Original Title: "EP30: In-depth Discussion on Stablecoins, Is Stablecoin Still a Blue Ocean?"

Original Source: Mint Ventures

· Host: Alex, Research Partner at Mint Ventures

· Guest: Mindao, Founder of dForce

· Recording Date: May 21, 2025

Disclaimer: The content discussed in this podcast does not represent the views of the guests' respective organizations, and the mentioned projects do not constitute any investment advice.

Hello, welcome to WEB3 Mint To Be initiated by Mint Ventures. Here, we continue to inquire and think deeply, clarifying facts, exploring reality, and seeking consensus in the WEB3 world. We aim to clarify the logic behind the hot topics, provide insights beyond the events themselves, and introduce diverse perspectives.

Alex: In this episode, we have invited our old friend Mindao again. Mindao has discussed many topics with us before, including bringing U.S. stocks onto the blockchain and DeFi. This time, the topic we are going to discuss is a recent policy hotspot, which is stablecoin, perhaps one of the most widely adopted products in the blockchain field. Let's first have Mindao greet us.

Mindao: Hello, I'm Mindao. I'm very pleased to be here today to share some insights on stablecoins.

Differences in Stablecoin in This Cycle

Alex: Alright, let's get into today's topic. We know that stablecoin has gone through many cycles, and in each cycle, key performance indicators and scale have reached new highs in this track. In your opinion, what are some of the key differences to pay attention to in this current cycle of stablecoin development compared to the previous cycle, apart from the regulatory policy issue that we will discuss later?

Mindao: We know that during the DeFi summer of 2021 and 2022, there were many types of stablecoins, especially on-chain stablecoins like TERA, and many completely on-chain algorithmic stablecoins. However, after the collapse of TERA in 2022, we saw a significant structural differentiation on the supply side of stablecoins. Needless to say, the dominant type is still fiat-backed stablecoins. We saw that the total minting volume of stablecoins actually dropped from around $180 billion in 2022, I remember, to about $1300-1400 billion during the Luna collapse, then gradually stabilized and began to rebound. Recently, I've seen data indicating it should be around $250 billion. So, the TVL of the entire DeFi space has not yet reached a new high, but we see that the minting volume of stablecoins has surpassed the previous volume. The driving force behind the minting of these stablecoins is not the native on-chain stablecoins we talk about, but rather algorithmic types, over-collateralized types. The only highlight in this cycle may have been Ethena's USDE.

We just talked about how USDT and USDC belong to the fiat payment type. In fact, Ethena should be considered a wealth management type, based on a valuation arbitrage of native Crypto assets, serving as both a profit-driven and wealth management stablecoin. Strictly speaking, it cannot be called a stablecoin because, for example, if I participate in Ethena's mining, the majority of the risk I bear is its price fluctuation risk, as it cannot achieve a one-to-one peg. Its entire exchange mechanism is not pegged to a USD stablecoin but rather based on its trading platform's valuation arbitrage position. Of course, recently it has undergone a transformation, as I have seen that it has invested approximately tens of billions of dollars in T-Bill assets. However, purely from the perspective of the previous cycle, we see the differentiation of stablecoins, with the payment type being very prominent, still dominated by fiat. Although DAI is considered the oldest decentralized stablecoin, its total minting volume has basically remained at the level of around 5 to 6 billion USD and has not made further breakthroughs.

I believe that since the DeFi summer, one relatively certain aspect is that the narrative of many stablecoins, at least at this stage, has been debunked, such as algorithmic stablecoins. With the clarification of regulations, I think many stablecoins of this type may gradually be phased out. We can see that many of the new stablecoins in the market are currently channel-driven, such as PayPal's stablecoin. Theoretically, USDT and USDC are also channel-driven stablecoins, both originating from a trading platform background. USDT was originally Bitfinex, and USDC is backed by Circle. However, despite the support of these two major trading platforms, they have both faced significant challenges in development. Binance has also supported many stablecoins, with abundant resources, yet none have taken off. So, I think each cycle has its own timing. From this perspective, it is likely that in the future, we will see more channel-driven stablecoin launches, especially through some compliant channels.

The Most Impactful Regulatory Policy

Alex: Understood. You just mentioned a significant change in terms of timing. In the past year, stablecoin legislation, especially progress in the United States regarding stablecoin legislation, has been quite rapid. Just yesterday, the U.S. Senate further advanced a bill called Genius through a procedural vote, which will later enter full Senate discussion and formal voting. In your opinion, which legislative or regulatory policies will have the most impact on the industry? How will they affect the current and future stablecoin industry?

Mindao: Actually, I think the entire regulatory landscape is divided into two major markets, one being the European Union and the other being the United States. In the EU, we have MiCA, which is the Markets in Crypto-Assets regulation, and from a regulatory framework perspective, it is relatively strict. On the U.S. side, last year we also had a much-discussed bill, FIT21, which stands for the Financial Innovation Technology 21st Century Act. This has already been passed in the House of Representatives and should be reviewed in the Senate this year. The bill you just mentioned, the Genius Act, was further advanced in the Senate yesterday. The Genius Act actually focuses on stablecoins, aiming to guide and establish a comprehensive regulatory framework for stablecoins in the United States, placing more emphasis on the stablecoin side.

