[TEMP CHECK v2] Arbitrum WETH Suppliers Retention Programme — Opt-in Liquidity Stabilisation

[mrjoaoms]

Author: @mrjoaoms
Co-shaped by: @MconnectDAO, @bobgodwinx
Date: 2026-05-12 Version: 2.0 (supersedes 2026-05-02 original)

Summary

This TEMP CHECK proposes the Aave DAO formally adopt an opt-in, time-bound Liquidity Stabilisation Programme for WETH suppliers on Aave V3 Arbitrum (and equivalent markets on Base, Mantle, Linea) as a complement to the DeFi United recovery effort and to the WETH Unfreeze AIP currently advanced by @LlamaRisk.

The Programme activates at the moment of unfreeze through a 7-day opt-in enrolment window. It includes a temporary APY boost differentiated by holding period, a uniform per-day withdrawal cap that applies to all suppliers regardless of opt-in status, and an optional vested AAVE component for participants committing to longer holding periods.

This is not retroactive compensation. It is a forward-looking incentive aligned with Aave’s “No Ghost Left Behind” doctrine, designed to prevent a post-thaw bank run that could undermine the recovery the protocol has spent $300M+ to coordinate.

Changelog v1 → v2

This version incorporates feedback from the discussion period:

• Mechanic 1: Reframed from automatic to opt-in (per @bobgodwinx). Now includes APY differentiation by holding period.

• Mechanic 2: Replaced tiered withdrawal precedence with uniform per-day caps applied to all suppliers (per @MconnectDAO; reinforced via synthesis with @bobgodwinx). Eliminates two-tier depositor precedent concern.

• Mechanic 3: Reframed as optional add-on for opt-in participants, not core component. Reduced parameters.

• Eligibility: Changed from “continuous net-positive supply” to pro-rated based on % of original position maintained throughout freeze (per @MconnectDAO). Rewards commitment without penalising prudent partial de-risking.

• Added: Scenario modeling section with three parametric cases (Baseline, Stress, Tail).

• Added: Reference to recent legal validation of the underlying premise (Aave LLC’s emergency motion of 4 May 2026 in SDNY).

Motivation

Context

I am a WETH supplier on Aave V3 Arbitrum. I have no rsETH exposure on any chain. I have not interacted with the Kelp bridge. My position has been frozen since the Guardian action of April 18, 2026. I write this as an affected non-rsETH user — exactly the cohort the LlamaRisk incident report flagged as “users without rsETH exposure who haven’t been spared either.”

The DeFi United coalition (currently $300M+ in formalised commitments across Stani Kulechov’s personal pledge, Frangella’s contribution, Consensys/Lubin’s 30,000 ETH, Mantle’s credit facility, Lido, EtherFi, the Arbitrum Security Council’s 30,766 ETH freeze, and the Aave DAO’s 25,000 ETH treasury proposal) represents the most significant cross-protocol coordination in DeFi’s history.

This proposal is not a critique of that effort. It identifies a structural gap that the recovery does not address by design: what happens at the moment the freeze is lifted.

The structural gap

The DeFi United stack is engineered to restore rsETH backing and prevent bad debt socialisation. Both are necessary. Neither is sufficient.

What the plan does not address is the liquidity dynamic at the moment of unfreeze. The WETH Unfreeze AIP by @LlamaRisk (filed 2026-05-09) correctly notes that “holding WETH LTV at 0 longer than necessary creates avoidable friction for end-users.” Equally true and equally avoidable is the friction created by an unconditioned full unfreeze, which produces outflow risk on suppliers who absorbed the freeze period.

Three reasons make significant post-thaw outflows a near-certainty without intervention:

(a) Collapsed yield during the freeze. CryptoQuant’s report of 29 April found that borrow rates for WETH and USDC “collapsed to near-zero, indicating a systemic withdrawal of activity, not just a liquidity crunch.” Aave’s Slope 2 adjustment to 1.5% for WETH on frozen markets — implemented to protect existing borrowers from runaway accrual — has the side-effect of nullifying the supply-side compensation that would normally accompany 100% utilisation. Suppliers on Arbitrum have been earning approximately 2.58% APY while their capital was inaccessible.

(b) Memory of the first run. Per MEXC News (30 April), WETH liquidity on Aave V3 collapsed from $689M to $1.5M in two hours on April 18-19. Total Aave TVL fell from $26.4B to ~$14.3B currently. Suppliers who didn’t withdraw fast enough were locked. That collective memory is now priced into every supplier’s decision function.

