In-depth discussion on stablecoins: Is Stablecoin still a blue ocean?
TechFlow Selected TechFlow Selected
In-depth discussion on stablecoins: Is Stablecoin still a blue ocean?
Tether is the most typical example of genuine natural network effects. These network effects will continuously strengthen and are not easily replaced by a single channel.
Navigating Web3 tides with focused insightsHost: Alex, Research Partner at Mint Ventures
Guest: Mindao, Founder of dForce
Recording Date: 2025.5.21
Disclaimer: The views discussed in this podcast do not represent the opinions of the institutions affiliated with the guests, and any projects mentioned do not constitute investment advice.
Hello everyone, welcome to WEB3 Mint To Be, initiated by Mint Ventures. Here, we continuously question and deeply reflect, clarifying facts, understanding realities, and seeking consensus within the WEB3 world. We aim to unravel the logic behind trending topics, provide insights beyond surface-level events, and introduce diverse perspectives.
Alex: For this episode, we’ve invited our good friend Professor Mindao once again. Professor Mindao has previously joined us to discuss various topics, including tokenized stocks and DeFi. Today’s topic is a recent policy hotspot—stablecoins—which may also be one of the most widely adopted products in the blockchain space. Let’s first invite Professor Mindao to say hello.
Mindao: Hello everyone, I’m Mindao. Very happy to be here today to share some thoughts on stablecoins.
What's Different About Stablecoins in This Cycle?
Alex: Great. Let’s dive into today’s main topic. As we know, stablecoins have survived multiple cycles, with core business metrics and scale hitting new highs in each cycle. In your view, how does the development of stablecoins in this current cycle differ from the previous one—aside from regulatory policies, which we’ll discuss later? What are the key differences worth noting this time around?
Mindao: During the DeFi summer of 2021–2022, there were actually many types of stablecoins, especially on-chain ones like TERRA and numerous algorithmic stablecoins fully based on blockchain. But after TERRA collapsed in 2022, we saw significant structural changes on the supply side of the stablecoin landscape. Dominant players remain fiat-backed stablecoins. We observed that stablecoin minting volume dropped from around $180 billion in 2022—down to about $130–140 billion during the Luna crash—before stabilizing and rebounding. Recently, I checked the data and it’s now reached $250 billion. While overall DeFi TVL hasn’t hit new highs yet, stablecoin issuance has already surpassed previous peaks. Importantly, this growth isn't driven by native on-chain stablecoins—neither algorithmic nor over-collateralized models. The only standout in this past cycle was Ethena’s USDE.
USDT and USDC fall under the fiat-backed payment category, while something like Ethena is more of a yield-generating or financial product, built on native crypto assets through arbitrage and yield strategies. Strictly speaking, it can't even be called a stablecoin because if you participate in Ethena mining, you’re mostly exposed to price volatility risk—it doesn’t guarantee a 1:1 redemption. Its redemption mechanism isn't tied to the dollar but rather relies on pricing and arbitrage positions on its trading platform. Although recently, they’ve started shifting—I noticed they purchased around ten billion dollars’ worth of T-Bills. But purely from last cycle’s perspective, we see a clear divergence: payment-focused stablecoins remain dominated by fiat-backed models. DAI, despite being the oldest decentralized stablecoin, has plateaued at around $5–6 billion in issuance without further breakout.
I think one thing that became clear post-DeFi summer is that many narratives around stablecoins have been disproven—at least for now—especially those around algorithmic models. With clearer regulations emerging, I believe many such stablecoin types may gradually phase out. Meanwhile, many newly launched stablecoins today are channel-driven, such as PayPal’s. Even USDT and USDC are theoretically channel-driven stablecoins backed by exchange ecosystems—USDT originally by Bitfinex, USDC by Circle. Yet even with these two major exchanges supporting them, their growth wasn’t easy. Binance has backed multiple stablecoins with strong resources, yet none succeeded. So I believe each cycle has its own timing. Given the current environment, I suspect future stablecoin launches will increasingly be led by compliant channels.
