Author: CoinFound
Takeaways
-
Core Insight One: Oligopoly Established, On-Chain Enters Asset Management Era. The Curator model represents a systemic concentration of discretionary management authority in on-chain asset management, marking the true starting point for crypto asset management. In just 6 months, the Top 4 concentration ratio in this sector has surged to 76.77%, far exceeding the 57% concentration of the US mutual fund industry, which took 20 years to form. The physical separation of "judgment rights, custody rights, execution rights" implemented at the smart contract layer is functionally isomorphic to traditional asset management, but because it is code-enforced, it builds a more thorough underlying infrastructure.
-
Core Insight Two: Fee Compression Drives Accelerated TradFi Entry, Pushing the Sector Toward Differentiation. The winning logic in the Vault sector has fundamentally shifted: compliance frameworks, off-chain structures, and role separation have become the sequential primary conditions for institutional capital access, while absolute returns (APY) have degraded to the last item in a multi-objective assessment. The invisible driving force behind this switch is the historical fee compression encountered by traditional asset managers. Based on this, this report anticipates that sector liquidity will further bifurcate into two main lines over the next 24 months: the "institutional compliance channel" and the "retail high-yield channel".
-
Core Insight Three: Vault Risk Is Forming. The four-link risk mechanism composed of "redemption mismatch, scaling, nesting contagion, concentration amplification" has historically, for the first time, simultaneously materialized at the current juncture. The Stream-Re7 incident in November 2025 constitutes a partial empirical case of this transmission path. The contagion mechanism for systemic shocks is no longer a theoretical assumption but an existing structural exposure.
-
Tracking Framework: Anchored to Quantitative Falsification Thresholds for Tail Risks and Extreme Scenarios. This report ties subjective judgments to quantitative thresholds: Top 4 concentration ratio exceeding 80%, a single Vault's exposure to non-immediately realizable assets exceeding 40%, nesting depth ≥3 layers coupled with top Curators holding each other as three extreme red lines. The report deduces five evolution scenarios covering the next 12-24 months (Baseline / Acceleration / Inflection / Reversal / Falsification). Triggering any threshold will force the report back to the evidence layer for falsification and reassessment.
-
Cycle Characterization: Regulatory Vacuum Game and "De Facto Takeover". The current moment for the sector can be considered "de facto takeover preceding de jure recognition." During the global regulatory vacuum where no specialized licensing framework for Curators has been introduced, institutional capital has substantively flowed into leading Curators along the compliance channel.
Chapter 1 Vault Market Overview and Conceptual Boundaries
Source: CoinFound, DeFiLlama, Morpho, data.morpho, Blockworks Research, BlackRock, Superstate; Data as of May 26, 2026*
By mid-2026, Vaults have become the largest structure by capital volume within DeFi. On Morpho, a single financial primitive alone, the size of the curated vault subset actively managed by Curators approached $6 to $8 billion; the protocol's TVL once touched $11.78 billion in mid-May, placing it at the forefront of the DeFi lending sector.
However, the term "vault" commonly used in the market does not refer to the same product: Maker/Sky's vault is essentially a Collateralized Debt Position (CDP), where users deposit collateral to generate stablecoins; Yearn's yVault is closer to a yield strategy account, where a strategy contract allocates funds to various DeFi yield opportunities; Mellow's vault primarily serves restaking scenarios, encapsulating yields and slashing risks from protocols like EigenLayer and Symbiotic; products like Superstate USTB are closer to tokenized fund shares, representing users' on-chain ownership of short-duration US Treasury fund units.
Therefore, the vaults discussed in this article do not cover all products labeled as vaults, but focus solely on one branch: on-chain asset management structures actively managed by independent Curators, operating on lending or yield-generating financial primitives, and typically encapsulated using the ERC-4626 standard. In other words, this article focuses on the asset allocation model represented by Curator vaults like those on Morpho and Euler, and does not discuss Maker/Sky's collateral debt vaults, Mellow's restaking vaults, or tokenized fund shares like Superstate USTB.
1.1 A New Protocol Stack Product: Curator as the Middle Layer
The operation of this type of vault depends on a three-layer stacked protocol stack (details in Figure 2).