Recently, everyone has become aware of the RWA track or, in other words, this particular niche has become particularly popular. Stablecoins, as the vanguard of RWA, also represent deposit certificates for the US dollar and tokenization of government bonds. So I think this has a very significant symbolic meaning. A key point here is the clarification and confirmation of the issuer's identity, such as financial institutions and non-financial institutions, but if they have applied for and obtained a license. Of course, with regard to whether technology companies can act as stablecoin issuers, there is no clear definition. There are restrictions on some large tech companies, but whether companies like Facebook or Google count, or whether they can issue stablecoins through their subsidiaries, is not explicitly stated. Basically, from the spirit of the entire legislation, the encouragement is for compliant financial institutions, which naturally includes some non-financial institutions, but not the kind of large tech companies that can enter this field.

I think another important point is regulatory segmentation. The first is at the federal level, where scales exceeding $100 billion are subject to federal regulation, while the rest fall under state-level regulation. Therefore, I believe that the implementation and clarification of regulation will play a very significant role in driving the issuance of stablecoins in the future. In addition, we refer to the United States as the so-called legislative beacon country, with many other countries and regions following frameworks similar to those of the United States. For example, we have seen Hong Kong recently vigorously promoting the so-called stablecoin sandbox, including a Hong Kong dollar stablecoin. During my recent visit to Hong Kong for a conference, I also saw that apart from the Hong Kong dollar stablecoin we mentioned, there are also many teams working on offshore renminbi stablecoins. So I think this U.S. legislation is not only a domestic legislation but also sets an example for other places, including Hong Kong, China, Singapore, and Dubai.

USDT Vs. USDC

Alex: I see. So, if we observe the market share of stablecoins, we find that in the previous cycle, Tether, or as we call it USDT, faced a significant challenge in terms of market share. Including during the last DeFi Summer, we saw a rapid growth in market share for USDC issued by Circle. However, as we entered this current cycle, we noticed that the overall growth rate of USDT is much faster than its competitors, including nearly doubling in total size from the previous peak period, reaching around $800 billion, and now growing to nearly $1.5 trillion to $1.6 trillion. However, looking at USDC, its market value has only increased by less than 20% from its peak in early 2022. As you also mentioned about other stablecoins, their growth rates have not been fast in this current cycle. What could be the reasons for this situation?

Min: Yes, this is a very interesting comparison. Because people have always been comparing USDC and USDT, saying that these two are the biggest beneficiaries of the stablecoin market coming up. But actually, Circle's recent financial report revealed many aspects of the business model that many people may not have seen, including the fact that the profitability of USDC has been overestimated, and a significant portion of the costs are actually channel costs. Additionally, I think many people treat USDT and USDC as an equivalent or identical stablecoin. In reality, these two are different. USDT is closer to a shadow dollar, while USDC is what we traditionally refer to as a stablecoin. The main difference in the growth of the minting volume of these two over the past few years, in my opinion, lies in the analogy that if we consider stablecoins as a river, the reservoir capacity downstream depends on how many tributaries there are upstream. USDC's tributaries, in terms of use cases, include one for financial management and another for deposits and withdrawals. Deposits and withdrawals may be USDC's biggest competitive advantage in the stablecoin space, as it can go one-to-one from trading platforms and mints to the US dollar.

But I think the market structure of USDT is completely different from that of USDC. For example, regarding financial management, there may be some, with various trading platforms accepting it for financial purposes. On the trading side, the use cases for USDT are much more varied than those of USDC. For example, most perpetual contracts on various trading platforms use USDT as collateral. Additionally, concerning OTC circulation, the volume of USDT may be tens to hundreds of times higher in the secondary market than that of USDC. So, as we can see, if we liken this to a river, USDT has far too many tributaries, and its reservoir volume below is much larger than that of USDC. Therefore, I believe this difference is entirely dependent on their respective use cases. USDT is closer to a shadow dollar or even an underground dollar, aligning more with our definition of money rather than just being a stablecoin pegged one-to-one to the dollar. This is because in practice, USDT's price does not match the dollar most of the time; it often experiences market fluctuations. For example, in some offshore markets, there may be premiums or discounts.

Of course, a benefit of the shadow dollar is that it has a vast number of counterparties and is accepted as a common equivalent, with a much broader moat compared to USDC. Another crucial point is that, as seen in Circle's disclosed financial report, its channel fees to Binance and Coinbase are very high; however, USDT is already completely free of any notion of "I have to pay you a channel fee," and many trading platforms have actively listed it. Therefore, in terms of use case and utility, I think USDT surpasses USDC by a significant margin.

Alex: OK. So, with USDT currently holding such a dominant position, I believe that for USDC, as well as many institutions planning to enter the stablecoin issuance field, they also hope to develop towards USDT's current position. In your opinion, is it highly likely for them to achieve a scale close to USDT and attain the same network effect? Taking USDC as an example, as its biggest competitor, do you think it currently has the potential to do so?