The Most Impactful Regulatory Policies
Alex: Understood. You just mentioned a major shift—the issue of timing. Over the past year, stablecoin legislation, particularly in the U.S., has advanced rapidly. Just yesterday, the U.S. Senate moved forward procedurally on the Genius bill, which will soon enter full chamber debate and formal Senate voting. In your opinion, which legislative or regulatory policies could have the greatest impact on the industry, and how might they shape the present and future of the stablecoin sector?
Mindao: Regulation mainly revolves around two major markets: the EU and the U.S. The EU’s MiCA (Markets in Crypto-Assets Regulation) is relatively strict in terms of regulatory framework. On the U.S. side, last year’s FIT21—Financial Innovation and Technology for the 21st Century Act—was widely discussed. It passed the House and is expected to be reviewed by the Senate this year. The Genius bill, recently advanced in the Senate, differs from FIT21 in that it specifically focuses on stablecoins, aiming to establish a comprehensive regulatory framework for U.S. stablecoins, thus centering more on the stablecoin aspect itself.
Recently, RWA (Real World Assets) has become extremely hot. As the vanguard of RWAs, stablecoins serve as digital representations of U.S. dollar deposits and Treasury securities. This is highly symbolic. Key aspects include clarifying and confirming issuer identities—financial institutions, or non-financial entities holding proper licenses. However, whether tech companies qualify as issuers remains undefined. Large tech firms face restrictions, but it’s unclear whether companies like Facebook or Google—or their subsidiaries—can issue stablecoins. Still, the spirit of the legislation encourages compliant financial institutions, and possibly certain non-financial entities (excluding large tech firms), to enter this space.
Another crucial point is layered regulation: federally regulated for issuers above $10 billion in scale, state-regulated below that. This clarity will significantly boost future stablecoin issuance. Additionally, the U.S. is seen as a "legislative lighthouse," with many other countries modeling their frameworks after it. For example, Hong Kong is aggressively pushing its stablecoin sandbox and HKD-pegged stablecoins. During my recent visit, I saw numerous teams working not only on HKD stablecoins but also offshore RMB versions. Thus, U.S. legislation won’t just affect domestic markets—it will set precedents for regions like Hong Kong, Singapore, and Dubai.
USDT vs. USDC
Alex: Understood. Observing stablecoin market share, we notice that in the previous cycle, Tether (issuer of USDT) faced significant challenges. During the last DeFi Summer, Circle’s USDC gained market share rapidly. But in this cycle, USDT’s growth has far outpaced competitors. Its total scale, previously peaking around $80 billion, has nearly doubled to over $150–160 billion. In contrast, USDC’s market cap has grown less than 20% from its early 2022 peak. Other stablecoins, as you mentioned, haven’t grown quickly either. What factors explain this shift?
Mindao: This is indeed an interesting comparison. Both USDC and USDT are often viewed as top beneficiaries of the stablecoin market’s rise. However, Circle’s recently disclosed financials revealed hidden aspects of their business model—including the fact that USDC’s profitability was overestimated due to high distribution costs. Many people treat USDT and USDC as equivalent, but they are fundamentally different. USDT functions more like a shadow dollar, whereas USDC fits the traditional definition of a stablecoin. The difference in minting growth stems largely from usage breadth—if we view stablecoins as rivers, downstream volume depends on upstream tributaries. USDC’s use cases include yield generation and on/off-ramping. Its biggest competitive advantage lies in off-ramping: users can seamlessly convert USDC 1:1 into USD via exchanges or issuers.
However, USDT and USDC operate in entirely different market structures. While USDT supports yield opportunities across platforms, its utility in trading vastly exceeds USDC’s. Most exchange perpetual contracts use USDT as collateral, and in OTC markets, USDT’s circulation volume is tens or even hundreds of times greater than USDC’s secondary market liquidity. Viewed as a river system, USDT has far more tributaries, resulting in much larger downstream reserves. This reflects fundamental differences in use case and utility. USDT occupies a position closer to “shadow” or “underground” dollars—more aligned with broader definitions of money rather than just a 1:1 dollar-pegged instrument. In reality, USDT rarely trades exactly at parity; it often floats based on market conditions, with premiums or discounts in offshore markets.
