Governance Innovation of On-Chain Credit Infrastructure: Industry Practices from DAO to Curator Model

Introduction: When DeFi Meets a "Midlife Crisis"

Since the second half of 2024, a noteworthy phenomenon has emerged in the DeFi lending sector: several leading protocols have simultaneously begun adjusting their governance structures, shifting from pure DAO governance to more flexible hybrid models. Behind this change lies a collective exploration by the entire industry in the face of institutionalization, regulatory pressure, and efficiency bottlenecks. Morpho's MetaMorpho curator system launched in 2024, Aave's attempt at an institution-only marketplace, and Compound's discussions surrounding Treasury all point in the same direction: the traditional "one person, one vote" governance model is revealing its limitations in the face of rapidly changing market demands. This article attempts to answer a core question by comparing and analyzing the strategic adjustments of several major protocols in 2024-2025: How should DeFi lending protocols find a balance between decentralized ideals and commercial efficiency? We will use Gearbox Protocol's Permissionless model launched in March 2025 as a primary case study and conduct a horizontal comparison with the practices of protocols such as Morpho, Aave, and Compound to evaluate the differences in performance of different governance models in actual operation.

DAO-managed "chronic disease"

Over the past year, the on-chain lending market has experienced a profound identity crisis. The narrative from 2020 to 2022 was "the democratization of permissionless finance," but by 2024, when traditional financial giants like Coinbase and BlackRock truly began deploying on-chain lending, the entire industry suddenly found itself no longer facing the challenge of "how to attract retail investors," but rather "how to serve institutional clients while adhering to the original principles of decentralization." The most direct manifestation of this predicament is the collapse of governance efficiency. Take Compound as an example: a proposal in 2024 regarding adding new collateral types took over eight weeks from discussion to final approval, during which time the market window had already closed. While Aave improved this somewhat through a fast-track approval process (Snapshot voting), the fundamental problem remained unresolved: when a protocol needs to serve 50 user groups with different risk appetites, the DAO's "all-encompassing" logic inevitably leads to "slowness."

Another underestimated challenge is regulatory compatibility. The EU's MiCA regulations and the US SEC's enforcement actions have made the entire industry realize that "pure decentralization" is neither a shield for regulatory exemptions nor conducive to integration with traditional financial institutions. Institutional funds require clearly defined responsible parties, clear risk control processes, and auditable decision-making records—precisely the weaknesses of traditional DAO governance. Against this backdrop, a wave of "governance model reconstruction" emerged in 2024. Morpho pioneered MetaMorpho, allowing professional teams to create vaults and manage risk parameters autonomously; Aave launched the Aave Arc program, providing institutional clients with a dedicated KYC marketplace; Euler launched EVC (Ethereum Vault Connector) to achieve more modular risk isolation. The common thread among these attempts is: while maintaining decentralization at the protocol layer, delegating operational decision-making power for specific markets to professional teams.

A "Species Atlas" of Four Governance Models

Compound: A staunch conservative, Compound still adheres to complete on-chain governance, requiring a full process of proposals, voting, and time locks for any parameter adjustments. The advantages of this model are high transparency and strong community participation, but the cost is a long decision-making cycle (typically 2-4 weeks) and difficulty in quickly responding to market changes. In 2024, the total number of Compound markets only increased from 32 to 41, significantly lagging behind its competitors. However, to be fair, Compound's positioning is "stability first," and its three-year record of zero bad debts proves the risk control advantages of this model. In an industry that generally pursues speed, Compound is like an old craftsman who insists on handcrafted brewing—slow, but with reliable quality.

Morpho: A Radical Decentralizer. Morpho's MetaMorpho system is perhaps the most radical attempt at decentralization to date. The protocol itself is only responsible for the underlying lending logic and liquidation mechanism; the creation of specific markets, asset selection, and risk parameters are entirely determined by the curators. This has led to an astonishing rate of expansion: in 2024, Morpho added over 200 markets, and TVL grew from $800 million to $2.8 billion (a 250% increase). However, problems are also evident. As of May 2025, Morpho's top 10 curators controlled over 65% of the TVL, and the risk of centralization cannot be ignored. More worryingly, some curators, in pursuit of higher returns, have begun accepting collateral with poor liquidity, and potential risks are accumulating. Morpho is like handing the steering wheel to the passenger, only responsible for applying the brakes—provided the braking system is sensitive enough.