Mindao: I actually think USDC may find it quite challenging in this respect, mainly because the upcoming competitors are essentially eating into USDC's market. For example, in terms of fiat on/off-ramps, if banks and financial institutions want to issue a stablecoin, this is their natural advantage, and in this aspect, I don't think USDC can even compete. Just like PayPal, its other fiat on/off-ramps, trade-side, or payment-side channels are actually much stronger than USDC. So, if stablecoins really become law through the House of Representatives this year, I believe it will open up the entire compliant stablecoin entrance, and the new players coming in are actually very strong in traditional channels. USDC relies on Coinbase and Binance, has a first-mover advantage, but this moat is relatively low. For example, Facebook, although currently large tech companies may not be allowed to issue directly, but what if they do it through partnerships or other means? I recently saw that they expressed such an intention, possibly to reboot the Libra project. Even Twitter's X also has significant ambitions in the payment and stablecoin fields. So, if they enter, I think in terms of channels, they will completely overshadow the channels USDC currently collaborates with.

Possibility of Traditional Financial Institutions Entering the Stablecoin Field

Alex: So, for these traditional financial institutions, including those with their own channel advantages in the Internet, do you think they have ambitions for Tether's on-chain, including scenarios beyond their existing traditional fiat on/off-ramp business? What are the possible approaches or methods for them to enter this field?

Mindao: Yes, companies like JD.com in Hong Kong are also preparing to issue an HKD stablecoin, and even Alibaba is making strides in this area, using their e-commerce business as a foothold. But I think Tether's current positioning is quite clever. Firstly, it has a genuine network effect, not like many USDC incentives that are subsidized and maintained through high channel costs. A truly natural network effect, I think Tether is the most typical. This kind of network effect will continue to strengthen and is not easily replaced by a single channel. For example, if JP Morgan were to issue a stablecoin, it might be more convenient for interbank on/off-ramps. However, secondary trading or circulation in OTC may not necessarily surpass Tether's current network effect.

So from this point of view, I think Tether actually has a rather unique competitive advantage in the current regulatory environment. Because the market it has captured is not easily accessible to other compliant stablecoins, or rather not fully covered by other channels. Moreover, it is now not only being used on the trading side, payment side, and OTC side, but basically it can cover almost all scenarios that other compliant stablecoins can cover. Apart from the direct exchange channel with banks, which cannot be opened directly in the European and American markets, everything else, such as secondary circulation and the payment side, has already been covered. So I actually think that these later compliant stablecoins may find it quite challenging to compete with Tether. In fact, Tether has faced many challengers. When it came out in 2015, no one took it seriously at first. Later on, you saw the Huobi stablecoin HUSD, and then OKX had a stablecoin, Binance has already launched its third stablecoin, but none of them have been as strong as Tether, or even reached the scale of USDC. This shows that the network effect and the channel cost required for subsequent promotion are very high.

Alex: Got it. When I saw the two parties discussing the Genius Act this time, one of the major opponents of the Democratic Party, Elizabeth Warren, mentioned a risk point. She believes that once this act is passed, the market size of stablecoins may grow from the current $200+ billion to tens of trillions within a few years. Although she is not satisfied with the current act, believing that the level of control is not sufficient, she has made this prediction. Do you think this prediction is reliable? Can the market really expand tenfold to tens of trillions within a few years? Additionally, is the majority of this ten trillion in new market share likely to go to established leaders like Tether? Or is it more possible that the additional market share will be taken up by some new players, including stablecoins issued by institutions like JP Morgan that you just mentioned, or stablecoins issued by large Internet companies?

Min: Yes, Elizabeth Warren has always been anti-crypto. I think her view on this matter is already biased and unjust. Let's set aside domestic U.S. political issues for now. For example, when this stablecoin act was being proposed, other U.S. lawmakers had previously held a meeting to reveal the so-called Trump family's crypto industry. I think this has a very strong political agenda. For example, in the stablecoin act, there is actually a strong opposition due to Trump's family also launching a stablecoin called USD1. So, I think there are many political issues involved. But setting politics aside, currently, around $180 billion of stablecoins are all invested in U.S. Treasury bonds, making stablecoins one of the top ten holders of U.S. debt. If this market continues to grow, the volume of U.S. debt will undoubtedly grow proportionally. Of course, I think achieving ten trillion is a very ambitious goal. The general estimate is that it may or may not break through the one trillion dollar mark by the end of next year. I think people may still feel somewhat uncertain about the overall growth of stablecoins. For example, when we were doing DeFi at the beginning of 2019, the total value locked (TVL) in DeFi was not even $100 million, roughly around $60-70 million. By the peak of the DeFi summer in 2020, it reached around $250 billion, growing two to three thousand times. The minting volume of stablecoins is similar, going from just over $1-2 billion all the way to over $200 billion today, growing by several hundred times. Of course, achieving such a significant growth purely based on the current players is likely to be quite challenging.

However, we have seen that BlackRock's on-chain volume of T-bills has also suddenly reached a scale of two to three billion US dollars in recent months. So I think if compliant institutions come in, it's not a slow growth; it may double in size within six months to a year. So I think Tether will definitely see a significant increase in market share within this compliant framework, but perhaps even greater growth will come from other new compliant institutions or stablecoin issuance by tech companies. This proportion will gradually rise. There is a very large payment Web2 company called Stripe, whose founder shared that in the past six months, after accepting a stablecoin payment platform or technology integrator called Bridge, they saw data growth several tens or even hundreds of times greater than traditional Web2 data.

