Within one year, three major DeFi protocols have abandoned the ve model. Can DeFi still continue to thrive?

Author: Pink Brains

Compiled by: Deep Tide TechFlow

Deep Tide Intro: In 12 months, Pendle, PancakeSwap, and Balancer all successively abandoned the ve token model. Together, the TVL of these three protocols once reached several billion dollars. This article provides the most systematic post-mortem dissection in the market: where exactly each protocol hit its breaking point, what replacement mechanisms were adopted, and whether the underlying failure logic is the same. The conclusion isn’t “ve tokens are dead,” but a more precise judgment—what kinds of protocols can use this model, and what kinds cannot.

The full text is as follows:

Three major DeFi protocols gave up the vote-custody model within 12 months. Pendle, PancakeSwap, and Balancer had different breakpoints, but they ultimately reached the same conclusion.

Vote-custody tokenomics (ve tokens) was supposed to be the ultimate solution for DeFi token economics. Lock tokens, gain governance power, earn fees, and permanently align incentives—no central governance required. Curve proved it could work, and dozens of protocols replicated the model between 2021 and 2024.

But that has changed.

In the 12 months of 2025, three protocols with combined TVL in the billions of dollars concluded that this mechanism’s downsides outweigh its upsides. Not because the theory was wrong, but because execution failed: low participation rates, governance captured, emission flows into unprofitable liquidity pools, and tokens collapsing in price while usage grows.

Pendle: vePENDLE → sPENDLE

Where it went wrong

The Pendle team disclosed that despite a 60x increase in revenue over two years, vePENDLE had the lowest participation rate among all ve token models—only 20% of the total PENDLE supply was locked.

This mechanism, designed to align incentives, excluded 80% of holders. The fatal blow came from pool-level data: over 60% of the pools that accepted emissions were not profitable.

A minority of high-performing pools subsidized the majority of pools that destroyed value. Highly concentrated voters meant emissions flowed to where large holders had positions—these were wrapper products first, and only then distributed to end users.

For comparison, Curve’s veCRV lock rate is around 50%+; Aerodrome’s veAERO lock rate is around 44%, with an average lock duration of about 3.7 years. Pendle’s 20% was too low. Compared with the opportunity cost of capital in yield markets, lock incentives were not compelling. And Aerodrome, as of March, had already distributed more than $44 million to veAERO voters.

Replacement: sPENDLE

14-day withdrawal window (or instant withdrawal, 5% fee)

Algorithmic emissions (cut by ~30%)

Passive rewards, only for key PPP votes

Transferable, composable, and re-stakable

80% of income → buy back PENDLE

sPENDLE is a liquidity staking token of PENDLE at 1:1. Rewards come from buybacks supported by revenue, not inflationary emissions. The algorithmic model reduces emissions by about 30% while directing capital toward profitable pools. Existing vePENDLE holders receive a loyalty bonus (up to a 4x multiplier, decreasing over two years starting from the Jan. 29 snapshot). A wallet associated with Arca accumulated more than $8.3 million worth of PENDLE in six days.

But not everyone agreed with this decision. Curve founder Michael Egorov believes that ve tokenomics is a very powerful mechanism for aligning incentives in DeFi.

PancakeSwap: veCAKE → Tokenomics 3.0 (burn + direct staking)

Where it went wrong

PancakeSwap’s veCAKE was a textbook case of bribery-driven incentive misallocation. The gauge voting system was captured by Convex-style aggregators—most notably Magpie Finance—which siphoned emissions while bringing very little revenue to PancakeSwap’s actual liquidity.

Pre-shutdown data: pools accounting for over 40% of total emissions contributed less than 2% of CAKE burn. The ve model created a bribery market in which aggregators extracted value, while pools that generated fees were not sufficiently incentivized.

However, this shutdown was deliberately designed. Michael Egorov called it “a governance attack that’s practically a textbook example,” arguing that CAKE insiders erased the governance rights of existing veCAKE holders and may force an unlock of their tokens after voting. Cakepie DAO, one of the largest CAKE holders, challenged the vote citing the existence of irregular conduct. PancakeSwap provided Cakepie users with up to $1.5 million in CAKE compensation.

Replacement: 100% fee revenue → CAKE burn

Team directly manages emissions

1 CAKE = 1 vote (simple governance)

~22,500 CAKE per day (target 14,500 CAKE)

100% fee revenue → CAKE burn, no dividends

Target: 4% annual deflation, reaching 20% in 2030

No-penalty unlock for all locked CAKE/veCAKE positions, providing a 6-month 1:1 redemption window. Revenue distribution changes to burning; key pool burn rate rises from 10% to 15%. PancakeSwap Infinity also rolled out in parallel, using a redesigned pool architecture.

Post-transition results: 2025 net supply decreased by 8.19%, continued deflation for 29 straight months. Since Sept. 2023, 37.6 million CAKE has been permanently burned; in Jan. 2026 alone, more than 3.4 million CAKE was burned. Cumulative trading volume: $3.5 trillion (with $2.36 trillion in Feb. 2025).

The deflation scheme looks good, but the CAKE price is still around $1.60, down about 92% from its all-time high.

Balancer: veBAL → liquidation of risk (DAO + zero emissions)

Where it went wrong

Balancer’s failure was a cascade of overlapping issues: governance capture, security incidents, and economic insolvency.

The war with whales came first. In 2022, the whale “Humpy” manipulated the veBAL system, steering $1.8 million worth of BAL into the CREAM/WETH liquidity pools it controlled over six weeks. In comparison, in the same period that liquidity pool generated only $18,000 in revenue for Balancer.

