions: the first is the trading function. We usually refer to a trading pair as a liquidity pool, and each pool generally consists of two assets, which can be exchanged within that asset pool. Uniswap also enables asset swaps between multiple liquidity pools. For example, if there are assets A and B, both of which have trading pools with USDT, but there is no trading pool for A and B. Then Uniswap can facilitate A to be exchanged for USDT, and USDT can then be exchanged for B, achieving the swap from A to B in a single transaction.

Uniswap also has another function, which is that users can add liquidity pools themselves, known as creating trading pairs in traditional exchanges.

Users can list their assets A and C on Uniswap, so I can add a certain amount of asset A and a certain amount of asset C, and the ratio of A to C will be the current price. Every year, many new projects start trading on Uniswap first before going to centralized exchanges. Currently, there are nearly 70,000 trading pairs trading on Uniswap, and we can see another data point: in terms of gas consumption on Ethereum, Uniswap generally ranks among the top three. About 5% to 10% of gas consumption is used on Uniswap.

Additionally, Uniswap has several other functions, one of which is flash loans. Users can borrow a certain amount of funds from Uniswap within a single transaction in one block to engage in arbitrage activities and then return the same amount of funds to Uniswap, solving the problem of arbitrageurs potentially lacking capital. Another function is the time-weighted oracle. Many DeFi protocols require an oracle to price assets.

Uniswap V2's time-weighted oracle solves the problem of price manipulation that previous oracles faced by introducing time weighting, allowing for an average price over a weighted time period, such as one hour or one day, to be used for asset pricing in third-party protocols. Another aspect is some secondary applications based on Uniswap, such as liquidity mining. Many friends may have heard of it, which is based on the liquidity pools of Uniswap's LPs for staking and mining. These are basically the main functions.

Xuanhao: So the main functions are trading, liquidity provision, flash loans, free token listing, and automatic routing of trades, along with oracle functions provided to other protocols. I think features like freely creating trading pairs are quite innovative.

If we say that Bitcoin has achieved the inviolability of private property from a technical perspective, then applying this sentence structure, Uniswap should achieve the inviolability of the right to free trade. No matter where you come from, your age, gender, wealth, or political affiliation, everyone is an equal trading user.

Moreover, features like flash loans are quite interesting (although Aave was the first to implement this feature). The interesting part is that the flash loan feature does not simply replicate products or logic from traditional finance on the blockchain; it has completely emerged on the blockchain, and such a thing does not exist in the traditional world.

In addition to AMM being a major innovation of Uniswap, on-chain protocol governance is also very interesting. For example, Uniswap's DAO is currently the largest DAO in terms of management scale, with an asset management scale of about 2.4 billion USD. I have also seen some governance-related news, such as the community voting proposal for the DeFi education fund. Can you talk about how this mechanism operates?

Chen Jia: Any UNI holder can submit proposals to introduce new features, which then go through a certain process to gain community approval before execution. The process is divided into several stages:

The first stage is a temperature check, requiring 25,000 UNI in favor to move to the second step.

The second step is a consensus check, requiring 50,000 votes in favor, and the final stage is the governance proposal. After passing the first two steps, this stage is more challenging, requiring 40 million $UNI votes to pass. Uniswap sets this mechanism mainly to consider that if a small number of people submit a large number of proposals, such as 10,000, the entire community's governance would become unmanageable, so several stages are set up. Each stage requires a certain threshold, and after initial discussions are approved, it moves to the second stage, then the third stage, and finally, the topics discussed by everyone are valid topics.

Xuanhao: So the basic model is divided into multiple stages, and the number of $UNI or staked tokens required for each stage increases. Voting is conducted through tools like Snapshot, and once the results are out, they may be executed by key community members or team members.

Chen Jia: Because most proposals cannot be executed automatically, they require some team members to implement them through multi-signature operations, or some may require developers from the team to deploy on other chains or modify contract parameters.

Xuanhao: In fact, protocols like this DAO have varying degrees of automation or decentralization. For example, the entire voting process is actually completed off-chain, and only after these off-chain voting tools produce a result do the actual multi-signature controllers or key community members execute the specific fund transfers or on-chain contract deployments.

Chen Jia: Yes, because from a technical perspective, many operations cannot be automated and can only be executed through manual means.

Xuanhao: We see that Uniswap operates partly on permissionless contract code on the blockchain, and another part is the front end.

