Even after our initial questions and @AaveLabs ‘s response, our position did not change significantly.
https://gov.uniswap.org/t/rfc-aave-s-cdp-for-uniswap-v4-positions/25568/24
We generally agree with the sentiment @0xuniact provided above, and would love to share a different version of the modification, which we believe is slightly better for the Uniswap DAO.
https://gov.uniswap.org/t/rfc-aave-s-cdp-for-uniswap-v4-positions/25568/34
TL;DR
- A success-based reward structure is preferable over fixed payments.
- Considering current market benchmarks, the target of $500M in borrow appears overly optimistic.
- While detailed debate about projections isn’t necessary here, it is critical to significantly reduce downside risk for Uniswap DAO.
- In bad cases, such as borrowing volume only temporarily hitting $100M (approx. $200M LP TVL), which we do not hope to happen, Uniswap DAO could lose more than $10M based on Aave’s request for payment ($3.3 M cash + 2 M UNI) (Calculation shown at “Huge downside risk” section)
- Hence, we propose a reward structure that includes an initial payment of $600K in UNI tokens and a revenue-based user incentive component. This approach provides smoother and more appropriate incentives for sustained growth and mutual success.
- For Aave’s participation in Uniswap DAO governance, we would propose 50k UNI delegation from the treasury to appreciate their presence in the DAO.
Problems in the current proposal
We see three issues that must be addressed for the proposal to be economically sound for the Uniswap DAO:
-
Over-optimistic TVL assumption
The primary use case for LP token collateralization typically involves leveraged farming strategies; therefore, it is instructive to reference current and historical market sizes from comparable leveraged-farming products.
In the Leveraged Farming category on DeFi Llama, TVL has hovered around $500 million over the past year. Even at its peak, it was only about $3 billion, and that was more than three years ago (September 2021).
Similar to that, prime brokerage protocols have similar use cases (although they work differently), such as Gearbox, of which TVL has been between $80M - $400M for the past year.
Thus, Leveraged-farming + “prime brokerage” TVL seems to have hovered around $0.5-1 B for the past year. Capturing 20% of that implies ~$200 M TVL and $100M borrow, which is just ≈ $4.5 M annual revenue for Uniswap DAO (if it sustains for a year).
Once, there was a case where MakerDAO reached $1.2 billion in TVL by allowing users to use G-UNI (a wrapper for Uniswap’s DAI-USDC LP tokens) as collateral to borrow DAI. However, after that, the scale shrank. It’s worth noting that DAI-USDC at the time was a notable trading pair, both of the tokens being the market’s favorite stablecoins. When we compare it to this proposed integration, the two points are clear.
a) Stablecoin pairs have a better chance of growing. On the other hand, volatile pairs have more difficulty in adoption as LP tokens.
b) GHO needs to become one of the best stablecoins in the market to achieve a huge growth like Maker’s G-UNI case, or forget GHO and focus on the current best stable pairs to grow this integration.
-
The past challenge in Aave
Although Aave has previously allowed Uniswap LP tokens wrapper (G-UNI) to be used as collateral, the program was closed midway due to a lack of liquidity, as it says in the Aave forum, showing that the utilization of LP tokens as collateral is not quite easy.
-
Huge downside risk
If borrowing peaks at $100 M (≈ $200 M supplied LP) for just one 90-day epoch, which we do not hope to happen, Uniswap would earn around $1.1 M (= $4.5M*90/365)in rev-share yet have already spent well over $14 M ($3.3M + 1.7M UNI) (suppose 1 UNI = $6.7)when the UNI outflow is marked to market, causing more than $10M loss to Uniswap DAO.
-
Competitive neutrality
Revert Lend, Morpho, Euler and others already service LP-CDPs. Uniswap should avoid subsidising one large player at the expense of an open, competitive hook ecosystem.
Revised Proposal
We propose a carefully designed incentive structure aimed at aligning the economic interests of both Uniswap DAO and Aave DAO:
-
Initial Costs
To minimize Uniswap DAO’s theoretical maximum loss, an upfront payment of $600,000 in UNI tokens will be allocated. Supposedly, this initial funding covers essential development and early-stage incentives, providing adequate resources for the bootstrap phase.
-
Revenue-Sharing Model
To ensure mutual success and balanced incentives, revenue sharing will be based on a fixed 50:50 split of revenue generated from the borrowing activities. This evenly split structure avoids scenarios where increased TVL growth could disproportionately disadvantage either party. Importantly, this is not limited to the GHO borrow profit, but includes all other assets borrowed by collateralizing Uniswap v4 LP tokens with this module. This is also the change that we believe to be made after the original proposal.
-
Additional User Incentives
Beyond the primary revenue split, Uniswap DAO commits to allocating an additional 10% of its received revenue as further user incentives. These additional incentives, however, will only trigger once Uniswap DAO surpasses its breakeven threshold, safeguarding the DAO from early losses.
Furthermore, incentive payments from this additional allocation are structured to occur only once per $50M increment in borrow volume. For example, achieving a borrow volume of $100M will trigger an incentive payout, but no further payments will occur until the next increment of $150M borrow volume is reached.
Detailed financial projections are outlined explicitly in the spreadsheet for reference.
-
Aave’s participation in Uniswap DAO governance
While participation by teams like Aave in Uniswap DAO governance is desirable, this integration does not justify granting as much as 550,000 UNI. As a compromise, we propose delegating 50,000 UNI tokens to the Aave DAO for a period of 2 years.
Although Aave DAO is projected to receive higher revenue in the provided financial model, the disparity is justified considering Aave’s responsibilities, including R&D expenses, collateral management, and operational costs. However, ideally, as indicated by @0xuniact, additional incentives from Aave DAO would further strengthen this integration’s appeal and economic viability.
Why This Works Better
This revised structure significantly improves upon the initial proposal by clearly limiting potential losses and ensuring incentives are explicitly tied to proven success metrics. Previously, the model risked significant upfront exposure for Uniswap DAO without sufficient alignment of long-term incentives.
With this new framework:
- The maximum potential loss for Uniswap DAO is explicitly capped at $600,000, reducing exposure substantially compared to previous proposals.
- Additional incentives are explicitly linked to actual revenue, avoiding scenarios where Uniswap DAO might face significant losses due to volatile borrow rates or insufficient growth.
- The fixed 50:50 revenue split ensures fairness and prevents situations where increasing TVL growth might negatively impact Uniswap DAO’s share of revenue.
- Structuring additional incentives around discrete borrow volume increments provides clear milestones and encourages consistent, sustainable growth.
We also considered milestone-based payments, but identified a potential issue; each DAO’s revenue would shift dramatically at the exact moment certain thresholds were reached, creating an unintended incentive to manipulate TVL (although we assume such manipulation would not realistically occur). Instead, to encourage steady, continuous growth and adoption of this integration, we concluded that a linear, or near-linear, incentive structure is preferable.
Thus, we opted for a revenue-sharing model as the basis for DAO rewards, with additional user incentives tied directly to the amount of revenue generated.
Adjustments to this proposal remain possible pending further clarity from Aave DAO regarding their contributions, audits, and detailed development costs. Nonetheless, based on current information, we view this incentive structure as optimal for Uniswap DAO, presenting a clear improvement over the original approach.
Concrete changes I’d need to flip to YES
