Liquid Buybacks: NEST execution with LDO/wstETH liquidity

Proposals
1

Summary

We are proposing an automated buyback mechanism that will deploy LDO/wstETH liquidity in a Uniswap-v2 style LP position across the full range and keep ownership of the position in Aragon Agent.

If voted in, it could be done approximately in Q1 2026. However, the purpose of this thread is to invite opinions about the mechanism, the proposed parameters and discussions about alternative options to achieve the same goal.

NEST execution for LDO/wstETH liquidity

The simplest version of a buyback would simply acquire LDO through NEST. However, referencing the slippage tables from the earlier proposal, these executions could take place 14 times over the course of a year in 350k clips without exceeding 2% price impact excluding the cost of gas. The key tradeoff is slippage and gas. Larger clips and fewer executions will result in larger slippage but consume less gas.

The risk is that as the USD value increases, all else being equal, the pressure on the liquidity of the pool increases as well and potentially meets a limiting factor with available on-chain liquidity for LDO/(w)stETH. The bottleneck is clearly the supply of LDO, which even a Cowswap solver could eventually exhaust tapping CEX order books and onchain liquidity alone. Our proposal aims to solve this bottleneck by increasing the depth of the order book while simultaneously achieving the goal of removing LDO from circulation with excess surplus.

The framework proposed aims to limit buybacks to moments when ETH price is relatively high in USD terms and when the USD value of the annualized revenue exceeds a defined parameter.

For the sake of the example, the proposed initial parameters would be to enact surplus distributions while:

  • ETH > 3000
  • Revenue in USD > $40m
  • Distribution: 50% of treasury inflows from staking over $40m
  • Frequency: Limit to 2% total price impact based on prevailing market liquidity
  • Cap: Maximum $10m on a rolling 12mo basis

If ETH price were to decline under 3000 or the USD revenue equivalent under $40m, the buybacks would not activate.

This model is anti-cyclical, in that the better ETH price does or the more USD revenues the DAO is able to generate through fees, the greater the absolute value is distributed and, conversely, in bear markets, the buybacks would naturally trend down until stopped to avoid over-distribution.

At time of writing, this would imply:

~$4m annualized distribution, executed over at least 12 trades or more throughout a 12mo period with up to 100 stETH deployed per execution (fewer if more frequent).

The LDO/wstETH LP alternative follows in the footsteps of the Maker Smart Burn Engine and deploys NEST orders for approximately half the value of LDO per clip but pair it in a Uniswap v2-style or Curve Cryptoswap LP position to progressively increase the liquidity depth for LDO onchain while still taking LDO out of circulation. Orders could thereafter increase in frequency over time, reducing the aggregate slippage loss per trade. An initial transaction directly from Aragon Treasury would mint the initial position and seed the pool with a 1-3bps fee accruing to the DAO as the LP holder.

image-1762853671051-g380nk1gf
image-1762853671051-g380nk1gf1200×742 31.8 KB

Chart assuming a $400k seed of a liquidity position (executed through a 50 stETH swap for 200k LDO and paired with another 50 stETH) along with the corresponding curve shifts with an increase in x of the pool size

The level of price impact improvement from relatively small increases in LP size through NEST would substantially improve the experience of using LDO onchain as well as effectively take LDO out of circulation.

The pitfalls are that development and execution are a bit more complex. Reliance on simple LP smart contracts, such as Uniswap v2, make the implementation a little bit easier at the expense of some efficiency for making the order book. The goal of this LP position is less to act as a profitable trader and more to increase utility to LDO while still achieving the goal of removing LDO from circulation through a buyback mechanism. It needs minimal maintenance in a set-and-forget approach that suits a DAO.

Technically speaking the process would look something like the following:

  1. NEST contract loaded through EasyTrack (either with a manual trigger from TMC or with a future automated execution)
  2. Some % is sold for LDO through Stonks v2
  3. Once the LDO is acquired, the balance of wstETH necessary to increase the LP position is wrapped with LDO into the pool
  4. The corresponding LP token is sent back to Aragon Agent smart contract
  5. If the amount of wstETH is insufficient, a new EasyTrack distribution can be made and the process restarted from #3

The ownership of the position would be the Aragon Agent smart contract, always in control by token holders.