Today, Uniswap may seem like a very mature application that everyone takes for granted, with around 7 to 8 billion USD locked, making it a well-known DeFi protocol.

However, I think that in the development history of a product or protocol, people's views on it may change over time. So if we look back at history, for example, in May or June of 2020, how did people view the development of the Uniswap protocol?

I have also seen some comments from Coinbase's former CTO Balaji, who was somewhat pessimistic about Uni in May 2020, and Multicoin and Dragonfly also initially did not invest in Uniswap, believing there might be some issues with the mechanism design. So I think at that time, not everyone believed that Uniswap would grow into a very prosperous or star protocol. As someone who experienced that period as a participant or practitioner in DeFi, how do you view the development process of Uniswap?

Chen Jia: The development process of Uniswap needs to mention the decentralized exchanges that existed before it, such as the Bancor protocol. When Bancor first came out, it was actually well recognized in the industry because the blockchain has always pursued the goal of decentralization, but there has always been a problem troubling the entire blockchain: the entire trading process is very centralized.

The creation of the Bancor protocol introduced decentralized trading to the entire blockchain, but due to some design issues with its model, it was not very usable. Later, when Uniswap came out, people compared Uniswap with Bancor and found that although Uniswap's trading model is very simple, it is very practical and effective in price tracking.

At that time, in the developer community, Uniswap was still viewed positively, but there were some inherent flaws in Uniswap, such as slow transaction speeds and high miner fees, which could not be resolved in the foreseeable future compared to traditional centralized exchanges. However, it can only be said that because of its decentralized characteristics, it inevitably brings some shortcomings.

So for Coinbase, which operates a centralized exchange, a transaction can be completed in basically one second. For Uniswap, which may require 15 seconds or even a few minutes for a transaction experience, they naturally see a significant flaw in the trading experience. However, for many blockchain practitioners and users, what they pursue is the entire decentralized process.

Thus, Uniswap's mechanism allows for completely decentralized trading on-chain, which is a significant advancement, and in this regard, the wait time is acceptable.

Xuanhao: I think, for example, as you just mentioned, there are issues of cost and speed, and there may still be some gaps in trading experience compared to centralized exchanges, and this gap may be determined by the blockchain itself, so there may not be any fundamentally good solutions in the short term.

What do you think the future trading volume distribution will look like between Uniswap and Binance (of course, most of the trading volume is still monopolized by centralized exchanges)? Will there be a large amount of trading volume going to centralized exchanges, or will both coexist in a long-term state?

Chen Jia: For major assets, there will definitely be more trading volume on centralized exchanges because the user experience on centralized exchanges is indeed better. However, for more long-tail assets and assets involving smart contracts, they will trade on decentralized exchanges like Uniswap.

Another point is that the usage threshold for Uniswap is relatively high for ordinary users; they need to understand the complete process of creating a wallet, miner fees, and wallet authorization, while it is much simpler on centralized exchanges. In the foreseeable future, centralized exchanges will still be the mainstream of the entire market, but the proportion of decentralized exchanges will continue to increase.

However, due to some decentralized reasons, there will definitely be a certain gap in experience, which will also deter some new players from using it. But with the development of technology, people's understanding of new technologies will deepen, and the proportion of decentralized exchanges will continue to rise.

Xuanhao: Because Layer 2 may have significant development in the future, it may also improve the issues of cost and speed on Ethereum Layer 1. For example, if protocols like Uniswap are deployed on Layer 2, will the issues of cost and speed see a significant improvement?

Chen Jia: There will definitely be improvements, but it may involve some issues. For example, its native assets can be used on Layer 2, but many non-native assets may not be able to cross over, which is a problem. Because many of the assets in Uniswap are non-native assets and long-tail assets.

Xuanhao: Are you saying that if these non-native assets need to migrate to Layer 2, the resistance may be quite large?

Chen Jia: Yes, because both Layer 1 and Layer 2 need liquidity pools, and if many assets' liquidity pools do not have high liquidity depth, then building two separate liquidity pools will incur relatively high costs.

Xuanhao: So there may still be some migration resistance, and it may require a process. Additionally, I think the competitive landscape and competitive factors of on-chain protocols are also quite interesting. For example, in traditional business, there is a concept of a moat, asking what the key competitive advantage of a product company is.

I think this question is also very interesting for on-chain protocols. First, on-chain protocols are mostly open-source, making it difficult to have any proprietary secret formulas or patents. Then, the switching cost for users seems to be low. In the Web 2 world, platforms like WeChat have very high switching costs.