Source:CoinFound,DefiLlama[2]、Morpho App Steakhouse USDC vault[6]、Vaults.fyi、Smokehouse USDC、Yearn Docs、Morpho official GraphQL API,2026-05-26
The bottom layer is the primitive layer, consisting of lending and yield protocols like Morpho Blue and Euler v2, providing immutable, composable financial primitives. The middle is the Curator layer, consisting of strategy managers like Steakhouse, Gauntlet, Sentora, Re7, and Block Analitica, a new middle layer that only formed in 2024, and the central focus of this report's analysis. The top layer is the application layer, the end-user interface, including Coinbase Earn, Bitwise's on-chain vaults, Kraken DeFi Earn, Apollo × Morpho. The relationship between the three layers is vertical: the application layer interfaces with users and invokes strategies from the Curator layer; the Curator layer makes active management decisions and deploys funds to the primitive layer; the primitive layer only handles underlying lending and liquidation logic.
Source:CoinFound,DefiLlama[2]、Morpho App Steakhouse USDC vault[6]、Vaults.fyi、Smokehouse USDC、Yearn Docs、Morpho official GraphQL API,2026-05-26
Therefore, this report will also adopt the "protocol stack" as the primary analytical framework. An example illustrates this choice. The Steakhouse USDC vault on Morpho and the Yearn v3 USDC vault are similar from an asset dimension—both are on-chain yield products for USDC—but their power structures are entirely different. The former distributes judgment rights (selecting primitives, setting parameters, rebalancing), custody rights (funds locked in the Morpho Blue contract, Curator has no withdraw right), and execution rights (automatic liquidation and rebalancing executed by contracts, Curator only issues instructions) to three independent contracts. The latter's strategy contract simultaneously holds all three powers, with strategy and governance coupled within the Yearn protocol[^6]. This difference is invisible from an asset dimension but becomes clear in the protocol stack dimension. This is also fundamental to gaining a deep insight into the Vault sector.
1.2 Four Characteristics of the Current Vault Market: Scale, Concentration, Assetization, Institutionalization
From the perspective of 2026, the Vault niche exhibits several significant characteristics:
The first is scale and pace. Morpho's protocol TVL increased about 15-fold over the past 18 months, from less than $1 billion in early 2024 to a historical peak of $11.78 billion in mid-May 2026, retreating to $7.627 billion as of the drafting date (details in Figure 3). This retracement was mainly caused by an industry-wide redemption wave following the Stream Finance xUSD depegging event in November 2025, with DeFi seeing a net outflow of about $10 billion in a single week. What's more notable is that the pace of this 15-fold growth was clearly driven by a series of TradFi integration events; each significant acceleration since Q4 2025 closely followed a new announcement of a traditional asset manager's entry.
Source: CoinFound, DefiLlama Morpho protocol page[2]、arXiv:2512.11976[1]、Various project official announcements. The curated vault subset accounts for approximately 60–80% of the protocol TVL (based on the paper's methodology combining the top eight curators, 2025-11 ≈ $72.7B).
The second is concentration. As of the drafting date, among the 53 curators under DefiLlama's "Risk Curators" category, the Top 4 collectively account for approximately 76.77% of this category's total[^1]; six months ago, the same measurement from a similar paper was only 64.6%. That is, the capital concentration in the curated vault market increased from about 65% to about 77% within 6 months. For comparison, the Top 4 in the US mutual fund industry (Vanguard, BlackRock, Fidelity, Capital Group) collectively account for about 57%. An industry that has developed for decades, repeatedly regulated and consolidated, sees its concentration surpassed by about 20 percentage points by an on-chain sector that rose within 3 years. This speed of concentration increase is a pattern almost never seen in traditional asset management.