A key advantage of shadow dollars is their vast network of counterparties and wider acceptance as a medium of exchange compared to USDC, giving it a much deeper moat. Another critical factor: Circle’s financials show high channel fees paid to Binance and Coinbase. In contrast, USDT doesn’t pay listing fees—many exchanges list it proactively. Therefore, in terms of usability and utility, USDT significantly outperforms USDC.
Alex: OK. With USDT now dominating so heavily, institutions like USDC and others planning to launch their own stablecoins likely aspire to reach similar dominance. Do you think it's still possible for them to approach USDT’s scale and achieve comparable network effects? For instance, taking USDC as the primary competitor, do you believe it still has this potential?
Mindao: I think USDC faces considerable difficulty here, as newer entrants primarily compete within USDC’s existing market—fiat on/off-ramps. Traditional banks and financial institutions launching stablecoins have a natural advantage in this area, one USDC cannot match. Take PayPal: its on/off-ramp infrastructure, trade, and payment channels are far stronger than USDC’s. If stablecoin legislation passes this year, it would open the door for compliant stablecoins, bringing in new players with powerful traditional distribution networks. USDC benefits from early partnerships with Coinbase and Binance, but its moat is shallow. Consider Facebook—even if large tech firms are barred from direct issuance, what if they partner or use alternative routes? They’ve recently signaled intentions to revive Libra. Twitter’s X also has big ambitions in payments and stablecoins. If they enter, their distribution power would completely overshadow USDC’s current partners.
Potential Entry of Traditional Financial Institutions into Stablecoins
Alex: Regarding traditional financial institutions and Web2 companies with strong distribution advantages, do you think they have ambitions beyond Tether’s current on-chain and traditional on/off-ramp use cases? What pathways or methods might they take to enter this space?
Mindao: Yes, JD.com plans to issue a Hong Kong dollar stablecoin in Hong Kong, and Alibaba is also making moves, leveraging their e-commerce ecosystems. But Tether’s current positioning is quite strategic. First, it enjoys genuine network effects—not artificially subsidized like USDC, which incurs high channel maintenance costs. Tether’s organic network effect is the most prominent example, reinforcing itself and hard to replace via single-channel competition. For instance, JP Morgan issuing a stablecoin might streamline interbank settlements, but it would struggle to match Tether’s dominance in secondary trading or OTC liquidity.
Thus, Tether holds a unique competitive edge in today’s compliant environment. It dominates markets where other compliant stablecoins can’t easily enter or fully cover. Moreover, Tether isn’t limited to trade, payments, and OTC—it already overlaps with nearly all use cases covered by compliant stablecoins. The only gap is direct 1:1 bank redemption in Western markets, which remains unavailable. Beyond that, it’s already active in secondary circulation and payments. Therefore, I believe new compliant stablecoins will find it tough to compete directly with Tether. Despite numerous challengers—HUSD from Huobi, OKX’s stablecoin, Binance’s third attempt—none have surpassed Tether, or even matched USDC’s scale. This demonstrates the strength of network effects and the immense channel costs required to challenge incumbents.
Alex: Understood. I noticed that during bipartisan discussions on the Genius bill, Democratic opponent Elizabeth Warren raised a concern: she believes the bill, if passed, could grow the stablecoin market from its current $200+ billion to several trillion within years. Though dissatisfied with the bill’s oversight level, she made this prediction. Do you find this credible? Could the market truly expand tenfold to several trillions in a few years? And would this new market share primarily go to leaders like Tether, or could newcomers—like JP Morgan or large internet firms—capture most of the growth?