Finally, on-chain protocols do not seem to have the same accumulation of data and content as WeChat; at least currently, platforms like Douyin and YouTube can accumulate a lot of content, making it difficult for other competitors to catch up. So what do you think the competitive advantages of on-chain protocols are built upon? You can talk about Uniswap specifically, or share any other thoughts.

Chen Jia: There are several core competitive advantages. First, the first point is brand recognition; its recognition in the global blockchain industry is quite high, and many new users and new funds will definitely prioritize Uniswap.

Second, Uniswap has a first-mover advantage and traffic advantage; its liquidity pool depth is currently the deepest. The deeper the liquidity pool, the better the overall trading experience, as slippage is lower. You can buy more assets for the same price, and more assets can sell for more money. This advantage means that the experience is much better than other competitors.

At the same time, many related protocols also rely on Uniswap, such as its time-weighted price parameters or liquidation processes that use Uniswap for liquidation. It’s a case of the strong getting stronger, so Uniswap, as a leading protocol, will continuously accumulate traffic, funds, and depth, forming a strong moat.

Xuanhao: There are actually quite a few DEXs that adopt AMM like Uniswap, such as Sushi and Curve. They may have made some innovations targeting specific application scenarios or short-term incentives. Of course, there are also order book-based DEXs like IDEX and 0x, which currently perform mediocrely. What do you think is the reason for this?

Chen Jia: The main reason is that decentralized order book exchanges have a significant gap in experience compared to centralized order book exchanges. At the same time, they do not have the conveniences of liquidity pools that decentralized exchanges offer. More people choose Uniswap-type exchanges on-chain and centralized order books off-chain, as both are relatively optimal choices in their respective domains.

Xuanhao: Since all on-chain protocols are open-source, if Uniswap makes some version updates or technical innovations, when other protocols copy or fork it, different DEXs may have different application scenarios and business contexts. Do you think this direct forking process will lead to other issues?

Chen Jia: There are many projects that fork Uniswap, and the vast majority will copy it verbatim, with a small number adding their own incentive mechanisms, such as trading mining, liquidity mining, or fee redistribution. From its core framework, 99% of them are the same as Uniswap.

These competing products will certainly affect Uniswap's traffic, but they mainly operate on non-Ethereum networks. On Ethereum, the main competitors are Sushi or similar smaller products. They will definitely affect Uniswap's traffic, but they also promote Uniswap as the standard.

Now we can see that basically any protocol involving decentralized exchanges is fully compatible with Uniswap. Those related to asset pricing and trading will also be compatible with Uniswap, making Uniswap the de facto standard.

Xuanhao: I find this quite interesting because there are similar examples in the traditional internet space. For instance, Google’s Chrome initially open-sourced its browser kernel Chromium, and then various different versions or browsers were derived from it. Ultimately, these browsers contributed to promoting the Chrome standard, leading Chrome to become the de facto monopolist in the browser space.

Chen Jia: Yes, Uniswap is now the standard for decentralized exchanges. Additionally, in the Uniswap V3 protocol, the liquidity reward certificates are NFTs, non-fungible tokens. In traditional Uniswap V2, the liquidity pool certificates are standardized tokens. If competitors copy Uniswap V3's features, it will lead to a problem: they are currently providing various liquidity mining products, and after migrating to V3, all these liquidity mining products will be unusable or must be done in a very cumbersome way.

This is also a significant reason why many of Uniswap's competitors have not copied Uniswap V3.

Xuanhao: Besides that, it seems that Uniswap is also backed by some top crypto VCs. Paradigm invested in Uniswap back in 2019, and VCs like Paradigm do not just invest and walk away. It seems that the VC team itself collaborates deeply with the Uniswap project team in product design and mechanism design. The backing of top VCs like this can also be a source of competitive advantage.

I think the ecosystem is also very important. If a protocol can grow a prosperous ecosystem, it may also be a point of competitive strength. Just like Ethereum, although public chains and protocols have some differences. But regarding Uniswap, what are the current protocols and applications surrounding it? Since DeFi often emphasizes composability, the most used building blocks may relatively hold more value, and Uniswap is still a foundational protocol. What do you think the landscape of the ecosystem looks like? What roles exist around Uniswap?