So I think in the future, stablecoins will greatly replace the existing interbank settlement and payment infrastructure we talk about, and the speed will be very fast, possibly fully deployed in the next two to three years. So how much stablecoin supply is needed to support the replacement of the new infrastructure? This amount is definitely very large. I don’t think we should understand this situation from the perspective of past DeFi cycles or cryptocurrency cycles because stablecoin development is no longer closely related to the crypto cycle. That's why, even after the DeFi summer crash, stablecoins hit new highs. In fact, the overall change in the crypto assets themselves has not been significant. Because a large amount of capital has come in, it doesn't necessarily mean it has to find arbitrage opportunities within crypto; there may be many on-chain opportunities to buy US Treasury bonds. So I think its growth curve is difficult to compare to the growth of DeFi or crypto assets in a zero-sum game manner. I think this growth, if opened up, could be exponential.

Possible Measures by Countries on Stablecoins

Alex: Got it. So let's take a broader view. The rapid growth of the US dollar stablecoin is evident, with many institutions entering the market. This is actually very helpful for the circulation of the US dollar globally and to cover more scenarios. On the other hand, many international markets are questioning the fundamental value of the US dollar and US bonds. Just the other day, I believe it was Moody's, that downgraded the rating of the US dollar or US bonds. It seems that there is a strong comparison between the two sides. The US has already begun to promote the global deployment of the US dollar stablecoin. Compared to the dollar, do other countries feel a crisis in terms of stablecoins? Or do they have any particular measures in this stablecoin matter? Based on your observations, please share your thoughts on this.

Mingdao: I think this crisis feeling is very strong. Including the US, we talk about MAGA now, MAGA's AI and encryption policy. In fact, I think its only imaginary enemy is China. All talks about promoting the US dollar stablecoin or the US dollar encryption market are benchmarked against China. For example, if the US does not promote the US dollar stablecoin, China's Belt and Road Initiative will do this currency swap. China is also very actively promoting the development of the digital RMB. I think the US not only focuses on the crypto market but also has a very clear understanding of the continuation of the dollar's hegemony. Even the current US Treasury Secretary has worked at the Soros Fund for many years, so he is the most understanding among the former treasury secretaries when it comes to the currency market. When he was at the Soros Fund, he also attacked the British Pound, so his understanding of the currency market is very deep. At the same time, the Secretary of Commerce is also one of the shareholders of Tether. I think they have a very clear understanding of the commercial logic and the relationship between the dollar's hegemony and stablecoins. It's not a superficial understanding but a deep comprehension of this mechanism. Here, I think a metaphor is quite apt: the US dollar stablecoin is a so-called room-temperature superconductor of the US dollar. In the traditional world, there are many electronic settlements for the dollar, with SWIFT, various transfers, but the friction is very high, with regulation from all over and different infrastructures of various systems. When you try to achieve this with a stablecoin ledger, you will find that the efficiency is very high. So I think this metaphor is very apt. It is a room-temperature superconductor with very low friction, yet not a high threshold. Now, as long as you join, as a payment institution, you no longer need to deal with the traditional banking infrastructure. So in this regard, I think it will greatly accelerate the superconductivity of the dollar's hegemony. For example, in the past, interest rate fluctuations in DeFi were very large. Everyone can see that recently, on-chain interest rates, although not entirely anchored to US Treasury bonds, have been very strong in the liquidity of the entire DeFi and in the transmission of the stablecoin interest rate market. This will only become stronger and stronger because more and more underlying protocols are integrating these T-Bill assets. This essentially becomes the US dollar interest rate policy.

Previously, it was actually very difficult to achieve global synchronization of the US dollar interest rate policy. Each bank may have a different rate, but on-chain, in the stablecoin market, I think the transmission of interest rates will become very efficient. In fact, I am not particularly optimistic about so-called stablecoins from small countries. Because transforming a fiat currency into a stablecoin does not necessarily make that stablecoin stronger, more efficient, or more liquid. The underlying factor still depends on the economic strength of the country. So, in the end, maybe only the United States, China, the European Union, and Japan, these few major countries that already have a significant market position in the foreign exchange market and are supported by a substantial economic base, their fiat currencies may enter what I consider to be a very important battleground in the stablecoin space. If you do not enter this battleground, you may gradually become dollarized. Dollarization used to happen in small countries that experienced political or economic instability. If the stablecoin pegged to the dollar continues to maintain this high level of monopoly, I think it may lead to dollarization in the European Union region and even in regions like China. The significant difference between this and traditional dollarization is that the friction cost of switching between this on-chain asset and the national currency is much lower than that of the traditional foreign exchange market. This will result in, if you want the dollar to continue its monopoly, I think regions like the European Union will also see this as a significant challenge to their monetary sovereignty. So, I think this will intensify the competition in the stablecoin race among major countries.