Mindao: Elizabeth Warren has long opposed crypto; her stance here seems biased and unfair. Setting aside domestic politics—for example, the congressional hearing on Trump family’s crypto ventures, which carried heavy political overtones—consider this: current stablecoins hold about $180 billion in Treasuries, already ranking among the top ten holders. Continued growth means proportional increases in Treasury holdings—this is certain. However, reaching $10 trillion is ambitious. Most estimates suggest hitting $1 trillion by next year-end. People still underestimate stablecoin growth potential. Back in early 2019, DeFi’s total TVL was under $100 million—around $60–70 million. By DeFi summer 2020, it peaked near $250 billion—a 2,000–3,000x increase. Stablecoin issuance grew similarly—from $1–2 billion to over $200 billion—hundreds of times over. Of course, achieving such multiples solely through existing players is difficult.
But consider BlackRock: in recent months, its on-chain T-bill volume surged to $20–30 billion overnight. If compliant institutions enter, growth won’t be gradual—it could double within six to twelve months. Tether will certainly gain more under this compliant framework, but even larger gains may come from new compliant institutions or tech company-issued stablecoins. Their share will gradually rise. Stripe, a major Web2 payments company, acquired a stablecoin payment integrator called Bridge. After integration, transaction data grew dozens or hundreds of times compared to traditional Web2 metrics.
I believe stablecoins will rapidly replace existing banking settlement and payment infrastructure—possibly completing deployment within two to three years. The volume of stablecoin issuance needed to support this transition will be massive. Don’t analyze this through past DeFi or crypto cycles, as stablecoin growth is now decoupled from crypto trends. That’s why stablecoins hit new highs after DeFi summer crashed, despite little change in overall crypto assets. Incoming capital isn’t necessarily seeking crypto arbitrage—it finds yield opportunities directly on-chain via U.S. Treasuries. Thus, comparing stablecoin growth to DeFi or crypto expansion is flawed. If opened up, this growth could be explosive and discontinuous.
Potential Actions by Countries on Stablecoins
Alex: Understood. Zooming out, the rapid growth of dollar-backed stablecoins, fueled by institutional entry, greatly enhances dollar circulation and expands its global footprint. On the flip side, many international observers are questioning the foundational value of the dollar and U.S. Treasuries. Recently, Moody’s downgraded either the dollar or Treasuries. These contrasting dynamics seem stark. The U.S. is actively pushing dollar stablecoin adoption globally. Compared to the dollar, do other nations feel a sense of crisis regarding stablecoins? What actions might they take? Please share your observations.
Mindao: The sense of crisis is very real. Even MAGA’s AI and crypto policies, in my view, ultimately frame China as the sole strategic rival. All discussions about promoting dollar stablecoins or U.S. crypto markets benchmark against China. If the U.S. doesn’t act, China’s Belt and Road currency swaps and aggressive push for RMB digital currency could fill the void. The U.S. clearly understands the need to preserve dollar hegemony—not just in crypto, but broadly. The current Treasury Secretary spent years at Soros Fund Management, making him the most knowledgeable about currency markets among recent secretaries. He previously participated in attacks on the British pound, indicating deep expertise. Meanwhile, the Commerce Secretary is a Tether shareholder. They deeply understand the interplay between commercial logic, dollar dominance, and stablecoins—not superficial awareness, but systemic comprehension.
A fitting analogy: dollar stablecoins are like room-temperature superconductors for the dollar. In the traditional world, dollar transactions involve electronic settlements, SWIFT, wire transfers—all burdened by friction, local regulations, and fragmented infrastructure. When implemented via a shared ledger using stablecoins, efficiency skyrockets. Hence, the “superconductor” metaphor fits: minimal friction, low barriers, and no need for legacy banking systems. This will dramatically accelerate dollar dominance. Previously, interest rate fluctuations in DeFi were extreme. Now, rates on-chain—while not perfectly pegged to Treasuries—are increasingly influenced by them, especially as more protocols integrate T-Bill assets. This transmission will only strengthen.