(c) No incentive to stay. There is currently no mechanism that rewards suppliers who held position through the freeze. The DeFi United funds restore rsETH backing. They do not change the supplier’s payoff for staying versus leaving on day one of unfreeze.

Legal validation of the underlying premise

On 4 May 2026, Aave LLC filed an emergency motion in the Southern District of New York to vacate the Gerstein Harrow restraining notice that had blocked the transfer of 30,766 ETH from Arbitrum. The motion’s central argument explicitly invokes harm to suppliers as legal justification:

“Continued delays risk cascading liquidations, sustained liquidity outflows, and irreversible changes to user positions.”

Judge Margaret Garnett’s 8 May order modifying the restraining notice implicitly endorsed this framing by allowing the recovery to proceed.

The retention mechanics proposed here address a related but distinct form of the same harm: not the freeze itself, but the unfreeze dynamic. If “sustained liquidity outflows” justified federal court intervention, the same dynamic at unfreeze deserves codified protocol-level response.

Why this matters for the protocol

A post-thaw bank run is not just a user welfare problem. It is a protocol-level risk:

It would invalidate the recovery. If unfreeze triggers a $500M-$1B outflow from Arbitrum WETH alone, the recovery’s primary objective — depositor confidence — fails on day one. The capital was deployed for nothing.

It would weaponise utilisation a second time. A coordinated withdrawal cascade pushes utilisation back to 100%, forcing emergency Guardian action, which re-freezes the market, destroying credibility. The protocol does not survive two of these in six weeks.

It penalises the loyal cohort. Suppliers who held position through the freeze are precisely the suppliers Aave most wants to retain — and the most vulnerable in a run because they didn’t get out the first time. Without retention mechanics, the protocol’s incentive structure rewards exit-on-rumour and punishes commitment.

It is cheap to fix. As specified below, the cost is in the low single-digit millions of USD. The cost of not implementing it, if it triggers a second run, is two to three orders of magnitude larger.

What “retention” means here

Retention is not compensation. It is not a refund. It is not an apology.

It is a forward-looking, opt-in incentive to align supplier behaviour with protocol stability during the highest-risk window of the recovery. The relevant precedent is depositor protection mechanisms in traditional finance: the FDIC does not compensate depositors for the inconvenience of an intervention, it guarantees the deposit so depositors don’t run. The point is structural: prevent the run, and the system survives.

Same philosophy as Aave’s “No Ghost Left Behind” policy, applied to the liquidity domain rather than the bad-debt domain.

Note on existing compensation framework

The Aave Foundation’s recently launched compensation checker tool covers rsETH and wrsETH holders affected by impairment of the underlying collateral. It does not address WETH suppliers on affected L2s, whose capital was frozen but never impaired in value. This proposal does not duplicate that framework — it complements it by addressing the cohort outside its scope.

Specification

Core principles (per @bobgodwinx’s framework)

1. No restrictions on withdrawals — protocol-level caps apply uniformly to all suppliers

2. No forced retention — opt-in is voluntary and reversible

3. Fully voluntary participation — eligibility is self-declared during enrolment window

4. Transparent and clearly defined incentives — all parameters published on-chain

Eligibility (uniform across mechanics)

WETH suppliers on Aave V3 Arbitrum, Base, Mantle, and Linea who:

• Maintained at least 50% of their pre-freeze WETH supply position continuously from the Guardian freeze date (2026-04-18) until the moment of unfreeze in their respective market

• Self-declare via on-chain opt-in transaction during the 7-day enrolment window following each market’s unfreeze

• Eligibility verified via on-chain aWETH balance snapshots at the moment of opt-in

Pro-rated benefits scale linearly with the maximum drawdown from original position during the freeze:

• ≥90% of original position maintained: 100% of benefits

• 70-90% maintained: 80% of benefits

• 50-70% maintained: 60% of benefits

• Below 50%: not eligible

Mechanic 1 — Opt-in APY Boost with Holding Period Differentiation

Eligible suppliers who opt in receive a temporary supply APY boost on their WETH position for 60 days following enrolment.

