You Don't Own
What You Think
You Own

1. The "DAO Governance" Myth

For most of the cycle, "DeFi governance token" was treated by institutional buyers as a coherent asset class. AAVE, UNI, MKR/SKY, MORPHO, AERO, LDO: all priced on similar multiples, all evaluated through similar frameworks, all assumed to convey similar rights. They do not. Each protocol has evolved a structurally distinct governance architecture, and the rights attached to a "governance token" differ so widely across them that comparing them on the same multiple is a category error.

The right frame is not how decentralized each protocol is. It is what each token actually buys you. One of these tokens entitles you to all of the protocol's net revenue. One entitles you to vote on parameters but no economic claim at all. One requires you to lock for four years before you receive anything. One can be vetoed by a separate token class that wasn't even part of your investment thesis. These are not variations of the same instrument.

Three findings from TI's protocol-by-protocol review made this concrete enough to publish. First, the dispersion in revenue rights across the six protocols below is wider than the dispersion in revenue multiples across all of large-cap DeFi. Second, the protocols with the most concentrated revenue rights (Aave post-AWW, Aerodrome ve-lockers) are not the same protocols with the most decentralized governance. Third, the architecture that determines who gets to act in a crisis is uncorrelated with the architecture that determines who gets paid in steady state.

2. Executive Governance: Aave

Economic decentralization. Operational centralization. A note on Hyperliquid.

Aave is the cleanest example of economic decentralization with operational centralization. The AWW proposal (April 2026) routes 100% of protocol revenue to AAVE holders; no Labs skim, no treasury intermediation. Cashflow rights are full. So are voting rights, nominally.

Execution power is a different story. Aave's architecture is not a single 5/9 multisig (as several third-party writeups have characterized it). It is a three-tier authority structure, visible in real time during the April 2026 rsETH event:

• Protocol Emergency Guardian (5-of-9 multisig, publicly named signers): pause and freeze authority. Used April 18 to freeze rsETH and wrsETH across 11 markets within hours of the depeg.
• Risk Stewards (Gauntlet + Chaos Labs functional role): parameter management. Used April 19 to cut WETH Slope2 borrow rate from 10.5% to 3%, preventing bad debt from compounding via accrued interest.
• Full Governance (token-weighted, on-chain, timelocked): asset listing, major upgrades, the AWW routing change itself. Required for any structural change.
• OneSig 7-of-10 (separate multisig for cross-chain operations): signers hash locally, preventing a compromised backend from slipping unauthorized changes through.

This is closer to a corporate-board model than a single-multisig dictatorship. The structural critique is not that 5 individuals control everything; it is that Guardian scope is broad, and the next stress event will test whether the boundary between emergency parameter change and structural policy change stays clean.

3. Representative Governance: Sky / MakerDAO

Built for being right rather than being fast.

Sky (the rebranded MakerDAO) is the cleanest representative-democracy example in DeFi. SKY/MKR holders vote on values and strategy, then delegate operational authority to specialized committees and recognized delegates. The Sky Core Council handles technical decisions; SubDAO structures split specific protocol functions into separate governance scopes. The multi-stage process separates discussion from formal proposals from execution.

The trade-off is information cost. Sky's governance produces high-quality technical decisions because the voters are domain specialists. But the complexity discourages broad participation, and the layered process makes Sky less responsive when conditions change quickly. The system is built for being right rather than being fast.

Revenue treatment is the subtle distinction. Sky does not route revenue directly to MKR/SKY holders the way Aave routes it to AAVE holders. Surplus flows through the surplus buffer and is allocated by governance. Token holders capture economic upside only by exercising the governance system itself.

4. Democratic Governance: Uniswap

High legitimacy. Low throughput. The fee switch is the test.

Uniswap is the high-threshold direct democracy. Any proposal requires 2.5M UNI ($5M+ at spot) to submit and 40M UNI (4% of supply) to reach quorum. Snapshot voting handles signal; on-chain with a 2-day timelock handles execution. No multisig acts unilaterally. No Risk Steward layer. No structured delegates. Every meaningful decision goes through the same broad-participation process.