The third is the changing composition of underlying assets. Stablecoin vaults still dominate, at about 65%; but the shares of Liquid Staking Tokens (LSTs) and RWAs have risen from marginal categories two years ago to about 20% and 8% respectively, with the trend continuing. This point is merely recorded as a current state in this chapter; its real implications will unfold in the risk mechanism discussion: stablecoins are instantly redeemable assets, LSTs have unbonding periods, RWAs have T+N or longer redemption windows—the shift in asset composition from "all instant" to "containing a significant proportion of non-immediate redemption" means a structural time gap is accumulating between the promise side (users can redeem anytime) and the fulfillment side (how quickly underlying assets can actually be realized).
The fourth is the intensive entry of institutions. From BlackRock launching BUIDL in March 2024 to Invesco taking over Superstate's USTB treasury fund in March 2026, 8 out of 10 representative TradFi entry events occurred within the 7-month window starting September 2025[^3].
Among these four characteristics, scale and institutionalization are mutually causal: TradFi capital inflow pushes up TVL, and the TVL volume makes the Curator sector more attractive for further TradFi entry. Concentration and assetization push the market's risk structure to a new position from two directions: the former amplifies the transmission radius of a single-point impact, while the latter introduces redemption mismatches more complex than stablecoins. In other words, today's vault market has evolved into a new structure: "crypto-native + institutional capital, stablecoins + LSTs + RWAs, concentrated in a few leading Curators." Every element of this new structure corresponds to a new dimension not yet widely recognized by the market.
Chapter 2 Curator Is the New Starting Point for Crypto Asset Management
The Curator role was not born from meticulous design but evolved. Returning to the state of DeFi's first-generation yield products around 2020, when the protocol layer and fund allocation layer were mixed, and governance was left to DAO-wide voting, governance apathy, slow decision-making, and unaccountable risk were inevitable outcomes. The Curator may seem like another layer of "middleman," but it is a role restructuring forced by the collapse of DeFi's first-generation asset management structure at the governance level.
2.1 From Strategy Aggregation to Curator: Evolution and Cost
Source: CoinFound
The fundamental reason for the emergence of Curators is that the market ultimately realized DeFi cannot replace subjective judgment. For the past decade, using DAO governance to handle such subjective judgments resulted in participation rates consistently below 10%, with Aave's historical voting participation long hovering at 2-3%. The "Golden Boys" governance attack on Compound in July 2024 and the Aave governance crisis in December 2025 both exposed the same structural problem: participation rates are too low to support the legitimacy of subjective judgments[^7]. In this state, the Curator model directly acknowledges the traditional governance failure, explicitly handing judgment rights to professional teams. The core user bet is no longer on the protocol but on the manager.
2.2 Separation of Powers in Curators: Isolation of Judgment, Custody, and Execution Rights
The true core of the Curator model lies in dividing the intermediary's power into three parts (details in Figure 5).
Source: CoinFound
This mutual-check relationship is akin to that of traditional asset management's "manager + custodian + auditor." But because custody and execution are in code, the separation is more thorough than in traditional asset management. A traditional custodian could theoretically collude with the manager to misappropriate funds (Madoff case), but the Morpho Blue contract will not "collaborate" with any Curator to misappropriate[^8]. The role played by the Curator here corresponds exactly to the manager in traditional asset management, albeit with a different legal identity (on-chain mandate vs. LP contract).
2.3 Top 4 Account for 77%, Concentration Exceeds US Mutual Funds
And the capital concentration in the Curator sector has already surpassed that of the US mutual fund industry. As of the drafting date, among the 53 Curators under DefiLlama's "Risk Curators" category, the Top 4 collectively account for approximately 76.77% of this category's total; six months ago, the same measurement was 64.6%, meaning concentration rose from ~65% to ~77% within 6 months[^1].
For comparison, the Top 4 in the US mutual fund industry collectively account for about 57%. An industry that has developed for decades, repeatedly regulated, and undergone multiple rounds of concentration adjustments, sees its concentration surpassed by about 20 percentage points by an on-chain sector that rose within 3 years (details in Figure 6). In fact, mutual fund concentration is also rising; in 2005, the Top 5 were only about 35%, reaching about 63% by 2024, but it took nearly 20 years[^9]. The Curator sector completed the same degree of concentration in just 6 months.