Chen Jia: Currently, there are several types of projects or protocols that rely on Uniswap. The first common type is popular decentralized DAOs (tokens), which often list on Uniswap first, like the recently popular decentralized investment DAO (Cult).

In their protocols, a portion of the transaction fees is automatically transferred to the community treasury, and this function is developed based on Uniswap. Many protocols rely on Uniswap's time-weighted oracle, and lending protocol liquidation bots also depend on Uniswap's trading functions.

The swap function of the MetaMask wallet uses Uniswap's pools. There are also some yield aggregators that may provide liquidity in Uniswap. So there are quite a few protocols that conduct business based on Uniswap, with different scenarios, making Uniswap a widely used and relied upon underlying protocol.

Xuanhao: So there are some other components in the on-chain DeFi system, such as lending protocols, liquidity aggregation protocols, or yield protocols, that may use Uniswap as a source of underlying yield.

Another part is that wallet software, which is closer to users, may have some built-in swap functions that directly route to Uniswap to complete transactions. The natural question is, what is the binding depth between Uniswap and these different participants in its ecosystem? For example, if a new DEX emerges that offers a better user experience than Uniswap, or a decentralized trading platform, how high would the cost be for these ecosystem participants to migrate to this new platform? Time costs, development costs, technical difficulties, etc.

Chen Jia: If it is completely based on Uniswap, then the migration cost may not be high. However, if it is entirely different from Uniswap, then the migration cost will be relatively high. The key point is that many protocols did not consider migrating to platforms outside of Uniswap during their development process, so their code cannot be modified to add support. Most protocols are basically unable to modify their support for Uniswap.

Xuanhao: So looking at it this way, its monopolistic nature in the ecosystem is quite strong, as many protocols are deeply bound to it. Once your code is deployed on-chain, it cannot be changed. If you later want to upgrade the version or make significant iterations, it still poses a high switching cost.

Chen Jia: Yes, because many protocols, especially simpler and more singular ones, have immutable contract code. Once deployed, they rely on Uniswap's pools and price oracles. At this point, it's like a high-speed train that is difficult to stop and change its wheels.

Xuanhao: Another topic is about MEV, as Uniswap itself also conducts transactions, and there may be some arbitrage based on on-chain trading mechanisms. Could you talk to us about the basic principles of this arbitrage?

Chen Jia: To put it simply, when people trade on Uniswap, there is an option for slippage. For example, if you set a 10% slippage, and the current price of Ethereum is assumed to be 3000 USD, then by setting a 10% slippage, you can accept a maximum price of 3300 USD for buying Ethereum.

Now, when you place an order to buy one ETH and submit this request to the blockchain, a bot or "sandwich" listens to your request. It finds that the current price is 3000 USD, quickly buys a certain amount of ETH in the pool, raising the price to 3300 USD, and its average purchase cost is around 3150 USD.

After the bot executes your transaction request, you effectively buy ETH at the price of 3300 USD. The bot then sells the ETH at 3300 USD, making a profit from the price difference, and the trading user ends up paying more. Originally, you needed 3000 USD, but now you need 3300 USD.

For this kind of "sandwich" (also known as sandwich attacks), there is a way to alleviate or optimize it. Users can set the price slippage to 1%. This means that when the price is 3000 USD, you can accept a maximum price of 3030 USD for buying one ETH, which eliminates arbitrage opportunities, and the transaction will not be affected by MEV.

Xuanhao: Do you think MEV arbitrage bots will exist in the long term? Because with future technological or protocol improvements, will this market space gradually shrink? We can also see some trading protocols that aim to resist MEV from the user's perspective, trying to avoid the participation of these arbitrageurs. What do you think about the long-term market space in this area?

Chen Jia: In the long term, I think the market space will actually be larger because the number of users on Uniswap is constantly increasing. More transactions are conducted with relatively high slippage, which creates arbitrage opportunities. Due to the nature of blockchain, a transaction must be broadcast first before being packaged, meaning that this problem will always exist: someone can always act before you. The only way to mitigate this is to more kindly remind users to set slippage to 1% or 0.5%, so that these "sandwiches" have no arbitrage space.

Xuanhao: So the MEV mechanism is still due to some characteristics of the blockchain itself, such as the order of transaction packaging. The operational mechanism of the chain indeed cannot be changed, so in the long run, MEV may always exist. Besides that, people may also be concerned about how the value capture mechanism of an on-chain protocol works.