Aave: A Pedestrian in the Dilemma. Aave has adopted a relatively moderate approach: the main pool remains strictly managed by the DAO, but a customized marketplace is offered to institutional clients through Aave Arc. The advantage of this "dual-track" system is that it retains the decentralized foundation while flexibly serving institutional needs. However, operational complexity has increased significantly; Aave's development costs increased by 40% year-on-year in 2024, and the Arc market growth was far below expectations (TVL accounted for only 8% of the total). A key reason is that institutional clients found that while Arc's Know Your Customer (KYC) requirements met compliance needs, they did not offer a significant advantage in interest rates, resulting in insufficient appeal. Aave is like a rider trying to ride two horses at once—theoretically feasible, but extremely exhausting in practice.

Compound V3: A Tinkerer's Attempt. Compound V3 attempts to improve upon the traditional DAO framework by introducing a "fast-track proposal" mechanism: for low-risk parameter adjustments (such as interest rate fine-tuning), a simplified process can be implemented, reducing approval time from 4 weeks to 1 week. This alleviates efficiency issues to some extent, but it essentially remains within the DAO framework, and 2024 practice demonstrated that its market expansion speed is still limited. A comparison reveals a growing consensus in the industry: pure DAO governance is ill-suited to the current market pace, but completely abandoning decentralization will create new problems. The key is finding appropriate "decentralization boundaries"—which decisions must remain at the DAO level (such as core protocol parameters and economic models), and which can be delegated to specialized teams (such as risk management in specific markets).

Gearbox's "Fast and Furious"

Gearbox's Permissionless model, launched in March 2025, can be seen as a variation of the Morpho Curator model, but with added constraints. The core logic is: the protocol layer provides standardized lending and leverage infrastructure, curators are responsible for creating specific markets and managing risk parameters, but curators must bear the first loss themselves (through a staking mechanism), and their operations are subject to on-chain transparency. Data shows that this model achieved significant initial success. From March to August 2025, Gearbox's TVL grew from $105 million to $329 million, a growth rate of 213%. Particularly noteworthy is the Lido dedicated pool, the first large-scale application of the Permissionless model, whose TVL grew from $72 million to $296 million. This growth rate is indeed impressive in the current market environment—for comparison, Aave's overall TVL growth rate during the same period was 45%, Compound's was 31%, and Morpho's was 89%.

What's even more interesting is the comparison of expansion speed. Gearbox deployed 42 new markets across 5 chains in 3 months, compared to only 41 markets in the entire year of 2024. This acceleration is driven by the improved decision-making efficiency brought about by the curator model: no longer needing to go through the DAO proposal process for each market, curators can act quickly based on market opportunities. In contrast, Compound's average launch cycle for each new market in 2024 was 3.2 weeks, Aave's was 2.1 weeks, while Gearbox shortened this cycle to an average of 5 days through the Permissionless model. However, rapid expansion has also raised questions. Gearbox currently has "activated" deployment capabilities on 28 EVM chains, but actually only operates on 9 chains. While this "wide net" strategy reduces operating costs, it also means that a large amount of liquidity is dispersed. Taking the Plasma pool as an example, although the TVL reaches $80 million, it is distributed across 4 different chains, and the actual liquidity depth on a single chain is not high. This is similar to the problem faced by Morpho: excessive market fragmentation can weaken capital efficiency.