Previously, U.S. monetary policy struggled to synchronize globally—banks had varying rates. But on-chain, in the stablecoin market, rate transmission becomes highly efficient. I’m not optimistic about small nations’ stablecoins. Simply tokenizing a fiat currency doesn’t automatically make it stronger, more efficient, or liquid—the underlying economy determines success. Ultimately, only major economies like the U.S., China, EU, and Japan—with established forex presence and economic scale—will compete seriously in the stablecoin arena. Failure to participate risks gradual dollarization. Historically, dollarization affected politically or economically unstable small nations. But if on-chain dollar stablecoins maintain high monopoly, they could lead to dollarization even in regions like the EU or China. The key difference from traditional dollarization is the drastically lower friction when switching between on-chain assets and local currencies—far below traditional FX market costs. This forces major economies to view U.S. dominance as a threat to monetary sovereignty, intensifying great-power competition in the stablecoin race.
Alex: Understood. Given this landscape, I believe all nations recognize the U.S. trajectory. Is it highly likely they’ll accelerate efforts to launch their own national currency stablecoins? Can we interpret it this way?
Mindao: Yes, the mechanics of stablecoin businesses are now well understood. At its core, stablecoins help sell government debt. Think about it: stablecoin holders ultimately become end buyers of Treasuries. Whether Tether allocates funds, USDC invests, or users buy directly on-chain, demand eventually flows into Treasuries. This lowers U.S. funding costs. Thus, American consumers, financial institutions, and the government become the ultimate beneficiaries. Similarly, promoting RMB stablecoins would reduce China’s financing costs. Everyone sees this clearly. It’s not merely about maintaining a currency’s dominance in clearing and settlement—it’s deeply tied to pricing power, capital flow, and funding cost control. These are two sides of the same coin.
Centralized vs. Decentralized Stablecoins
Alex: Last cycle, we saw many algorithmic and decentralized stablecoins. This cycle, however, new projects like Ethena and PayPal’s PYUSD are clearly linked to centralized entities. Does this suggest institutional, centralized stablecoins are better suited to this product category? Or can we conclude that decentralized stablecoin experiments are unlikely to succeed?
Mindao: In DeFi, stablecoins date back to BitShares around 2014–2015, proposing CNY and USD-pegged versions. MakerDAO introduced the first large-scale decentralized stablecoin in 2015. Since 2019, I’ve worked extensively in this space, focusing on decentralized models. Looking at the evolution, the positioning and narrative of decentralized stablecoins have shifted significantly—many original assumptions have been disproven. For example, earlier models emphasized payment and transaction mediums as key use cases. Without this, the logic for decentralized stablecoins weakens. But in this cycle, new stablecoins focus primarily on yield generation. Ethena, for instance, saw fastest minting growth when offering basis arbitrage yields—mining returns reached double digits, sometimes 20–30% with积分 and Pendle PT incentives, making sUSDE highly attractive.
Yet toward the end of the last cycle, its underlying yield fell below T-Bill rates. Thus, I believe only one or two decentralized stablecoins may survive, serving purposes distinct from fiat-backed ones. DAI, for example, retains users due to its lack of blacklisting features. DAI cleverly solved two issues: most of its yield comes from Treasuries, and it maintains on-chain reserves allowing 1:1 conversion with USDC. Of course, this reserve is finite—when low, it sells Treasuries to replenish reserves. Hence, a common saying emerged: regardless of design, all decentralized stablecoins ultimately become wrappers of USDC—precisely describing DAI’s model. While technically a wrapped version of USDC, DAI’s censorship resistance, absence of blacklists, and unique collateral models give it niche relevance in the crypto ecosystem.