Boost parameters:

• Base boost: +200% relative to baseline supply APY

• Maximum effective APY cap: 8% (gross)

• Differentiation by holding period: full boost for ≥30 days continuous post-opt-in supply; pro-rated for earlier exits

Funding source: accumulated protocol fees from the high-utilisation period during the freeze (per aixbt labs estimate, Aave generated ~$1.7M/day during the crisis week, a portion of which is attributable to affected pools).

Anti-arbitrage logic: boost is calculated on opt-in-time supply balance, not on subsequent top-ups. New capital arriving post-unfreeze cannot capture the boost retroactively.

Mechanic 2 — Uniform Per-Day Withdrawal Caps (Protocol-Level)

For the first 14 days following unfreeze, withdrawals from the WETH reserve are capped at 5% of total pool liquidity per day, applied uniformly to all suppliers regardless of opt-in status.

Rationale: a per-day cap applied uniformly across all participants prevents coordinated exit cascades without creating two depositor tiers. It is a protocol-level circuit breaker, not a user-level discrimination.

This mechanism applies independently of Mechanic 1 — even non-opt-in suppliers operate under the same throttle. The cap is symmetric, time-bounded, and lifts automatically at day 14.

Sub-parameter: if any single day’s withdrawal demand exceeds the cap, the excess queues to the next day on a first-in-first-out basis.

Mechanic 3 — Optional Vested AAVE Add-on

Opt-in participants who maintain at least 75% of their post-opt-in supply for a continuous 90 days receive AAVE token rewards equivalent to 0.25% of supply value at moment of opt-in, vested linearly over the following 90 days.

Forfeiture if supply drops below 50% of opt-in value during the vesting window.

This component is optional in the proposal: it can be ratified independently of Mechanics 1 and 2. Community can adopt the full package, the first two mechanics only, or any subset.

Sizing estimate

For Arbitrum WETH (current pool ~$300M-$400M of locked supply):

• Mechanic 1 (APY boost): assuming 50% participation rate and average $200M qualifying supply, 60-day boost at 8% effective APY versus 2.58% baseline = ~$1.8M cost

• Mechanic 2 (withdrawal caps): zero direct cost, operational adjustment only

• Mechanic 3 (vested AAVE): assuming 30% of opt-in participants reach 90-day threshold, ~0.25% × $60M = ~$150K in AAVE tokens

Total estimated cost for Arbitrum alone: $1.95M, with negligible incremental cost for Base/Mantle/Linea (proportional to respective pool sizes).

Compared to the $58M Aave is committing to DeFi United (25,000 ETH), this is a 3-4% incremental investment in the success of the primary recovery effort.

Scenario modeling — outflow risk

Per @bobgodwinx’s request for serious modeling alongside the broader recovery package. Three parametric scenarios for Arbitrum WETH outflow at unfreeze:

Scenario A — Baseline (no retention programme)

Assumptions:

• 50% of pre-freeze suppliers withdraw within first 7 days of unfreeze

• 30% withdraw within day 8-30

• 20% maintain position

• New capital inflow during first 30 days: 30% of outflow volume

Projected impact:

• Day 1-7 outflow: ~$150M

• Day 8-30 outflow: ~$90M

• Net 30-day TVL change: -$170M

• Probability of utilisation hitting 100% again: high (~75%)

• Probability of re-freeze requirement: moderate (~40%)

Scenario B — Stress (no retention, accelerated panic)

Assumptions:

• 70% withdraw within first 7 days triggered by social signal

• 20% withdraw within day 8-30

• 10% maintain

• New capital inflow: 15% of outflow volume

Projected impact:

• Day 1-7 outflow: ~$210M

• Day 8-30 outflow: ~$60M

• Net 30-day TVL change: -$240M

• Probability of utilisation hitting 100%: very high (~95%)

• Probability of re-freeze requirement: high (~70%)

• Cascading risk to other L2s: high

Scenario C — With retention programme

Assumptions:

• 30% opt in to Mechanic 1

• Day 1-7 outflow throttled by Mechanic 2 to ~5%/day max = $75M maximum

• 60% of opt-in cohort maintains position through 30-day boost

• New capital inflow: 40% of outflow volume (attracted by boost-adjusted yields)

Projected impact:

• Day 1-7 outflow: ~$60M (throttled)

• Day 8-30 outflow: ~$50M

• Net 30-day TVL change: -$45M

• Probability of utilisation hitting 100%: low (~20%)

• Probability of re-freeze requirement: very low (~5%)

These scenarios are illustrative starting points, not predictions. Welcome refinement from @LlamaRisk, @ChaosLabs, and @TokenLogic with access to more granular data on supplier composition and historical behaviour.