Another noteworthy metric is the quality of the curators. Gearbox currently collaborates with five curators, including Invariant Group, Re7, and Maven11, managing a combined asset size of over $1.5 billion, with four of them ranking among the top 15 DeFi curators. This certainly reflects a certain level of institutionalization, but a comparison with Morpho reveals a gap: Morpho's curator ecosystem includes over 30 professional institutions such as Gauntlet, Steakhouse Financial, and Block Analitica, managing a total asset size exceeding $5 billion. The breadth and depth of the curator ecosystem directly determine the model's resilience. Risk testing is another crucial dimension. Gearbox officially emphasizes that it achieved zero bad debts during the extreme market volatility of October 10, 2024, which is indeed a positive sign. However, it should be noted that the extreme nature of this test was relatively limited (ETH experienced a 12% single-day drop). A true stress test should refer to the Terra crash in May 2022 or the March 12, 2020 black swan event. For comparison, Aave experienced approximately $5.3 million in bad debt during the March 12, 2020 event, but this was covered by insurance and community funds; Compound experienced approximately $8.6 million in liquidation delays, but ultimately did not result in actual losses. Morpho, due to its later launch, has not yet undergone a true systemic risk test, which is one reason why the market remains cautious about its rapid expansion.

The strategic transformation of "fighting with our backs to the wall"

In August 2025, Gearbox DAO passed proposal GIP-264, deciding to completely abandon the traditional DAO governance pool and fully transition to a permissionless model. This decision sparked intense debate within the community, as it signified a complete departure from the protocol's earlier concept of "community-driven decision-making for specific markets." From a business perspective, this choice is understandable: the permissionless pool already accounted for over 70% of the total TVL, and its growth rate far exceeded that of the DAO pool; maintaining two systems was clearly not cost-effective. The DAO pool's TVL only grew from 82 million to 98 million in 2024 (a 19% increase), while the permissionless pool grew by 180% during the same period. More importantly, the operating costs of the DAO pool (including the time and manpower costs of the governance process) were actually higher. This is akin to a company finding its traditional business unit not only growing slowly but also continuously consuming resources, ultimately deciding to make a drastic decision.

However, the objections raised by the opponents are not without merit. First, is a 70% TVL (TVL) percentage sufficient to represent "users voting with their feet"? It's worth considering that the Permissionless pool enjoys preferential resources from the protocol in marketing and curator relationships. Second, will completely abandoning the DAO pool cause the protocol to lose its "decentralized" narrative foundation? In the current regulatory environment, this might actually be a disadvantage. Uniswap has attracted SEC attention due to its overly centralized front-end control, while Compound's pure DAO governance has become a point of defense. From an industry comparison perspective, Gearbox's choice is quite radical. Aave opted for a dual-track approach, Compound still insists on DAO dominance, and even Morpho retains DAO governance at the protocol layer (although the market layer is completely delegated to curators). Gearbox is the first mainstream protocol to completely "bet" on the curator model. This could be a turning point for the industry, or it could be a costly experiment. The design of the migration process is noteworthy. To ensure a smooth transition, Gearbox deployed two new institutional curators (Maven11 and KPK) to take over funding from the original DAO pool. However, a subtle issue arises: are users being "guided" to migrate or migrating "voluntarily"? If support and optimization of the DAO pool gradually cease, users are essentially passively accepting the new model. This "soft coercion" could sow the seeds of future problems—if issues arise with the new model, the community might feel deprived of their right to choose.

The Curator Economy: A Two-Sided Mirror

The rise of the curator model is essentially an inevitable result of the "specialization" in the DeFi field. However, we need to be soberly aware that this model is far from perfect. The bright side is obvious: professional teams can provide more refined risk management, faster market response, and more user-friendly interfaces for institutional connections. Morpho's data shows that markets managed by its curators have an average capital utilization rate 23% higher than those governed by DAOs. Gearbox claims that its curators can complete the deployment of new markets within 5 days, while the traditional process takes 3-4 weeks. These efficiency improvements are real, just as a professional chef can indeed make more refined dishes than a home kitchen.