Over recent years, its minting volume has remained stagnant. Going forward, I see two viable paths for decentralized stablecoins: payment-focused models seem unviable—they won’t reach USDC-scale volumes, and this path appears blocked. Instead, success may lie in specific niches. One is yield aggregation—combining diverse income sources. Ethena, for example, no longer relies solely on arbitrage; it now includes T-Bill yields, creating a hybrid yield product. DAI similarly integrates Ethena-like strategies, becoming a composite yield vehicle. Centralized stablecoins struggle here—recent U.S. GENIUS Act explicitly bans interest-bearing stablecoins. Thus, decentralized yield-focused models have opportunities, especially in strategy composition and flexibility unmatched by traditional fiat stablecoins.
Yield-centric approaches offer promising entry points. Another role is internal accounting within DeFi protocols. For example, we have sUSX—an internal lending protocol used as an accounting token across our systems. Aave’s GHO, while not strictly a stablecoin, serves as a liquidity accounting mechanism between protocols. This allows efficient cross-chain liquidity management—similar to an internal dollar-equivalent bookkeeping tool among banks. Such functionality proves necessary in protocol design. Curve’s crvUSD aims to become a unified liquidity pool across Curve’s DEX, improving capital efficiency. Thus, decentralized stablecoins may evolve from competing with fiat to serving specialized roles—yield aggregation and inter-protocol equivalents—making direct comparisons with fiat currencies less relevant.
Impact of Stablecoin Legislation on Top DeFi Protocols
Alex: There’s a view that as stablecoin market size grows, even if—as you noted—much capital doesn’t flow into crypto directly, some TVL will inevitably enter our sector, potentially boosting protocols like Pendle and Aave. Some argue that passing the stablecoin bill would benefit leading DeFi projects. What’s your take? If you agree, which projects stand to gain the most?
Mindao: Two perspectives here. Yesterday’s market reaction showed Aave rising 20–30%. But I don’t think Aave rose because it has a stablecoin. Rather, as a DeFi bank, more stablecoins mean better liquidity and more fiat-backed stablecoins entering the ecosystem. Protocols like Aave or lending platforms clearly benefit from increased liquidity. Moreover, stablecoins entering the system will eventually leverage crypto or RWA assets. Thus, protocols indirectly related to native stablecoins—lending platforms, DEXs like Uniswap (which needs stablecoin pools), or RWA asset protocols—stand to gain significantly.
In contrast, native stablecoin protocols like Ethena face headwinds. First, Ethena’s basis arbitrage is increasingly captured by traditional Wall Street firms. Unlike before—when arbitrage required access to centralized or offshore exchanges—today, many hedge funds execute directly via ETFs and CME, extracting profits efficiently. Hundreds of billions in capital now operate here. Long-term, basis arbitrage margins will compress, converging toward or even below Treasury yields. Even in bull markets, traditional players now enter faster, squeezing spreads aggressively. Thus, scaling purely on arbitrage strategies is nearly impossible—unimaginable to reach USDC or USDT scale. Consequently, yield-focused stablecoins like Ethena face major challenges. This explains why Ethena shifted ~$10 billion into T-Bills. Whether more shifts depend on arbitrage returns. For non-yield-focused stablecoins like DAI—held by users who don’t seek yield—this news is neutral-to-negative, not outright positive. But for Bridge, lending protocols, and DEXs, it’s a major tailwind. With stablecoins arriving as vanguards, RWAs follow. Recently in Hong Kong, all my contacts from traditional finance are obsessed with RWA—similar to the NFT breakout moment. Every institution discusses RWA strategies. This enthusiasm is clearly tied to U.S. regulatory clarity and stablecoin advancements.
Alex: Stablecoins exhibit strong network effects or Matthew effects, with Tether dominating market share. With regulations nearing clarity and many institutions preparing to enter, do you consider the current stablecoin landscape—whether centralized or decentralized—still a blue ocean at this juncture?
Mindao: Yes, I still see stablecoins as a blue ocean—but while the table remains open, the players have changed. For native crypto teams, it’s no longer blue; for traditional financial institutions and large Web2 enterprises, it remains wide open. Currently, no major player has entered successfully—PayPal tried, but not decisively. For native startup teams, simply getting a seat at the table may be impossible. Hence, the focus should shift to niche verticals—not direct competition with fiat in payments or as mediums of exchange.