What this proposal explicitly does not include

To pre-empt foreseeable objections:

• No retroactive compensation for the freeze period itself

• No special treatment of large suppliers — mechanics scale linearly with supply

• No legal claim or indemnification — Aave’s existing terms of service govern

• No precedent for every freeze — criteria scoped specifically to multi-week freezes that imposed asymmetric costs on a non-causal cohort

• No interference with the WETH Unfreeze AIP — this programme is complementary and can be ratified before, during, or after the unfreeze itself

Implementation pathway

This proposal is designed to integrate with the existing Unfreeze AIP, not delay it:

1. TEMP CHECK discussion period: minimum 5 days from this v2 posting

2. If consensus: escalation to Snapshot for TEMP CHECK vote (320,000 YAE quorum, 80,000 differential)

3. If Snapshot passes: ARFC drafting with technical specification, parameter calibration with service providers

4. ARFC Snapshot: ratification by Aave DAO

5. Implementation via Risk Steward / AIP: can be deployed before, during, or shortly after WETH unfreeze depending on timing

Critical timing note: the WETH Unfreeze AIP (@LlamaRisk) is on a fast track. For this Retention Programme to be operational at unfreeze, the TEMP CHECK needs to advance in parallel rather than sequentially. Requesting prioritisation by service providers given the time-sensitivity.

Disclaimer

The author is a WETH supplier on Aave V3 Arbitrum and would benefit directly from the implementation of this proposal. The author maintains the same wallet and WETH supply position continuously from before the freeze, with no intention of withdrawing post-unfreeze. The author is not affiliated with any service provider, DAO, or third party. No compensation has been received for drafting this TEMP CHECK. The author has no rsETH exposure on any chain and has not interacted with the Kelp bridge.

Acknowledgements

Substantial credit to @MconnectDAO (governance precedent framing, eligibility pro-rated structure) and @bobgodwinx (opt-in framework, four core principles, scenario modeling request) for shaping this v2. The proposal in its current form is materially better than the v1 of 2 May 2026 due to their engagement during the discussion period.

Asks of the community

• @LlamaRisk: would you consider an addendum to the Unfreeze AIP referencing this Retention Programme as a complementary measure, or co-authoring an integrated specification?

• @ChaosLabs: are the scenario modeling assumptions above reasonable as starting points? Would you have access to historical Aave outflow data from prior stress events that could refine the calibration?

• @TokenLogic: does the 25,000 ETH ARFC scope permit integration of a retention component funded from accumulated freeze-period protocol fees? If not, is a parallel ARFC the right path?

• @MarcZeller / ACI: would the Aave Chan Initiative consider sponsoring this proposal as a constructive supplement to the recovery framework?

• Other forum participants: feedback from suppliers in equivalent positions on Base, Mantle, and Linea is particularly welcome. The mechanics are designed to apply uniformly across all affected L2s.

Next Steps

1. Gather final community feedback during the discussion period (target: 5 days from v2 posting, closing 2026-05-17)

2. Refine eligibility criteria and sizing estimates based on input from service providers

3. If consensus, escalate to Snapshot for TEMP CHECK vote

4. If Snapshot passes, proceed to ARFC with technical specification

5. Coordinate implementation timing with the Unfreeze AIP execution to ensure operational availability at unfreeze

Copyright

Copyright and related rights waived via CC0.

[EzR3aL]

Hi, please follow the official guidelines for posting a proposal. Check the TEMP CHECK for this.

[mrjoaoms]

Thank you for the guidance, @EzR3aL. The proposal has been reformatted as a TEMP CHECK following the official framework. Same thread, updated content above.

[MconnectDAO]

From my side, I agree that a carefully managed unfreeze is necessary here, but I am not convinced the current Retention Programme needs to be this complex or heavy.

A few safer, simpler levers could already go a long way in reducing bank‑run risk:

• A narrow, capped APY boost funded from protocol fees, without extra layers like vested AAVE.

• Equal, chain‑wide throttling / phased unfreeze (per‑day withdrawal caps for everyone) instead of priority queues that create two depositor tiers.