Alex: Got it. In the current situation, I believe that countries should all see the trend regarding the US dollar. So, accelerating the practice of their own national currency stablecoins behind it should also be a high probability event, right? Can we understand it this way?

Mindao: Yes, in fact, everyone understands how stablecoins work now. Essentially, the underlying logic of stablecoins is still to help sell national debt. Think about it, what do we stablecoin holders ultimately do? Isn't it just becoming the ultimate purchasing power of US Treasuries? Whether it's Tether to match, or USDC to match, or our on-chain purchases, they will all eventually be converted into the purchasing power of US Treasuries. In this way, you are effectively reducing the financing cost of the US dollar. So, ultimately, whether it's consumers, financial institutions, or governments in the US, they may be the biggest beneficiaries. For example, the promotion of the Renminbi stablecoin will definitely reduce the financing cost of the Renminbi. I think everyone has understood this point. It's not just about maintaining the monopoly position of a currency in terms of settlement and clearing, but it's actually about linking pricing power, capital flow, and cost of funds completely together.

Centralized Stablecoins Vs. Decentralized Stablecoins

Alex: In the last round, we saw many algorithmic stablecoins that you just mentioned, including many decentralized stablecoins. However, some stablecoin projects that have emerged this time, such as Ethena and PayPal's PYUSD, are actually more typical of stablecoins associated with centralized institutions. So, looking at the current situation, does this mean that institutionalized, centralized stablecoins are actually more suitable for a product like stablecoin? Or, based on the current situation, can we basically determine that the exploration of decentralized stablecoins is very difficult to succeed?

Ming Dao: In DeFi, stablecoins actually started around 2014 or 2015 with Bitshares. At that time, the concept of RMB-pegged stablecoins and USD-pegged stablecoins was already introduced. By 2015, when MakerDAO emerged, it was the first large-scale decentralized stablecoin. I myself have been involved in the stablecoin field since 2019, mainly focusing on decentralized stablecoins. Looking at the evolution of this field, I see that the positioning and narrative of decentralized stablecoins have undergone many changes. Many previous assumptions have been debunked. For example, earlier decentralized stablecoins considered transactional medium and payments as a crucial use case. Without this, the logic behind decentralized stablecoins would be hard to justify. However, in the most recent cycle, several stablecoins that came out mainly focused on financial returns. For example, Ethena, when its supply was increasing rapidly, was primarily driven by the arbitrage spread's returns. At that time, we saw mining returns in the range of over ten percent, and with additional incentives like earning points or Pendle's PT, the returns could even reach twenty to thirty percent, making sUSDE returns very high.

However, towards the end of the previous cycle, the underlying returns had dropped below T-bill yields. Therefore, I believe that in the decentralized stablecoin space, there may ultimately be one or two competitors left. Of course, I think its value proposition is quite distinct from fiat-pegged stablecoins. For instance, DAI is still held by many people because its key feature is the lack of an on-chain blacklist. Yet, DAI has cleverly addressed several issues: Firstly, the majority of its underlying returns come from national debts. Secondly, it has an on-chain reserve with USDC, enabling a one-to-one redemption. Of course, this reserve is quantified, and when it is low, it sells national debts to replenish the on-chain reserves. There was a saying that, in reality, all decentralized stablecoins, regardless of their design, are essentially a wrapper for USDC, meaning they follow the DAI model. While to some extent it is a tokenized version of USDC, its lack of blacklisting, censorship resistance, its ability to incorporate collateral, and some differences in its economic model or monetary policy from traditional fiat currencies give it a unique place and necessity in the crypto world.

Looking back over the past few years, its total supply has remained relatively stable, with little change. I believe that the future of decentralized stablecoins will likely be divided into two categories: Payment-oriented ones, which I think are fundamentally challenging to achieve, it is challenging to scale like USDC, and this model seems unsustainable. The other two categories will find specific use cases. For example, as I just mentioned, the financial products types will be able to aggregate various revenue streams. For example, Ethena has evolved beyond a singular arbitrage strategy; it now includes income from T-bills, making it a mixed-income product. DAI follows a similar path, bundling numerous Ethena strategies underneath. Thus, from the perspective of DAI, it also consists of a mixed strategy. However, you can hardly imagine a centralized stablecoin doing the same. For instance, the recent GENIUS Act in the U.S. expressly prohibits interest-bearing stablecoins. Therefore, financial products types of decentralized stablecoins have an opportunity, especially because they can excellently combine and assemble various strategies. This flexibility sets them apart from traditional fiat-pegged stablecoins.

Starting from the wealth management side, I think it's a great entry point. Another point is to act as an internal accounting system within a DeFi protocol. For example, in our case, we have our own sUSX, which is an internal lending protocol, serving as a ledger asset between different lending protocols. For instance, Aave has something called GHO, which strictly speaking is not a stablecoin, but it serves as a ledger system for liquidity between protocols. The benefit here is that I can use this stablecoin to allocate liquidity across different chains, somewhat like a dollarized equivalent in internal bank accounting. I believe this aspect is necessary in protocol design. Of course, including Curve's crvUSD, which is also intended to be part of a unified liquidity allocation within the entire Curve DEX pool, in order to increase capital efficiency. Therefore, I think that within decentralized stablecoins, we might gradually shift from competing directly with fiat currencies to specific scenarios, such as the wealth management category we just discussed, including inter-protocol equivalents. In this way, the overall positioning and market will be much less comparable to fiat currencies in the future.