The Chemical Reaction Between DeFi, AI, and Stablecoins
Alex: Let’s explore a cross-domain topic. As you mentioned, the U.S. seeks innovation in two sectors: crypto and AI. This cycle has seen many AI projects emerge, yet few align product-market fit. Some propose a new category called PayFi, suggesting significant potential. What’s your view on PayFi’s relationship with stablecoins? Could AI and stablecoins generate synergies?
Mindao: PayFi feels like a narrative invented by blockchains and project teams—similar to SocialFi or GameFi. Whether it logically holds or practically works remains questionable. Current PayFi projects emphasize two narratives: combining payments with yield—how to earn while paying. But isn’t that essentially what stablecoins already do? If stablecoins permeate payment systems—USDT or USDC earning yield in DeFi protocols—how much room remains for PayFi startups? Perhaps narrow niches exist—e.g., accounts receivable solutions addressing specific cash turnover scenarios. But the concept itself struggles to form a standalone major category. AI combined with crypto is another matter—partially overlapping with PayFi but not identical. AI agents in DeFi require integrated payment systems.
Traditional payment integration lacks the seamlessness of stablecoins or native DeFi protocols. Thus, I’m bullish on AI-driven automation in fund aggregation and investment decisions. We’re building AI-DeFi integrations ourselves. A major bottleneck in traditional DeFi is that 100% of logic must reside in smart contracts, limiting scalability. That’s why since 2019, core DeFi protocols remain unchanged—AMMs, DEXs, lending, and stablecoins. New contract-based models haven’t achieved Uniswap-level dominance in AMM/DEX. DeFi’s evolution has been slow across cycles, primarily because contracts must be fully on-chain, incurring high audit and operational costs. Integrating AI agents could transform development—future on-chain logic might only cover 10–20%, with 80% handled off-chain by AI.
Early use cases are emerging. As AI reasoning improves, its flexibility and extensibility surpass traditional trading systems. Recently, we’ve seen AI-agent integration with stablecoins as the most natural application—AI agents paying for services via stablecoins instead of traditional media offers greater coherence. Another exciting angle: AI agents forming closed-loop systems. Revenue and expenses occur entirely on-chain, fully automated without human intervention. As agents grow more autonomous, integrating on-chain revenue collection and payment frameworks becomes essential—making stablecoins and DeFi the most natural partners for agent ecosystems.
Alex: Understood. You mentioned DeFi-AI integration—a topic I’m also keen on. You noted that DeFi built on smart contracts evolved slowly because most logic executes on-chain. You added that future systems might shift 80% of logic to AI, leaving 20% on-chain. What exactly would this 80% entail?
Mindao: Let me illustrate. We’re building a cross-chain yield aggregator—a product impossible with traditional DeFi. First, cross-chain operations involve information synchronization, preventing atomic transactions. While protocols like LayerZero enable bridging, managing funds across chains and protocols introduces complexity too great for any single contract deployed across chains. Traditional DeFi can’t handle this. But with AI Agents, traditional DeFi handles only user deposits/withdrawals and on-chain strategy execution—specifically, cross-chain routing and protocol allocation. Essentially, on-chain logic ensures verifiable, transparent settlement—the cryptographic layer. But intermediate logic—e.g., withdrawing from Aave on Ethereum, bridging to Base, depositing into Morpho, then leveraging and looping within Morpho—involves dynamic decision-making too fluid for static contracts.