• Softer eligibility that does not punish prudent partial de‑risking, while still rewarding users who remain through the unfreeze window.

In my view, starting from these simpler tools would better balance protocol safety, user fairness and long‑term governance precedent, before we commit to a bespoke, cohort‑specific package.

@mrjoaoms

[mrjoaoms]

Thank you @MconnectDAO, this is precisely the kind of governance-level critique the proposal needs. Three responses inline:

On chain-wide throttling vs tiered precedence (your point 2): You are right, and the principle behind your point is more important than the mechanism. Creating two depositor tiers introduces a precedent the protocol shouldn’t set lightly — once you accept differential withdrawal rights based on past behaviour, you’ve crossed a line that’s hard to walk back. A flat per-day cap (e.g., 5% of pool liquidity) applied uniformly achieves the anti-bank-run goal without that institutional cost. Will revise Mechanic 2 in v2.

On softer eligibility (your point 3): Agreed, and this is the strongest of your three points. The current “continuous net-positive” criterion penalises rational risk management during a multi-week freeze, which contradicts the proposal’s own logic. A pro-rated structure makes more sense: full benefits for continuous holders, partial benefits scaling with how much of original position was maintained throughout. This rewards commitment without punishing prudence.

On vested AAVE rewards (your point 1): Here I’d push back, but acknowledge the simplicity argument has weight. The APY boost addresses immediate post-thaw behaviour. The vested component addresses a different problem: distinguishing loyal capital from opportunistic re-entries that arrive after unfreeze, capture the boost, and exit at day 61. Without some medium-term hook, the programme effectively subsidises arbitrage rather than retention.

That said, the form matters less than the function. Open to alternatives: stkAAVE rewards (built-in alignment via Safety Module), a smaller vesting fraction (0.25% instead of 0.5-1%), or removing Mechanic 3 entirely if the community judges the precedent risk outweighs the retention benefit. Curious what you’d see as the line between “appropriate retention incentive” and “bespoke cohort-specific package” — it’s a real tension and I don’t think the proposal resolves it perfectly yet.

Will draft v2 incorporating points 2 and 3 over the coming days. Your feedback on Mechanic 3’s framing — keep, modify, remove — would be valuable before that revision. The simpler-is-better instinct is often correct in governance, and a leaner proposal that establishes the principle is more useful than a comprehensive one that creates resistance.

[bobgodwinx]

This makes a lot of sense.

A lot of the current discussion has understandably focused on restoring solvency and unwinding the immediate liquidity constraints, but that only addresses the balance sheet side of the issue. It doesn’t fully address the behavioral side: supplier confidence.

Once withdrawals are re-enabled, it’s entirely rational to expect a significant portion of locked WETH suppliers to exit immediately. Many users have had capital effectively trapped while earning suboptimal yield and carrying elevated protocol risk. Even if the underlying insolvency concerns are resolved, confidence doesn’t automatically recover at the same time as solvency.

The point about this not being retroactive compensation but rather a forward-looking liquidity stabilization mechanism is especially important. Preventing a disorderly withdrawal event should be viewed as protocol risk management, not a bailout.

A temporary retention program could help smooth the transition back to normal market conditions, provided it is:

• clearly time-bounded
• transparently funded
• targeted specifically at affected suppliers
• calibrated to avoid over-subsidization

The withdrawal-priority concept is also interesting, since it directly recognizes the users who absorbed the most uncertainty during the freeze.

At minimum, I think this is worth serious modeling alongside the broader recovery package. Solvency restoration alone may not be sufficient if supplier trust and liquidity depth are not actively rebuilt.

[mrjoaoms]

Thank you @bobgodwinx. The “balance sheet vs behavioural side” distinction is sharper than my original framing — solvency restoration and supplier confidence are genuinely separate problems with separate remediation paths, and conflating them is precisely how recovery efforts have failed in past DeFi crises. Will adopt that framing explicitly in v2.

Your four criteria, time-bounded, transparently funded, targeted, calibrated against over-subsidisation,will be hard requirements in the revised specification. The fourth is where the current draft is weakest: without explicit anti-arbitrage logic, the APY boost risks attracting opportunistic capital arriving post-unfreeze rather than retaining the locked cohort. Working on language that ties eligibility to pre-unfreeze position holding, not just post-unfreeze behaviour.