Source: CoinFound
2.4 Leading Platforms and Curators Are Differentiating at Both Ends
From a broader market perspective, curated vaults run on four leading platforms: Morpho, Kamino, Veda, Pendle (not a typical Curator model, but its PT/YT split provides fixed-income tools for Curators). Morpho alone carries the vast majority of curated vault volume and is the central focus of this report's analysis.
Source: CoinFound
The seven leading Curators show clear differentiation in strategy style and institutional background (details in Figure 7).
Source: CoinFound
The seven curators operate across two different competitive dimensions: strategy-focused ones (Steakhouse / Gauntlet / Re7 / MEV Capital / Apostro) continue to compete on APY and risk control; institution-channel-focused ones (Block Analitica / Sentora / part of Gauntlet's business) begin to compete on compliance integration and off-chain structure. The funding sources, fee levels, and risk appetites of these two lines are diverging.
Chapter 3 DeFi Competition Standard Shifts: From APY to Compliance Channel
Over the past three years, this sector settled winners based on retail logic: whoever could generate more stable, risk-adjusted superior APY on public primitives would capture the market. However, since Q4 2025, institutional capital has entered with visible density, and the competitive framework has been replaced. Compliance integration, off-chain structure, and role separation now rank before APY. We believe this is the first fundamental shift in winning logic since the Curator sector formed.
3.1 From Retail to Institutional
The Past: Retail capital dominated, competition settled at the strategy layer.
Looking at the capital composition from 2023 to 2025, Curators' funding primarily came from crypto-native users. This money evaluated Curators based on straightforward standards: APY ranking, brand recognition, and user experience smoothness. This capital structure determined that the competition among Curators was settled at the strategy layer. Gauntlet served as Aave's risk advisor for four years, with fees once reaching $2 million annually[^31]. After leaving Aave for Morpho in February 2024, its modeling capabilities directly translated into vault performance; on October 30, 2025, its USDT vault absorbed $775 million in a single day, with TVL growing 40-fold without destabilizing[^32]. Steakhouse leveraged research capabilities to turn conservative stablecoin vaults into the sole curator position for Coinbase Earn. MEV Capital publicly states its rebalancing threshold is when an alternative venue can consistently provide ≥75bp extra yield[^33]; such parameterized strategy narratives themselves attract retail capital.
This logic has held true in retail scenarios and will continue to do so. However, starting Q4 2025, Curators began to see a completely different type of capital.
Present: Institutional capital enters, competition standards become fourfold.
Institutional capital evaluates Curators on a completely different logic. When institutions select Curators, they filter in the following order: First, can it fit into a compliance framework? That is KYC/AML, controllable whitelists, auditable on-chain records. Second, can the off-chain structure be integrated? That includes fund management agreements, custodian connections, periodic reporting systems. Third, is role separation clear? Who is responsible for investment decisions, risk control, compliance, and can they be independently accessed during due diligence. Only after passing these three criteria do they look at APY.
Applying these four criteria to the existing seven leading Curators immediately reveals stratification (details in Figure 10).
Source: CoinFound
We tend to believe that over the next 24 months, the Curator sector will bifurcate along this fault line: the institutional channel (Block Analitica, Sentora, Gauntlet's institutional business) and the retail channel (Steakhouse's pure on-chain business, Re7, MEV Capital, Apostro). The AUM growth rates, fee levels, and risk appetites of the two sides will diverge noticeably.
3.2 Root Cause of the Shift: TradFi Accelerated Entry Since 2025
Why did TradFi accelerate its entry so densely within the window from Q4 2025 to Q1 2026? The market offers three explanations.
The surface explanation is regulatory shifts, stablecoin legislation, and interest rate cycles. This is repeatedly discussed in mainstream media and won't be elaborated here. Academic papers often cite the real drivers as infrastructure land grabs, distribution channels, and settlement efficiency. These are valid, but they don't explain the specific timing question: why concentrated in Q4 2025 to Q1 2026, and not 2024 or 2027?
We believe the real invisible driver is fee pressure.
From January 2024 to March 2026, fees on the active management side of TradFi have been compressed to a new range by ETFs and passive products. Vanguard's and Schwab's short-term bond ETFs are already around 3bp, and BlackRock's iShares Short Treasury Bond ETF (SGOV) is at 9bp[^37].