A core issue in traditional DeFi is business logic rigidity. Platforms like Binance iterate products in days or weeks—new features deploy immediately, reflecting centralized service agility. But DeFi requires all logic in immutable on-chain contracts. Look at Uniswap, Aave, MakerDAO—foundational DeFi protocols update every two to three years. This lag exists because DeFi logic is static, while markets evolve dynamically—strategies change daily. AI Agents excel at extending dynamic logic. Traditional AI required exhaustive rule enumeration. Modern inference models don’t need explicit rules. For example, an agent understands gas cost distribution over days and APY impact without manual rule-setting. Thus, beyond on-chain settlement logic, most operations can be delegated to Agents. Unlike Binance, which requires constant team intervention, DeFi Agents could autonomously optimize simple operations without human input. Next, this extends to yield aggregation, lending, and swapping—gradually replacing all with Agent-driven systems. This cycle, Ethena blurs lines—it’s hard to call it pure DeFi, as all capital operates within exchange-controlled environments running arbitrage strategies. But going forward, underlying DeFi protocols will undergo AI transformation, gradually replaced by Agents. This trend is unmistakable. Next-cycle DeFi products won’t resemble today’s fully on-chain-contract-based models.
Alex: Understood. For next-gen DeFi products integrating AI modules—where AI handles 80% of execution logic—how can we ensure verifiability and determinism of AI outputs?
Mindao: This is currently the biggest challenge. AI models still suffer from hallucinations. For example, given identical inputs—funds, data, APIs—most AI models return different strategies across 10 identical queries. Clearly, issues persist. But compared to three months ago, progress is rapid—hallucination rates have dropped sharply. Strategies now pass cross-validation checks. While consistency remains imperfect—100 identical requests won’t yield 100 identical results—the outputs are generally reliable, solving basic arithmetic correctly. Mitigation involves breaking agent workflows into finer tasks, reducing hallucination frequency. As inference models improve, this concern diminishes. Within a year, models like Grok 3.5—based on first principles or physics-inspired reasoning—may resolve many hallucination issues. Eventually, a single prompt could yield consistently high-quality outputs. Model improvement addresses this core challenge.
We observe many DeFi projects developing MCPs (Model Control Protocols). MCPs inherently decompose workflows. Future specialized MCPs may emerge—e.g., one dedicated to lending arbitrage strategies, delivering executable, verified outputs. The exciting prospect: while traditional DeFi emphasized modularity—composing AMMs from Aave or Uniswap code—AI-DeFi development may spawn professional MCPs. Each MCP acts as a module; combining them unlocks novel functionalities. Modularity evolves rapidly at the MCP layer. New risks arise—e.g., poisoned training data in AI environments. But relative to scale, these are manageable. DeFi security, after years of maturation, has proven resilient.
In fact, AI may simplify many DeFi security challenges. For example, traditional financial institutions hesitate to enter DeFi because multi-million-dollar audits still can’t eliminate risk—unlike banks, where losing all funds is unacceptable. In DeFi, hacks often drain entire protocol pools. But if future DeFi logic is AI-managed, traditional Web2 risk controls—like withdrawal limits—can be embedded via agent logic. This enables finer-grained security gradation. Traditional DeFi can’t implement granular controls due to gas costs and new bug vectors. AI integration makes this feasible.
Alex: Understood. One final question, stemming from our discussion. Previously, extensive code required auditing. Based on your observation, with AI advancement, is the security cost for a DeFi project increasing or decreasing? Has there been a noticeable decline in security spending over the past one to two years?
Mindao: Cost reduction depends on several factors. A major DeFi problem is that audits can’t cover all possibilities—neither code audits nor formal verification can exhaust all edge cases. That’s why traditional DeFi avoids complex products—one flaw triggers cascading failures. Comparing traditional DeFi to AI-enhanced DeFi, the latter offers more controllable audit costs and security boundaries. I lack exact figures—it’s too early—but my sense is that DeFi-AI products achieve richer functionality with lower audit overhead. Their capabilities and operational logic differ fundamentally.
Alex: Understood. We’ve covered a lot today—from stablecoins to their intersection with DeFi, AI, and the convergence of AI and stablecoins. Thank you, Professor Mindao, for joining us and sharing such profound insights and analysis. Appreciate it.
Mindao: Thank you.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
Mint Ventures