However, the darker side cannot be ignored. First, there's the risk of centralization. In both Morpho and Gearbox, the top 5 curators control over 50% of the total value (TVL). If these curators make a systemic misjudgment (like the collapses of 3AC and Celsius in 2022), the entire protocol will be severely impacted. A more insidious risk is potential collusion among curators—although theoretically each market is independent, if several major curators use similar risk control models, systemic risk can be amplified. Second, there's the ambiguity of liability. When bad debts occur, should the curators or the protocol bear the burden? Morpho's solution involves curators staking to cover the initial loss, but the staking ratio is typically only 5-10% of the market size, meaning large bad debts will still be passed on to the protocol. Gearbox's mechanism is similar and hasn't undergone true stress testing. In contrast, while Aave is slower, its liability boundaries are clear: the protocol bears all risks, covered by an insurance fund.

Third is information asymmetry. Although all operations are on-chain, it's difficult for ordinary users to assess the curator's risk exposure in real time. Morpho has seen cases where curators quietly increased the proportion of risky assets, only to be discovered after someone exposed it on a forum. Gearbox currently relies on third-party dashboards like Dune for transparency, but this is far inferior to the protocol's native risk monitoring system. Fourth is the lack of an exit mechanism. Theoretically, if users are dissatisfied with the curator's decisions, they can withdraw their funds and move them to another market. However, in practice, withdrawals often require a waiting period (due to liquidity management needs), and switching markets incurs gas fees and potential slippage losses. This "stickiness" actually makes curators lack sufficient external constraints, making them resemble a monopoly without competitive pressure.

The "bloated" trap of multi-chain expansion

Both Gearbox and Morpho heavily promote their multi-chain deployment capabilities, but this strategy warrants closer examination. Gearbox claims to be "activated" on 28 EVM chains, but actually operates on 9. Morpho operates on 12 chains. In contrast, Aave focuses on only 6 mainstream chains, and Compound only has the Ethereum mainnet and a few L2 chains. Which strategy is superior? The logic behind supporting multiple chains is to capture more incremental markets, especially early opportunities in emerging L2 chains. Gearbox's pools on Plasma and Etherlink have indeed capitalized on the early liquidity gaps of these chains, achieving a TVL of $80 million and $17 million respectively. This is successful from the perspective of a single chain, but problems emerge when viewed at the protocol level as a whole.

First, liquidity fragmentation severely weakens capital efficiency. When the same asset type is distributed across 10 chains, the depth of a single chain is insufficient to support large-scale lending, necessitating higher interest rates to compensate for liquidity risk. Aave's data shows that its average utilization rate for USDC on the Ethereum mainnet is 75%, while on some smaller L2 chains, the utilization rate of pools with the same parameters is only 40%. Second, cross-chain risks are systematically underestimated. The collapse of Multichain in 2024 resulted in losses exceeding $120 million for multiple protocols, most of which were multi-chain protocols. Each additional chain adds reliance on cross-chain bridges, which remain the most vulnerable link in the entire crypto space. Although Gearbox claims to use "official bridges," official bridges also cannot guarantee absolute security—the Ronin Bridge is a stark example. Third, there is a hidden increase in operating costs. Although Gearbox says "activation" is not the same as "operation," each chain requires oracles, liquidation bots, and monitoring systems. More importantly, there is technical debt: the compatibility of different EVMs is not 100%, and when contracts need to be upgraded, the coordination cost of 28 chains is much higher than that of 6 chains.

In contrast, Aave's "selective mainchain" strategy, while seemingly conservative, offers superior capital efficiency and security. Its Ethereum mainnet single-chain TVL exceeds $12 billion, far surpassing the single-chain scale of any multi-chain protocol, implying greater liquidity depth and lower lending costs. Compound's strategy is even more extreme: focusing solely on Ethereum, but maximizing security and stability, resulting in three years of zero bad debts and making it the preferred choice for institutional investors. Multi-chain expansion is essentially an "efficiency illusion": superficially, the market quantity and chain coverage appear extensive, but the true moat lies not in breadth, but in depth. The network effect of DeFi is the network effect of liquidity; decentralization weakens this effect. Like a chain restaurant, opening 100 mediocre branches is less competitive than concentrating resources on 10 boutique stores.