In fact, active management products have not escaped this compression curve: Apollo Diversified Credit's retail interval fund (CRDIX) has total annual fees of 2.84%-4.51%, net 3.11%-4.11% after fee waiver[^38]; Franklin Templeton's on-chain money market fund FOBXX fee is only 20bp[^39]. Active products either become passive or sell at lower fees—neither is an exception.
In other words, the fee space TradFi asset managers can capture in their existing businesses has been compressed to a historically low level.
The fee structure on the DeFi vault side is entirely different. Curators' mainstream performance fee takes 8%-15% of annualized yield (mostly set at 15%); converted to an AUM-equivalent fee based on a 5%-7% net APY, it's 50bp-100bp[^40]; Morpho Vault V2's newly added management fee field has an upper limit of 500bp, still not yet topped out[^41].
There are two layers to this: First, in absolute numbers, the equivalent fee for DeFi vaults is 5-11 times that of SGOV and 2.5-5 times that of FOBXX. Second, users' perception mechanism of fees is different. Curator vault users see a number like 6% APY on the front end; the performance fee field requires clicking into the details page to be displayed[^42]. What users perceive is not a "management fee" but a "yield spread." For the same capital amount under similar compliance requirements, a TradFi asset manager can only charge 10bp in its existing business but can charge 80bp in a Curator vault. This is a fee space they can no longer find in their own business. In other words, the density of entry starting in Q4 2025 wasn't driven by any single event but by fee pressure reaching a tipping point where not entering meant continued passivization.
Source: CoinFound
As for why entry must be through Curators and not directly into Aave or Morpho pools, it's because DeFi-native pool architecture is incompatible with TradFi compliance frameworks on three levels.
Pool models share risk, compliance requires risk isolation (the same LP cannot share risks across different funds); pool models are passive management, compliance requires explicit mandates (funds have written investment strategies and boundaries); pool models have anonymous LPs, compliance requires KYC[^43].
The Curator model happens to solve all three constraints: MetaMorpho/Vault V2's isolated market architecture is designed for risk isolation[^44]; a Curator's public mandate is equivalent to a fund prospectus; permissioned vaults plus KYC solve the whitelist issue. The Curator model wasn't designed for TradFi, but it's currently the only interface TradFi can connect to.
In fact, the fit of this interface has been validated by three events: Apollo×Securitize deploying ACRED on Morpho (Polygon Compound Blue + Gauntlet as curator) on January 30, 2025[^45]; Coinbase Earn USDC integrating through Steakhouse on September 18, 2025, with cumulative borrowing reaching $2.17 billion by Q1 2026; Apollo-affiliated entities signing an agreement with the Morpho Association on February 13, 2026, to purchase up to 90 million MORPHO governance tokens (approx. 9%) over 4 years[^46]. A traditional asset manager directly holding governance rights at the protocol layer is the strongest signal of the interface thesis. Out of 12 representative TradFi entry events from BlackRock BUIDL in March 2024 to Invesco taking over USTB in March 2026, 9 occurred within the 7-month window starting September 2025 (details in Figure 8). This density cannot be coincidental.
Source: CoinFound
However, not all signals point to unilateral acceleration. To date, institution-channel-focused Curators have not been issued any specialized licenses by any jurisdiction, and there is no public statistical classification for the scale share of permissioned vaults[^47]. Nevertheless, we believe these observations do not undermine the main thesis direction; they only determine that the transition's form will be "de facto takeover preceding de jure recognition." The pace of regulatory framework establishment will noticeably lag behind the pace of capital arrival.
The shift has already occurred, and the winning framework has been rewritten. Yet within this new framework lies an unpriced risk: when institutional capital flows along compliance channels, concentrated in leading Curators, and repeatedly repackaged through vault nesting, the concentration of Curators itself becomes an amplifier for systemic shocks.