The fee switch is the cleanest historical test of what the system can and cannot do. Debate began in 2020. The mechanism existed. Activation was proposed multiple times. Legitimate concerns (securities classification, technical execution, treasury vs holder routing) repeatedly slowed the path to a vote. The UNIfication proposal finally activated the fee switch in December 2025, routing protocol fees through a buyback-and-burn mechanism with an initial 100M UNI retroactive treasury burn. Per DefiLlama, cumulative holders revenue since activation reached $14.9M by June 2026, running at ~$3.7M per 30 days. UNI holders went from "nominal economic rights, effective zero" to actively receiving deflationary buyback flow.

That activation took more than five years from first debate. The strength of Uniswap's model is that it cannot be subverted by a small group, and UNIfication demonstrated that democratic governance can deliver economically consequential decisions when consensus eventually crystallizes. The weakness is the timescale: a theoretical Uniswap equivalent of the April 2026 rsETH event would have required a 7-day proposal cycle, longer than the window in which the affected positions would have been unwound by market liquidations. The fee switch shows the system can move on multi-year strategic questions. It does not show it can move on multi-day crises. The trade-off is structural, not a defect.

5. Stakeholder Governance: Aerodrome

Conditional rights, locked for four years.

Aerodrome is the cleanest example of conditional governance rights. AERO is two assets. Unlocked AERO trades freely but earns no fees and gets no votes. veAERO (AERO locked for up to four years) earns 100% of trading fees from voted pools, captures external bribes, and controls gauge-weight emissions. Voting power decays linearly across the lock, so the longer you lock, the more weight you have today.

This produces a specific alignment. Voters are by construction long-horizon, and the flywheel (protocols bribe veAERO voters → emissions to voted pools → fees flow back to veAERO voters) is mechanically self-reinforcing. It also concentrates governance power among patient capital in a way that nominal holder counts obscure. 40% of initial veAERO was distributed to existing veVELO lockers on Optimism: a "no traditional VC sale" signal that is also a concentration of governance among a self-selected operator class from launch.

The risk Aerodrome is exposed to is the inverse of Uniswap's: ve-lock protects against short-term manipulation but enables long-term capture by committed actors.

6. Guardian Governance: Morpho

Dual gates. Both directions of capture checked.

Morpho's structure is the most restrictive on paper and the most balanced in practice. Three gates: (1) proposer must hold 500K MORPHO ($1M+ at spot), (2) proposal must clear off-chain Snapshot voting, (3) a guardian role plus a 5/9 multisig for execution must consent. Large holders cannot unilaterally pass proposals; the guardian cannot act without significant token-holder support. Both directions of capture are checked.

The cost is that most MORPHO holders effectively cannot participate. The 500K threshold prices out individuals entirely; off-chain Snapshot reduces transparency; the 5/9 multisig has signers who are not all publicly identified.

The revenue framing is the most stark in the cohort. Morpho's fee switch is not active. The design caps it at 25% of borrower-paid interest maximum, and even at activation the revenue would flow to the DAO treasury, not directly to MORPHO holders. MORPHO holders today have neither cashflow rights nor unilateral governance rights. The token's value rests on optionality that one or both might change. The market is pricing that optionality at a meaningful discount, which is mechanically consistent with what the holder actually owns.

7. Bicameral Governance: Lido

Two token classes. Each can check the other.

Lido is the structural novelty in this set. LDO holders control the Lido DAO: validator-set membership, fee parameters, treasury allocation, upgrades. stETH holders, since the Dual Governance mechanism was approved in June 2025, can delay or veto LDO governance proposals through a separate review process. A literal bicameral legislature: two distinct token classes with overlapping but not identical interests, each able to act within their own authority and each able to check the other.

The motivation was a real disagreement. LDO holders (largely VCs and early investors) could in principle make decisions that benefited LDO at the expense of stETH stakers; the protocol's safety properties ultimately depend on stETH being trusted. Dual Governance is the structural fix. It is the only DeFi governance model where a token class outside the "governance token" holds direct check power.

For LDO holders specifically, this produces a more constrained ownership picture than the protocol's nominal scale suggests. LDO does not receive direct protocol revenue. LDO controls the DAO but is checked by stETH on any decision that materially affects stakers. Closer to an equity stake in a company whose shareholders need approval from a separate class of customer-creditors before changing the business model. Novel; also a reason to be careful pricing LDO as comparable to AAVE.

8. Governance Under Stress

The test that matters happens in a crisis.