Impact of Stablecoin Regulations on Top DeFi Businesses

Alex: I've also heard a view that as the stablecoin market gradually grows, even though, as you mentioned, many stablecoins have not yet entered the crypto scene, there will still be some Total Value Locked (TVL) flowing into our industry, potentially enhancing projects like Pendle and increasing TVL in protocols like Aave's lending. Therefore, some voices believe that once stablecoin regulations are passed, it will be a positive development for some of the top DeFi businesses. What's your take on this viewpoint? If you agree, which projects do you think will see a significant improvement in their fundamentals?

Mindao: I think it can be viewed from two perspectives. I saw yesterday that the market also reflected this, with Aave rising by 20% to 30%. But I don't think Aave rose simply because of stablecoins. I think it's because as a DeFi bank, the more stablecoins, the better the liquidity, and the more fiat stablecoins can come in. In this regard, I believe protocols like Aave or those involved in lending markets like us are definitely in a favorable position because it means more liquidity is coming in. And as stablecoins come in, ultimately, they will also need to go into crypto, leverage assets like Real World Assets (RWA). Therefore, I believe that the so-called projects that are more inclined towards DeFi stablecoins but not directly related can benefit more from this. For example, lending protocols, DEXs, and UNISWAP. Stablecoins also need pools, or you need an exchange pool between the Euro, as well as RWA assets. However, for protocols focused solely on native stablecoins, like Ethena's stablecoin protocol, this could be a significant blow. Because firstly, if Ethena relies solely on the basis arbitrage, this business segment has now been largely consumed by traditional Wall Street financial institutions. They now don't need to go to centralized exchanges or offshore trading platforms like we used to for arbitrage; now many hedge funds on Wall Street directly trade in ETFs, including the CME, to capture these profits. There may be hundreds of billions of dollars being deployed in this manner.

So I think in terms of basis, in the long run, it will definitely be more and more squeezed, ultimately ending up aligning with government bond yields or perhaps even slightly lower. Even when a major bull market arrives, the efficiency of traditional funds entering this market is increasing, and it will squeeze this basis significantly. So, if you rely purely on arbitrage-type strategies on a large scale, I think it's very difficult to scale up, and it's basically unimaginable to reach the scale of USDC or USDT. Therefore, I think that instead, if you were to use stablecoins in a more DeFi management style, it would be very challenging. That's also why I mentioned that Ethena later converted a portion of its assets, roughly tens of billions of US dollars, into T-bills. Whether more will move in that direction may depend on how high the arbitrage return on the basis is. For stablecoins like DAI, which are not purely treasury-based, many holders of DAI do not earn interest and hold it on-chain. For stablecoins like this, I think this news should probably only be considered somewhat bearish and not entirely bullish. But I think it's a huge bullish trend for Bridges, lending protocols, and DEX. And with the arrival of stablecoins, as we just discussed, stablecoins are just the vanguard, with RWA assets to follow. Recently, I've been in Hong Kong, and all my friends from traditional financial institutions are particularly enthusiastic about RWAs, and I don't know why. It's a bit like when the NFT market took off. Everyone in traditional institutions is discussing how to get involved with RWAs. I think this is definitely closely related to the clarification of regulation in the United States as well as the development of stablecoins.

Alex: Stablecoins now have a very strong network effect or what we call the Matthew Effect as a very powerful business category, with Tether already holding a very large market share. Now, the regulatory policies of the market are almost clear, and many institutions want to step in. From the perspective of the current market environment, whether for centralized or decentralized stablecoins, do you think the current state of stablecoins can be classified as a Blue Ocean?

Mindao: Yes, I think stablecoins are still a Blue Ocean right now, but even though the table is still a Blue Ocean, the players have changed. I think it's no longer a Blue Ocean for native crypto teams, but it is still a Blue Ocean for traditional financial institutions and traditional Web2 enterprises. No one has really entered yet; PayPal made an attempt, which can't be considered very successful. But for these native startup teams, I think they might not even necessarily get a seat at the table. I think that's a fairly significant issue, and going forward, the focus might shift to what I just mentioned—branching out into other specific race tracks, not purely competing with fiat in terms of payment or acting as a trading medium.

DeFi, AI, and the Alchemy of Stablecoins

Alex: Okay, then let's take a look at another cross-domain issue. Like you mentioned earlier, in the U.S., there is a stronger desire to drive innovation in two industries, one being crypto and the other being AI. In terms of AI, in this current round, we have seen many so-called AI projects emerge, but the number of projects that truly have product demand and market fit in the industry race is actually quite low. Many people mention a viewpoint, or they have defined a new industry race name called Payfi, believing it has great potential. What is your view on the relationship between this Payfi definition and stablecoins? And could there be some kind of chemical reaction between AI and stablecoins?