Chapter 4 Vault Four-Link Risk Chain: Risk Structure Already Formed
When redemption mismatch, nesting contagion, and judgment rights concentration occur simultaneously, risk itself transitions from a "possibility discussion" to a "structural discussion." We believe the current vault form is accumulating not just one new risk, but a chain of incidents composed of four links. Each link alone is not fatal; each corresponds to a major incident in DeFi history. However, all four links being in place simultaneously is a state that only first materialized after Q4 2025.
Source: CoinFound
4.1 The Risk Chain Consists of Four Links
The first link is redemption timing mismatch. Vaults promise users instant redemption, but whether underlying assets can be instantly realized depends on the assets' own settlement windows (details in Figure 11). Tokenized US Treasuries settle T+0 to T+1, private credit redeems monthly-quarterly (Apollo ACRED quarterly redemption cap 5%, parent fund only processed 45% of redemption requests in March 2026[^64]), real estate assets under Reg D have an initial 1-year lock-up, LSTs have 1 to 14-day unstaking periods[^65], and Curator rebalancing itself carries slippage on low-liquidity vaults. The time gap between the promise side and the fulfillment side is the prerequisite for any bank run event. This link alone is not fatal; the March 12, 2020 Maker Black Thursday and the ~7% stETH discount in June 2022 were events triggered by this single link, with impact limited to one protocol or one asset type.
Source: CoinFound
The second link is the scaling of mismatch. On-chain RWA scale grew from ~$15.2 billion at the end of 2024 to ~$35 billion in November 2025, expanding about 3.4x in 2025[^66]; LST's share within DeFi TVL rose from about 17.6% in Q4 2022 to about 51% in early 2026 over three years[^67]. These two curves turn "non-immediately redeemable assets" from a minority category within vaults into the systemic foundation of the vault industry. The consequence is that during the next bank run, the pressure surface will no longer be confined to one asset type or one protocol—a large swath of vaults will simultaneously face the hard mismatch between settlement windows and underlying redemption cycles.
The third link is the opening of nesting contagion paths. ERC-4626 makes vaults holding each other a standard action. Morpho Vault V2 formally supports V2 → V1 MetaMorpho double-layer nesting through the Adapter system, and Yearn V3 structurally allows further nesting[^68]. The change brought by this link is the real structural risk; risks accumulated in the first two links are no longer confined to a single vault but will transmit upward along the nesting chain. A user deposits into a vault A that appears to hold only stablecoins, but the funds may actually hold non-immediately redeemable assets indirectly through vault B; vault A's front end only shows its own risk score, not vault B's redemption window. For reference, the collapses of Iron Finance, Wonderland, and UST/Anchor from 2020 to 2021 relied on nested structures to transmit; that generation's nesting was based on yield farming token loops—different form, same contagion mechanism.
The fourth link is judgment rights concentration. As of the drafting date, the Top 4 Curators control about 77% of the curated vault market's capital (see section 2.3), with Steakhouse alone accounting for about 28% of Morpho's total. Under the three-power separation framework argued in section 2.2, judgment rights reside with the Curator. When the first three links are all in place, the fourth link determines the upper limit of incident scale; any parameter error, rebalancing error, or risk control threshold error by a leading Curator will amplify from "one vault damaged" to "a batch of vaults under simultaneous pressure."
These four links combined imply an already-existing causal chain: Mismatch → Scaling → Nesting Contagion → Concentration Amplification.
All that remains is an external trigger: a sharp market downturn, stablecoin depeg, underlying asset default, a regulatory black swan event... And the frequency of any of these occurring in the crypto market is hardly rare.
4.2 Risk Formation Timeline: When Did All Four Links Simultaneously Materialize?
To determine if the combined chain has already formed, we must examine the time window when each link individually materialized (details in Figure 12).
Source: CoinFound
Our judgment is that the current juncture is precisely the first time all four links have simultaneously materialized. This is the full meaning of "the mechanism is already formed": not predicting a specific incident will happen, but acknowledging that all four necessary links for an incident chain are already in place. Whenever an external trigger appears, the chain already possesses a complete transmission path.