Institutionalized "Faustian trade"

The overarching context of this governance model revolution is the "institutionalization" wave in DeFi. However, we need to question: is institutionalization truly the future of DeFi? Supporters have a clear logic: institutional funds, with their large size, high stability, and professional risk control, are essential for DeFi to scale up. Data supports this: in 2024, institutional funds accounted for 29% of DeFi's total TVL, up from 12%, contributing the vast majority of the incremental growth. Gearbox's ability to attract institutional curators like Maven11 and KPK, and Morpho's collaboration with top risk control companies like Gauntlet, have indeed enhanced the professionalism of the protocols. But opponents' concerns are equally valid: wasn't the original intention of DeFi to "remove intermediaries"? When we introduce institutional curators, KYC access, and professional risk control teams, how much difference remains between DeFi and traditional finance? A more realistic risk is that institutionalization might turn DeFi into a "licensed club," systematically excluding small teams and individual users.

A worrying trend is the "oligopolization" of the curator ecosystem. Of Morpho's top 15 curators, 9 also provide risk consulting for protocols like Aave and Compound. What does this mean? It means that risk decisions for the entire DeFi lending market are effectively controlled by fewer than 20 institutions. Once these institutions adopt similar models and assumptions (which they likely already are), systemic risk will be greatly amplified. Wasn't the lesson of the 2008 financial crisis that "everyone uses the same risk control model"? Another overlooked issue is the disappearance of regulatory arbitrage. A major advantage of early DeFi was the flexibility afforded by regulatory gray areas, but as protocols actively embrace institutionalization, introduce KYC, and establish clearly defined responsible parties, they are essentially relinquishing this advantage. The EU's MiCA and the potential US stablecoin legislation both explicitly require licensed institutions to comply with traditional financial rules such as capital adequacy ratios and liquidity coverage ratios. If DeFi protocols ultimately also have to comply with these rules, will their cost structure become similar to traditional finance, thus losing their competitive advantage?

The more fundamental philosophical question is: what kind of financial system do we really want? If the answer is "more efficient traditional finance," then institutionalization is fine. But if the answer is "a truly decentralized, censorship-resistant, permissionless financial system," then the current direction is questionable. Uniswap's success precisely demonstrates that minimalist, truly decentralized protocols can also achieve huge market share (TVL exceeding $4 billion, and completely without curators or governance). This is like asking: do we want a faster horse-drawn carriage or a car? If DeFi is merely moving traditional finance onto the blockchain, what's the point?

There are no perfect answers, only the art of weighing options.

By comparing and analyzing Gearbox, Morpho, Aave, and Compound, we can draw several conclusions. First, traditional DAO governance does indeed have efficiency bottlenecks, but its "slowness" isn't necessarily a bad thing. Compound's three-year record of zero bad debts proves that caution and transparency have irreplaceable value in risk management. Speed isn't always good, especially when dealing with other people's funds. Second, the curator model improves efficiency, but at the cost of centralized risk and ambiguous responsibility. The rapid expansion of Morpho and Gearbox is impressive, but it hasn't been tested over a full cycle. A key test will be: how will the curator and the protocol share the losses when large-scale bad debts first occur, and will the community accept it? Third, multi-chain expansion seems appealing, but it may actually be an efficiency trap. Capital efficiency and liquidity depth are more important than market size; Aave's curated strategy may be superior in the long run. Fourth, institutionalization is a double-edged sword. It brings professionalism and scale, but it may also cause DeFi to lose its fundamental advantages. The industry needs to find a balance between embracing institutional funds and maintaining decentralization, rather than pursuing the former one-sidedly.

Finally, Gearbox's full transition to a permissionless model is a bold experiment, and its success or failure will profoundly impact the future direction of the entire DeFi lending sector. If successful, we may see more protocols follow suit; if it fails, it will serve as a cautionary tale of "over-centralization." Regardless, this experiment deserves close attention from the entire industry. The future of DeFi will not be the victory of any one model, but rather the coexistence of multiple models in different scenarios. Some users need Compound's extreme security, some need Morpho's flexibility and efficiency, and some need Aave's robust balance. The protocol's task is not to find the "perfect model," but to clearly tell users: what trade-offs are you making, what risks are you taking, and what returns are you getting? Transparency and honesty are ultimately more important than any governance model.