Then came the exploit event. A rounding flaw in the Balancer V2 swap logic was exploited across multiple chains, resulting in about $128 million being withdrawn. TVL fell by $500 million within two weeks, and Balancer Labs again faced legally risky exposure that it could not afford.

Replacement: 100% fees → DAO treasury

BAL emissions set to zero

100% of fees allocated to the DAO treasury

Buy back BAL at a fixed price to enable exits

Focus: reCLAMM, LBP, stable liquidity pools

Maintain a lean team via Balancer OpCo

In old DeFi models, token-reward-centered structures are being phased out. Despite token-economic issues, Martinelli pointed out that Balancer “is still generating real revenue”—over the past 3 months it exceeded $1 million: “The problem isn’t that Balancer doesn’t work; the problem is that the economics around Balancer don’t work. These are fixable.”

Can a lean DAO maintain $158 million TVL without incentives is an open question. Notably, Balancer’s market cap ($9.9 million) is currently lower than its treasury ($14.4 million).

Underlying mechanism

The three exits above are symptoms; the root causes are structural.

A recent analysis by Cube Exchange outlines three scenarios in which ve token models may fail.

Assumption 1: Emissions must remain valuable. If the token price crashes, emissions lose value → LPs leave → liquidity, trading volume, and fees decline → more sell pressure. This is the classic reverse flywheel (seen on CRV, CAKE, and BAL).

Assumption 2: Locking must be real. If locked tokens can be wrapped into liquid versions (Convex, Aura, Magpie), then “locking” loses meaning and creates exploitable inefficiencies.

Assumption 3: There must be real allocation problems. The premise for ve working is that the protocol continually needs to decide where incentives flow (e.g., for AMMs). Without that, gauge voting becomes unnecessary overhead.

Diagnostic test: Does the protocol have a real, recurring allocation problem such that community-led emissions can create materially more economic value than team-led allocations? If not, ve tokenomics is only adding complexity without adding value.

Fee-to-emissions ratio

The fee-to-emissions ratio is the protocol’s fee dollar value divided by the dollar value of emissions it distributes. When this ratio is above 1.0x, the protocol earns more revenue from liquidity than it spends to attract it. Below 1.0x means activity is subsidizing losses.

Here’s a detail revealed by Pendle’s exit: the overall ratio masks the true picture of each pool. Pendle’s total fee efficiency is above 1.0x (revenue exceeds emissions). But when the team breaks it down pool by pool, more than 60% of the pools are unprofitable on their own. A small number of high-performing pools (most likely large stablecoin yield markets) subsidize other pools. Manual gauge voting routed emissions to pools that benefited major voters—not to pools that generated the most fees.

The same situation happened with PancakeSwap, but it shows up in CAKE burn.

Liquidity-locking contradiction

Ve tokenomics creates a problem: capital locking is inefficient. Liquidity locking products solve this by wrapping locked tokens into tradable derivative assets. But while they improve capital efficiency, they also create a governance centralization problem. This is the core paradox of every ve tokenomics system.

In Curve’s case, this paradox produced stable (though concentrated) outcomes. Convex holds 53% of all veCRV, with StakeDAO and Yearn holding additional shares. Through Convex, individual governance is effectively intermediated via vlCVX votes. Convex’s incentives are aligned with Curve’s success, and its entire business depends on Curve running well. Centralization is structural, not parasitic.

In Balancer’s case, this paradox was destructive. Aura Finance became the largest veBAL holder and the de facto governance layer. But with no other strong competitors, a hostile whale (Humpy) independently accumulated 35% of veBAL and manipulated gauge caps to extract emissions.

In PancakeSwap’s case, Magpie Finance and its aggregators captured gauge voting through bribes and routed emissions to pools that brought PancakeSwap very little value.

Ve tokenomics needs locked capital to function, but locked capital is inefficient. So intermediaries appear to unlock it—and in doing so, they concentrate governance power that was supposed to be distributed. The model creates the conditions for being captured by itself.

Curve’s rebuttal on why ve tokenomics still matters

Curve’s conclusion: the amount of tokens continuously locked in veCRV is always about three times the amount that an equivalent burn mechanism might remove.

Scarcity based on locking is structurally deeper than scarcity based on burning, because it simultaneously creates governance participation, fee allocation, and liquidity coordination—not just reduced supply.

In 2025, Curve’s DAO removed the veCRV whitelist and expanded eligibility for DAO governance. The protocol metrics were equally impressive: trading volume grew from $119 billion in 2024 to $126 billion in 2025, pool interaction volume more than doubled to 25.2 million transactions. Curve’s share of Ethereum DEX fees rose from 1.6% at the start of 2025 to 44% in December—an increase of 27.5x.

But here’s a rebuttal to the rebuttal: Curve occupies a unique position as a backbone of stablecoin liquidity on Ethereum, and 2025 was the year of stablecoins. Gauge-driven liquidity is real, market-driven, and organically demanded. Stablecoin issuers like Ethena structurally need Curve pools. This creates a bribery market rooted in real economic value.

The three protocols that left ve tokenomics don’t have these. Pendle’s value proposition is yield trading, not liquidity coordination; PancakeSwap’s is a multi-chain DEX; Balancer’s is programmable liquidity pools. None of them provides a structural reason for external protocols to compete for their gauge emissions.

Conclusion

Ve tokenomics isn’t universally dead. Curve’s veCRV and Aerodrome’s ve (3, 3) are working well. But this model only works when gauge-driven emissions can create liquidity with real economic demand. Meanwhile, other protocols are choosing alternatives to ve tokenomics such as revenue-backed buybacks, deflationary supply mechanisms, or liquid governance tokens.

Maybe it’s time for DeFi to have a new incentive mechanism that benefits both the protocol and token holders’ long-term interests.