4.3 Stream-Re7 Incident: Partial Empirical Evidence of Three Links Simultaneously Activated
The Stream Finance xUSD depeg and multi-protocol bad debt cascade from October to November 2025 is an event where three out of the four links were simultaneously activated, creating an industry-wide trust crisis (details in Figure 13). Re7 was a party affected, not the primary liquidator; the scale did not reach systemic levels, but it brought the concept of "contagion path" from theory to empirical evidence.
Source: CoinFound
First, a brief recap. Stream Finance founder Caleb McMeans handed strategy control to a trader Ryan DeMattia with "no formal relationship"; after DeMattia's personal ETH leveraged position was liquidated on October 10, 2025, he allegedly misappropriated Stream's funds[^70]. On-chain analyst CBB0FE issued a warning on October 28; on November 3, Balancer suffered a $1-1.28 billion attack coupled with xUSD withdrawal delays; on November 4, Stream disclosed a $93 million loss, with xUSD dropping 77% that day[^71]. Re7's exposure on the Plasma chain xUSD/USDT0 cluster was about $14.65 million, plus $12.75 million in sdeUSD/deUSD positions on Euler and Morpho, totaling about $27.4 million[^72]. Related bad debts totaled about $285 million; DeFi saw a net outflow of about $10 billion in a week; total curator vault TVL fell from the October 2025 peak of about $10 billion to about $6 billion (~-40%)[^73]. Other affected parties included MEV Capital (~$34M), TelosC ($123-124M), Silo ($15.4M), Varlamore (over $19M), and Elixir's deUSD project itself collapsing 98%[^71].
Source: CoinFound
Mapped onto the four-link model, the event activated the first, third, and fourth links; the second link was established as background but not directly triggered by this event. First link: xUSD could not actually be realized instantly, redemption queues lengthened, a bank run occurred. Third link: xUSD → Stream → deUSD → multi-protocol vault contagion along nested chains, a single asset depeg diffused to multiple vaults. Fourth link: Leading Curators' xUSD-related exposures (Re7, MEV Capital, etc.) directly translated into systemic shocks, with multiple leading Curators affected simultaneously deepening the industry trust crisis. The second link (large-scale mismatch scaling of RWAs/LSTs) was established during 2024-2025, but the direct trigger asset in this event was xUSD (a yield-bearing stablecoin structure), not RWA or LST, so the second link existed as systemic background in the event but was not directly activated[^71][^73].
Chapter 5 Tracking Framework and Scenario Analysis
To transition this report's three core judgments from argumentation into a state of continuous observation, a set of tracking indicators is needed. We build the tracking framework upon each judgment; changes in indicators necessitate revisiting the evidence layer for reassessment, while scenario analysis provides falsifiable boundaries.
5.1 Tracking Indicators
Source: CoinFound
The tracking indicators for Judgment One are the concentration trend and regulatory progress.
For Top 4 Curator share growth trend, new entrants, and regulatory progress, we use data.morpho.org's "Assets under curation" data (approx. $2.9 billion as of drafting) as a numerator proxy sample, matched against DefiLlama's full DeFi TVL as the denominator. The denominator is real, the numerator is a partial representative sample, so the magnitudes cannot be directly combined into an absolute value, but the direction of change (MoM) is stable.
On the regulatory side, all five public channels are tracked: SEC Crypto Task Force, OCC, HK SFC, EU MiCA, MAS[^84]. As of the drafting date, no jurisdiction has advanced specific regulation for Curators.
The tracking indicators for Judgment Two are whether the AUM divergence between institutional channels and retail channels can be further evidenced in data.
As the industry lacks a unified classification, we initially treat permissioned vaults and CEX-sourced vaults as institutional channels, and publicly available, freely distributed on-chain vaults as retail channels. If TVL for institutional channel vaults like Coinbase × Steakhouse continues to grow while ordinary public on-chain vault growth lags, it indicates institutional capital is indeed concentrating toward specific channels. Additionally, monitoring whether the fee advantage or fee pressure of Curator vaults changes. Currently, Curator vaults' equivalent fees are significantly higher than traditional low-fee products like SGOV and FOBXX. If the fee gap between them narrows noticeably in the future, it suggests the business model and attractiveness of Curator vaults may be shifting, requiring a reassessment of the true reasons for institutional capital's choice.