Mindao: I think Payfi is a concept created by public blockchains and project teams to tell a story, somewhat similar to SocialFi or GameFi. Whether this thing actually makes logical sense or can be implemented, I still have doubts. Currently, typical Payfi projects in the market are actually based on two narratives: one is how to combine wealth management and payments? How can you have better earnings while making payments? But isn't this perspective essentially the stablecoin business? If stablecoins penetrate various channels, such as USDT or USDC in various payment systems, and can also be deposited into various DeFi protocols for yield farming. For Payfi companies, how much of this opportunity can they capture? There may be some very specific asset categories within this, such as accounts receivable, which may address specific scenarios related to fund turnover; I think this might have some potential. However, I find it hard to categorize this concept as a standalone entity. Of course, the combination of AI and Crypto is another matter, which may overlap a bit with Payfi, but not entirely. This is because we know that AI agents in DeFi require intervention from the payment system.

Indeed, in terms of traditional payments, in terms of integration, it is not as seamless as stablecoins, or native DeFi protocols. So, in this regard, I am more optimistic about the future development of AI in automating fund aggregation and investment decisions. We are also working on something that combines AI and DeFi. I think a significant barrier to the development of traditional DeFi is that all previous DeFi required all logic to be written into the smart contract, making scalability particularly challenging. This is also why from 2019 to now, there hasn't been much change in the foundational protocols of DeFi; it's still those few AMMs, DEXs, lending protocols, and stablecoins. Although there have been contract-based protocols recently, no player has truly captured as much market share in this space as Uniswap has in AMM or DEX. The evolution of DeFi over the past few cycles has actually been slow, and one of the biggest reasons for this is that all logic is 100% on-chain, making auditing costs and other aspects very high. In the future, with the combination of AI agents and DeFi, I think a significant change will be the transformation of the entire development model, where perhaps only 10% to 20% of on-chain logic will remain, and the remaining 80% of the logic will be implemented by AI agents.

We have also seen some early use cases slowly emerging in the market. And as AI's inference capability continues to strengthen, many things we introduced back when we were trading in the traditional sense are very different from the flexibility and scalability of LLM now. So recently, we are seeing, I think in the combination with DeFi, and stablecoins are definitely the core use case, such as an AI agent paying for various services, whether to use a traditional payment medium or stablecoin for payment, stablecoin is certainly a more coherent way. Another interesting point is that everyone is now talking about this AI agent eventually forming a peer-to-peer closed loop. Your income and costs are both on-chain, and at the agent level, the entire loop can be completed autonomously, without any human intervention. In this context, as the agent becomes more and more automated, you must integrate a framework that can aggregate your on-chain income and facilitate on-chain payments, which is the most natural thing for stablecoins and DeFi in conjunction with agents.

Alex: Got it. You mentioned earlier the combination of DeFi and AI, which is actually a topic I have been particularly interested in recently. You mentioned that the development speed of smart contract-based DeFi was relatively slow because most of its logic needs to be executed on smart contracts and on-chain. You also mentioned a point that perhaps 80% of the logic will be implemented by AI in the future, and the remaining 20% will be within the on-chain smart contracts. What does this 80% correspond to?

Min: Let me give you an example. For instance, we are currently working on a cross-chain yield aggregation product. Traditionally, this is not possible in DeFi. First of all, across different chains, involving information synchronization, atomic transactions are definitely not feasible. You can use some cross-chain protocols, such as LayerZero, to perform some integrations, but the logic required for different sources of funds across different chains and configurations in different chains and protocols has become so complex that no single contract can be written and deployed across different chains to solve the problem, which traditional DeFi cannot achieve. But through the AI Agent, traditional DeFi in this area only handles user deposits and withdrawals, as well as on-chain strategies, and then moves to different protocols for cross-chain transactions. Essentially, it goes back to what we were saying, that Crypto's settlement logic needs to be implemented on-chain to ensure that the ledger is at least verifiable and transparent. However, all the intermediate logic, such as withdrawing from Aave on Ethereum, crossing over to Base to deposit into Morpho, then leveraging in Morpho, and looping again, involves a lot of logic changes. I think a major issue with traditional DeFi lies in business logic. Iterations of products like Binance or CeFi are calculated on a daily or weekly basis. When something new comes out, it can be launched immediately, which is the centralized service logic of a centralized exchange platform like Binance. However, all the logic in DeFi must be completely on-chain. We see Uniswap, Aave, MakerDAO; these are the basic DeFi protocols we talk about, and the iteration cycle is calculated in two to three-year intervals. The reason it takes such a long time is that all DeFi logic is static, but business changes are dynamic. Today, you have to reveal one strategy, tomorrow, you have to reveal another strategy; this is dynamic.