Governance systems are revealed during crises, not normal conditions. The April 2026 rsETH event provided the cleanest cross-protocol test in DeFi's history.

Aave's Protocol Guardian froze rsETH across 11 markets within hours. Aave's Risk Stewards adjusted WETH borrow rate parameters the next day. Aave's full Governance was simultaneously active on the AWW proposal. Three different speeds, three different authority levels, operating in parallel. The structure handled the stress as designed; the underwriting discipline that put 98.5% LST-funded WETH collateral on the books is the part that failed (covered in TI's same-pool funding risk report).

A theoretical Uniswap equivalent could not have responded the same way. No Risk Steward. No Emergency Guardian. Every parameter change would have required a Snapshot signal, on-chain proposal, 40M-UNI quorum, and 2-day timelock. The rsETH depeg unwound over hours; Uniswap's governance operates over days minimum. Sky's emergency procedures (ESM, GSM Pause Delay) sit between the two. Aerodrome treats most actions as a weekly emissions vote. Morpho's dual-gate structure is the slowest by design.

Governance architecture determines what the protocol can do, not just who is nominally in charge. A token whose protocol cannot respond to a crisis quickly is a different asset from one whose protocol can.

9. Every System Protects Against One Risk. Every System Is Exposed to Another.

The cleanest way to summarize the cross-asset picture is to name what each architecture protects you against and what it leaves you exposed to. There is no architecture that protects against everything. The right question for an investor is not which protocol has "the best governance"; it is which trade-off the investor is willing to underwrite.

10. The Question No One Asks

The DeFi governance debate has spent a decade on decentralization, voting mechanisms, and proposal thresholds. It has spent very little time on the more fundamental question: what rights should a governance token actually convey?

The choices are not abstract. They include (1) cashflow rights, (2) governance rights, (3) veto rights, (4) emergency-action rights, and (5) treasury rights. The six protocols above answer these five questions in six different ways, and the variance in answers is wider than the variance in valuation.

The natural next step for institutional DeFi analysis is to stop treating "governance token" as a uniform category and start asking, for each token, which specific rights the holder actually owns. That is closer to how equity is analyzed: shareholders of different companies hold different rights, and a fair valuation prices the specific rights, not the share-class label.

11. TI Reads by Type

• Aave (Hold): Economic decentralization is real; operational concentration is real. Watch the boundary between Risk Steward and Guardian authority on the next stress event. The structure handled rsETH; the question is whether the same scope holds under different stress.
• Sky (Hold): Most sophisticated governance system in the cohort; representative structure produces high-quality decisions but at the cost of participation and speed. Watch SubDAO scale-out.
• Uniswap (Hold): Pure democratic legitimacy. UNIfication (Dec 2025) finally activated the fee switch after 5+ years of debate, proving democratic governance can make economically consequential decisions on multi-year timescales. The remaining open question is whether the system can do that again on the next consequential decision without a similar half-decade lead time.
• Aerodrome (Hold): ve-lock alignment is real; Base ecosystem volume is the variable that resolves whether the flywheel re-accelerates. Long-term capture risk is the structural concern.
• Morpho (No rating): Fee switch off; revenue rights inactive; participation threshold high. Token's value rests on optionality that fee switch activates and routes to holders rather than treasury.
• Lido (No rating): Bicameral structure is genuinely novel; LDO ownership picture is more constrained than the protocol's scale implies. Watch the first Dual Governance veto event to test the model in practice.

12. Falsifiers

• The "six different things" thesis falsifies if the dispersion in revenue rights and emergency-action rights across the six protocols collapses within two years (most likely path: regulatory pressure forces convergence on a single template). If the architectures stay this varied through 2027, the thesis stands.
• The "executive vs democratic" trade-off falsifies if Uniswap's democratic governance can produce a second economically consequential decision within 12-18 months of UNIfication's December 2025 activation. UNIfication itself took 5+ years from first debate, which is consistent with structural throughput limitation; a faster cycle on the next consequential vote would suggest UNIfication unlocked the process, not just resolved that specific question.
• The Lido bicameral model falsifies if the first stETH veto produces a coordination failure that materially harms LDO or stETH holders. If the model resolves a real disagreement cleanly, it becomes a reference architecture for other dual-token systems.