The tracking indicators for Judgment Three are the proportion of non-immediate assets, nesting depth, and a single Curator's concentration in a single primitive.
Currently, there is no unified disclosure standard for these three types of risks. Therefore, we employ alternative verification methods: using a self-built indexer to track vault supplyQueue, observing the actual nesting layers funds pass through; using Blockworks Research's Average Concentration HHI and data.morpho.org data on different vaults' combined exposure to the same collateral to cross-check; in stress testing, not citing a single standardized number but referencing already-occurred real risk incidents and risk materials self-disclosed by leading Curators. Stream/Re7 is currently a relatively complete real-world stress event case study, while Gauntlet VaultBook and Block Analitica risk reports represent risk observation lenses self-disclosed by leading Curators.
5.2 Scenario Analysis: Four Scenarios and Falsification Conditions
Source: CoinFound
Under Baseline and Acceleration scenarios, institutional capital continues to concentrate toward leading Curators along compliance channels, stablecoin vaults' equivalent fees converge with off-chain money markets, and RWA vaults retain credit and liquidity premiums. The difference between Baseline and Acceleration lies in regulatory pace: Baseline assumes "de facto takeover precedes de jure recognition," while Acceleration assumes "regulatory framework precedes major incident," the latter pushing leading shares to a higher tier within 12 months.
Inflection and Reversal scenarios correspond to two qualitatively different external shocks. Inflection refers to the full-chain accident described in section 4.3, where a major mistake by a leading Curator triggers the complete chain; Reversal refers to accelerated stablecoin settlement layer formation, where issuers/exchanges/wallets/native lending protocols directly absorb stablecoin yield demand, snatching the core asset base from Curators externally. Our judgment is that both probabilities are non-zero, but their transmission forms differ: Inflection amplifies a single-point impact into a systemic shock, Reversal is a structural migration of underlying assets; the former unfolds over weeks, the latter over quarters.
The Falsification scenario is reserved independently. If any of the following three are triggered—institutional capital does not concentrate toward leading Curators, the yield spread between stablecoin vaults and RWA vaults does not compress as capital enters, Top 4 concentration stops rising—then this report's core judgments regarding "Curators becoming institutional allocation gateways, yield spreads compressing, sector moving toward concentration" become invalid, and the corresponding chapters must be re-examined against the evidence layer. It's worth noting that this scenario is not currently supported by existing data.
Conclusion
The Vault model did not merely create a new yield tool; it completed a switch in the underlying logic of on-chain asset management: systematically concentrating the "subjective judgment rights" that were previously dispersed across the protocol layer, DAO governance, and end-users upward into a new entity, the Curator. This on-chain replication of traditional finance's "principal-agent structure" significantly reduces the due diligence and integration friction for institutional capital, providing the structural prerequisite for it to accommodate the great migration of TradFi funds. Within this framework, traditional observation metrics face obsolescence; simply tracking TVL expansion or nominal APY can no longer measure the underlying real exposures. The core factors determining a Vault's long-term commercial value and survival probability have substantively shifted toward the Curator's risk management boundaries, asset transparency capabilities, and funding structure clarity. In other words, the market's pricing anchor is converging from crude "yield generation" toward "institutional-grade risk management capabilities."
What demands vigilance is that the Vault sector currently finds itself in a vacuum period where "capital scale" and "risk control systems" are misaligned. Systemic fragility during this phase is being amplified implicitly: high capital concentration amplifies the impact radius of leading Curators' misjudgments, while complex underlying nesting may cause front-end APY to fail in pricing tail risks. The entire market urgently needs to establish on-chain stress testing and regulatory disclosure frameworks capable of matching the current multi-billion-dollar scale.
Ultimately, Vaults represent a historic compromise between decentralized technology and traditional capital at the boundary of efficiency and risk. When the decentralized utopia gives way to pragmatic active management, in this irreversible great capital migration, those who truly hold pricing power are no longer just the underlying protocols, but the Curators capable of simultaneously explaining sources of yield, asset boundaries, compliance channels, and risk transmission paths.