So I think AI Agents are particularly suitable for extending many dynamic logics. In the past, when developing AI, we almost had to exhaustively consider every rule. But now, many reasoning models basically don't need to exhaust rules and scenarios. For example, when we talk about the cost issue of cross-chain transactions, the AI Agent can understand how Gas fees are spread over a certain number of days, how they affect APY, without us needing to redefine a very detailed rule. So I think in the future, apart from the on-chain settlement of funds in and out that I just mentioned, most of the logic can be implemented through Agents. The biggest difference between this and traditional CeFi, and the biggest difference with Binance, is whether human intervention is needed. For example, if Binance doesn’t have a team to optimize and intervene, the business will definitely not run smoothly. So I think that in DeFi, there may be some relatively simple logic that can be fully optimized and implemented by an Agent without the need for team intervention. The next step may expand to different areas, from yield aggregation to lending, to swaps. I think eventually all of these will be implemented by Agents. In fact, in this cycle, we have seen projects like Ethena, it's hard to say it's a DeFi project; all funds are inside the trading platform and everything is controlled by the team to run arbitrage strategies. But in the future, I think all underlying DeFi protocols will be transformed using AI, all to be gradually replaced by Agents. I think this trend is very clear. If the so-called DeFi product of the next cycle appears, it will no longer be our traditional fully on-chain contract logic-based DeFi product.

Alex: Got it. So for products like this next-generation DeFi overlaying an AI module, if AI accounts for 80% of the execution logic, how should the verifiability of this AI module or the determinism of its output results be ensured?

Mindao: Yes, I think this is currently the biggest issue. The illusion with AI models is still very obvious. For example, with the same request, providing you with funds, data, APIs, you give me a strategy. Right now, with most AI models, if you ask them the same question 10 times, they will give you different strategies each time. There are certainly some issues that need to be resolved. But I think compared to testing these models three months ago, progress is very, very rapid now, and the decrease in illusion is very significant. You'll find that its strategy still conforms to our cross-validation, although there is an issue with consistency; it is difficult to ensure that 100 identical requests will have 100 identical results. But at least I think for the most part, it gets the arithmetic right. One way to address this issue is to further refine the workflow of the agent for each task to reduce the frequency of illusions. I think as reasoning models improve, this problem should not be overly worrying. Maybe in another year, including the recent Grok version 3.5, if it really is based on first principles or so-called physical principles for reasoning, many of the illusion problems will slowly be resolved. Perhaps in the future, just by entering a large prompt, very high-quality results can be output. So the benefit of this problem lies in the improvement of the foundational models.

We can see that in DeFi, there are actually many projects slowly developing MCP. Once MCP is established, it essentially further subdivides the workflow. I think we may see more specialized MCPs in the future. For example, if you ask me about a strategy, my MCP may specialize in running a lending arbitrage strategy, and I can provide you with an executable strategy that has been validated in various ways. So, I think the interesting point in the future is that, previously in DeFi, we talked about modularity or composability, where you could combine AAVE's code or Uniswap's code to create a different type of AMM. Looking at it from the perspective of AI and DeFi development, there may be many specialized MCPs emerging. MCP itself is a module, and when combining different MCPs, many new functions can be achieved. I think modularity is developing very fast at the MCP level. Of course, this will give rise to new security issues, such as whether the AI environment has been poisoned, leading to various risks. However, compared to scalability, I see these as minor issues. Because in reality, DeFi security has evolved over these many years of development.

Therefore, I believe that AI might actually make it easier to solve many of DeFi's security issues. Let me give you an example. Currently, many traditional financial institutions are not very willing to enter DeFi, and a major reason is that, for example, we spend millions of dollars on audits, but I cannot guarantee that there is no risk involved, I can never vouch for it. However, traditional financial institutions do not have such issues; they can't say all the money in my bank has been stolen. Whereas, in DeFi protocols, in the event of a hack, all assets within the protocol are compromised, and most of the time, when a problem arises, the entire pool is drained. However, if the logic behind future DeFi projects is largely AI-driven, then at the agent level, many traditional Web2 risk control methods can actually be incorporated. For instance, withdrawal limits could be enforced through an agent. Therefore, it might be easier to fine-tune security in a more granular manner. Traditional DeFi cannot achieve such granularity due to issues like gas fees and new bugs and security concerns.

Alex: Understood. One last small question, also based on the topics we've discussed here. You mentioned earlier that a lot of code used to require auditing by audit firms. From your observations, with the advancement of AI, is the security cost of a DeFi project increasing or decreasing? Has there been a noticeable decrease in security costs over the past year or two?

Mindao: I think when it comes to cost reduction, we need to consider a few factors. One major issue with DeFi is that it's very challenging for audits to cover all possibilities, not only in code audits but also in formal verification. Even formal verification finds it difficult to exhaustively cover all security boundary logic. That's why in traditional DeFi, everyone is reluctant to develop overly complex products because, with complexity, logic bugs and various edge cases arise. In comparison, a DeFi product versus a DeFi product enhanced with AI, I think the latter's security audit costs and security boundaries are much more manageable. I don't have specific numbers because it's still very early, but my personal feeling is that DeFi AI products can achieve more logic, reducing the need for extensive audit costs. The capabilities and possible logical scenarios between the two are quite different.

Alex: Got it. We had a comprehensive discussion today, starting from stablecoins and expanding to the connection between stablecoins and DeFi, DeFi and AI, and various intersections involving AI, DeFi, and stablecoins. Thank you very much, Professor Mindao, for being our guest on the show today and sharing so many insightful perspectives. Thank you.

Mindao: Sure, thank you.

Original Article Link