One tension worth surfacing: @MconnectDAO raised a valid concern earlier in the thread about Mechanic 2 (withdrawal precedence) creating two depositor tiers and setting a precedent the protocol should be cautious about. You see the asymmetry as the justification for asymmetric treatment. Both views have merit. A possible synthesis: uniform per-day withdrawal caps for everyone (addresses the precedent concern) combined with differentiated APY treatment based on holding period inside Mechanic 1 (preserves the recognition of suppliers who absorbed the freeze). The withdrawal mechanism itself stays neutral; the recognition lives in the yield structure. Curious if you see this as a reasonable middle ground or a dilution of the retention logic.

The call for serious modeling alongside the broader recovery package is the right next step. v2 will include scenario analysis on outflow risk under different parameter choices.

[MconnectDAO]

Thanks for the thoughtful synthesis. The middle ground uniform per-day withdrawal caps for everyone + APY differentiation based on holding period feels like a reasonable balance. It addresses the two-tier precedent concern while still recognizing suppliers who absorbed the freeze. I’d support this direction moving into v2. Looking forward to the scenario modeling on outflow risk. @mrjoaoms

[mrjoaoms]

Appreciate the explicit endorsement, @MconnectDAO. That gives v2 a clear structural direction:

• Mechanic 2: uniform per-day withdrawal caps for all suppliers (no precedent issue)

• Mechanic 1: APY boost differentiated by holding period (recognition lives in yield, not in withdrawal mechanics)

• Eligibility: pro-rated based on position maintained throughout freeze (per your earlier point)

Mechanic 3 (vested rewards) remains the open question. Will draft v2 with all three above, present Mechanic 3 as an optional component for the community to decide on rather than embed it as core. Outflow scenario modeling included, as flagged by @bobgodwinx.

Worth noting in this context: Aave LLC’s emergency motion filed May 4 in the Southern District of New York (Morrison Cohen LLP) explicitly invokes the harm to non-rsETH suppliers as core legal justification for vacating the Gerstein Harrow restraining notice — the filing argues the freeze is causing “irreparable harm” via “cascading liquidations, sustained liquidity outflows, and irreversible changes to user positions.” The legal language essentially mirrors the structural concern this proposal addresses, which suggests the bank-run dynamic and asymmetric supplier costs are now being recognized institutionally as material protocol risks rather than abstract scenarios. Reinforces the case for codifying retention mechanics into the recovery framework rather than leaving the post-thaw window unaddressed.

Targeting v2 within the discussion period window. Will tag both of you when posted.

[bobgodwinx]

Let me reiterate that the protocol should aim to restore normal functioning as closely as possible to its pre-incident state.

The protocol remains solvent, and bad debt has been accounted for. However, even in this context, liquidity dynamics at the point of unfreezing remain an important consideration.

A key risk is not solvency, but the potential for clustered withdrawals driven by reflexive behavior, as users may choose to exit simultaneously once withdrawals are re-enabled.

To address this, I would suggest exploring an opt-in, time-bound retention incentive programme. Such a mechanism would allow users who choose to remain supplied during the transition period to be rewarded for providing liquidity stability during a potentially fragile phase.

The important principles here are:

• No restrictions on withdrawals
• No forced retention
• Fully voluntary participation
• Transparent and clearly defined incentives

This would allow the protocol to respect user autonomy while potentially smoothing liquidity dynamics during the post-unfreeze period.

I think AAVE community delivered a blow to the North Korea Lazarus group and the unity demonstrated will pay-off.

So let people opt-in in the proposal. For me I am staying but I understand those who panic and go to a more risky protocol.. just saying my one cent

[mrjoaoms]

Thanks @bobgodwinx — this is a cleaner formulation than what I had drafted in response to the earlier synthesis. The opt-in framing genuinely improves the proposal in three ways:

(a) It dissolves the moral hazard objection. Users actively choose to participate, with full transparency about what they’re committing to. No protocol-imposed retention, no defaults, no implicit lock-up. Pure user agency.

(b) It reinforces @MconnectDAO’s earlier point about precedent. Opt-in mechanics don’t create two depositor tiers in any meaningful sense — they create one tier of users who chose to take an explicit deal, and another of users who chose not to. Both choices are equally legitimate and equally available to anyone.