For example, it seems that Uniswap is currently the most important leading protocol in the DeFi ecosystem, but what do you think about the value capture mechanism of the $UNI token? Because Uniswap currently has nearly 80% market share in the DEX space, forming a de facto monopoly, but the price of the governance token is not ideal, which is also an interesting phenomenon. I wonder how you view this issue.

Chen Jia: Currently, Uniswap mainly serves as a governance token, with the primary right being to vote on certain decisions, without involving buybacks or dividends. The interests of UNI holders are not strongly aligned with Uniswap, so in the short term, its price has a bottleneck. This may depend on the community and what more mechanisms or products can empower $UNI through voting, allowing some value from Uniswap to transfer to $UNI.

Xuanhao: There are indeed defects in the value capture mechanism, yet it has become widely used as infrastructure. There are quite a few examples of this, even extreme ones, like Ethernet standards and TCP/IP. People use them every day, but they have little to do with value capture. In the future, perhaps opening up fee-sharing through governance voting or some buyback proposals may bring changes to the market, but it seems that we have not seen much of that yet.

Chen Jia: Yes, there are voices in the community, but they have not gained much support. The $UNI holder community needs to reach a proposal with more supporters through internal discussions.

Xuanhao: Regarding the issue of token dividends, there are also some opinions that Uniswap may have some games with regulators because if you open a dividend mechanism, you may be subject to SEC regulations or influences, so this remains a somewhat ambiguous issue. Speaking of this, the game between the crypto world and the traditional world is also a quite interesting topic.

The crypto world is praised mainly for its characteristics of censorship resistance, decentralization, and permissionless access. Now, for a protocol, I think we can discuss censorship resistance. For example, how might Uniswap be affected by regulatory bodies or enforcement agencies, and to what extent? A specific example is that last July, some tokens were delisted from Uniswap under pressure from the SEC. What do you think about this?

Chen Jia: First of all, as we mentioned at the beginning, Uniswap is a program based on smart contracts running on Ethereum that cannot be modified. It has already been deployed, and its development team cannot stop it from running.

This ensures that Uniswap can continue to operate. However, the front end of Uniswap is controlled by the Uniswap team, which can block certain assets due to policies or even potentially shut down Uniswap's front end due to regulations. However, Uniswap's front end can be self-built; for example, many users in China use third-party mirrors of Uniswap.

This means that even if the official Uniswap team shuts down the front end due to policy reasons, it does not affect users' ability to use Uniswap. Moreover, the tokens that are blocked on Uniswap can still be traded on the blockchain; they just cannot be traded through the official Uniswap front end.

Xuanhao: So for users, the main interaction is with the front end, but in reality, the front end is still relatively open. There may be some third-party front ends or front ends hosted on servers in different countries or regions. Therefore, if you want to block all front ends, it is theoretically quite difficult. Additionally, regarding domain name resolution bans, for example, blocking Uniswap's resolution domain in a certain area is also a feasible approach.

However, it seems that this issue can also be resolved by adding other domains or migrating domains.

It appears that the impact on users is not that significant. For other on-chain protocols based on Uniswap, or from a developer's perspective, if faced with a contract that has already been deployed on-chain, it is indeed impossible to stop it. The businesses based on Uniswap will also not be affected.

Chen Jia: Yes, Uniswap runs on Ethereum as a smart contract, which cannot be shut down or deleted. Therefore, Uniswap can continue to operate, and the front end can be built by any third party.

Xuanhao: The tokens that were previously delisted, although they cannot be traded on Uniswap's front end, can still be traded on 1inch (a DEX aggregation protocol).

In fact, 1inch has become a broad front end for Uniswap, which is quite anti-fragile. I think we have covered a lot of content, discussing Uniswap from the perspectives of product, track, competition, and regulation, which gives us a fairly comprehensive view. Finally, looking ahead, do you think Uniswap will continue to maintain its lead or monopoly? Or do you think that five years from now, Uniswap will still be the most mainstream and foundational on-chain trading application?

Chen Jia: Personally, I am quite optimistic about the continuous development potential of the Uniswap team. Uniswap V2 is a great version and a groundbreaking product that solved many issues from V1, but based on V2, V3 was later launched, which has more advanced features. I believe that in the future, Uniswap may also have V4, V5, and may maintain its leading position in the decentralized exchange space for a considerable time.

Xuanhao: Alright, then we will leave the rest to time to verify whether Uniswap will continue to stand at the forefront or be surpassed by newcomers. Thank you, Chen Jia.