(c) It surfaces honest signal. Users who opt in are voluntarily declaring “I am staying.” Users who don’t are voluntarily declaring “I want flexibility.” Both are useful information for the protocol. Your own statement that you’re staying is exactly this kind of signal made personal — and for what it’s worth, I’m staying too. Same wallet, same WETH supply on Arbitrum since before the freeze, no intention of moving.

One technical tension worth flagging: opt-in mechanisms historically have lower take-up than automatic ones — behavioural research is consistent on this. Risk: a fully opt-in retention programme captures only the cohort that would have stayed anyway, missing the marginal supplier whose decision could genuinely tip toward exit. Mitigation: a clear, well-communicated opt-in window during a defined post-unfreeze period (e.g., 7-day enrolment), so the choice is salient rather than buried in UI.

Synthesis for v2 — combining everything that’s emerged in this thread:

• Mechanic 1: opt-in APY boost, available during a defined post-unfreeze enrolment window, transparent terms, no lock-up beyond the boost period itself

• Mechanic 2: uniform per-day withdrawal caps for all suppliers (per the @MconnectDAO synthesis), independent of opt-in status — this is protocol-level, not user-level

• Mechanic 3 (vested AAVE): reframed as optional add-on for opt-in participants who commit to longer holding periods, rather than core component

• Eligibility: opt-in itself becomes the eligibility criterion, with secondary verification of supply position at the moment of opt-in

This version respects user autonomy fully, removes any “forced retention” reading, addresses the bank-run dynamic by giving the protocol a tool to incentivise (not require) stability during the fragile window, and creates clean information signal in both directions.

Will incorporate into v2. Three contributors, one coherent design. Genuinely how governance discussion should work.

[mrjoaoms]

Update: Judge Margaret Garnett issued a two-page order on May 8 modifying the restraining notice to allow the Arbitrum onchain governance vote and the subsequent transfer of the 30,766 ETH to a wallet controlled by Aave LLC, while preserving the underlying claims for future resolution.

Two implications relevant to this thread:

First, the recovery timeline is now legally cleared. The Arbitrum DAO Snapshot has overwhelming support, the onchain vote can proceed, and the technical unwind plan can execute. The post-thaw window — the focus of this proposal — is now a near-term planning horizon, not a hypothetical scenario.

Second, the bench has implicitly endorsed the structural framing this proposal rests on: Aave Labs argued in its emergency motion that the freeze was causing “irreparable harm” via “cascading liquidations, sustained liquidity outflows, and irreversible changes to user positions.” The judge’s modification respects that argument by allowing the recovery to proceed. The retention mechanics this thread proposes are the natural extension of the same logic — preventing a different but related form of harm at the moment the freeze ends.

Will accelerate v2 drafting given the timeline now seems concrete rather than speculative.

[mrjoaoms]

Critical update on timeline: LlamaRisk just published a Direct-to-AIP yesterday (May 9, 9:09 UTC) proposing immediate WETH unfreeze on Arbitrum, Base, Mantle, and Linea, alongside LTV restoration to pre-exploit values. Direct-to-AIP path means execution within ~10-14 days, not the 4-6 week window we were planning around.

This significantly compresses the runway for this proposal. Three implications:

First, the post-thaw window is no longer a hypothetical scenario for late May or June — it’s a near-immediate operational reality. The bank-run dynamic this thread addresses needs to be considered alongside the unfreeze AIP itself, not after it.

Second, the structural argument is reinforced by the LlamaRisk specification: the proposal correctly identifies that “holding WETH LTV at 0 longer than necessary creates avoidable friction for end-users” — but it doesn’t address the equally avoidable friction (or worse) that immediate full unfreeze creates via outflow risk on suppliers who absorbed the freeze period.

Third, this changes v2 timing. I’ll accelerate v2 drafting to this week rather than next, so it can be considered in parallel with the unfreeze AIP discussion. If retention mechanics aren’t codified before the Risk Guardian lifts the freeze, the structural opportunity to align supplier behaviour with protocol stability during the most fragile window is lost.

Will tag @LlamaRisk in v2 specifically given the operational adjacency. Open to coordinating directly if there’s appetite to integrate retention mechanics into the unfreeze sequence.

[mrjoaoms]

Note: v2 of this proposal was just posted above (edited the original post).

Tagging service providers separately given the time-sensitivity of the parallel Unfreeze AIP: @LlamaRisk @TokenLogic

Additional context for @ChaosLabs and @MarcZeller in the v2 body itself.