Call to action

This scaffold proposal aims to seek community feedback prior to formal enactment through a Snapshot vote. In particular, we would like to hear feedback on the framework, the parameters proposed and if other alternatives might be more suitable or effective.

Prior proposals

19 Likes
2

As an individual stakeholder, I think the Lido DAO should prioritize strengthening its governance before focusing on boosting the protocol token’s value. Currently, only about 5–6% of the circulating supply participates in votes, which is quite low and could lead to governance or security vulnerabilities. Increasing participation should be the first priority, and that’s where revenue should be directed, in my opinion. Once voting participation reaches more optimal levels, like 15–20%, then it might make sense to shift attention toward increasing the token’s value.

I actually wrote about this earlier: [RFC] Adjusting Delegate Incentivization Program , but I guess not many share the same view.

4 Likes
3

In my talks with token holders I very rarely encounter the desire for more active involvement in the governance or more active governance in general, so yeah, it’s not widely shared. I think you’re right we as a DAO want more participation but I think that natural path to it is via increasing the importance of governance by increasing the value of product lines Lido DAO has, but that will take time and is no reason to postpone thinking about token value.

9 Likes
4

Agree with the general view that it would be better / more secure for Lido DAO to face more participation. Would add these are not mutually exclusive goals. Generally we have a bias against “pay to play” voter incentivization but that’s a discussion for the other thread indeed.

3 Likes
5

Thanks for this proposal.
In general, I support the idea of ​​buybacks for projects where the token only has governance utility—this reduces the risk of a power grab in the DAO due to the stability of the governance token’s value.

  1. I have a question about revenue calculations.
    Annual revenue can be accurately determined at the end of the year.
    The question is, given the stated buyback frequency, how do we know that this condition has been met? By extrapolation, or will the last 12 months be taken into account?
  2. How do you plan to avoid intentional token pumps before buybacks? Will the 2% impact on the price be calculated after a buy position is opened?
  3. (Please correct me if I misunderstood or miscalculated.) Why was the target of $40 million chosen, and at the same time, the ETH price should be above $3,000?
    Judging by the protocol plans from the community call, the goal is to achieve at least 1.8k ETH revenue in stVaults and 3k ETH in other modules - in dollars this comes out to only 4,800 * 3000 = 14.4 million
    image
    image1004×734 99 KB
6

We can’t know next year’s revenue or spending before it ends, but we can correctly identify what is spending on buybacks that is reasonable in the worst case. If ETH price is high (>3k) and revenue this day, if annualized is over >$40m we can afford it. Next year or if market weather changes dramatically, this parameter can be adjusted.

Each event will be fairly small in size in this scheme, around 50k/day at current prices. 2% impact is a safeguard that is not supposed to be hit regualrly.

It’s extra revenue, not total. If nothing else changes (e.g. staking rate, staking share of stETH etc) and it works out, we’re talking about 13k (current revenue) + 1.8k (vaults) + 3k (margin increase) for a total of 17.8k.

6 Likes
7

Based on private converstations, I found out that these details are not clearly stated.

In this proposal the NEST (auction to buy LDO, essentially) is supposed to be run daily, with 50% of that day’s treasury inflows (half of 10-11 stETH on most days). It will have an expected price based on oracles (e.g. current LDO price is about 0.0002406 ETH). If the resulting price in an auction will be more that 2% different from current oracle price, auction will be failed and called off.

Based on an expected size of a clip and LDO liquidity, 5-6 stETH can’t move LDO price 2% naturally. 2% price impact means that either:

  • oracle is faulty and we can’t trust it to set up expected price; good reason to call off auction
  • auction mechanism is faulty, which is good reason to call off an auction.

It’s a technical safeguard against fault mechanisms more than economical reason.

3 Likes
10

It’s great to see this type of approval at this time; I hope something close to this can be implemented.

I agree with the thresholds ($40m in revenue; ETH > $3000) but I do think it would be a good idea to have a third threshold which relates to $LDO (e.g, buybacks will happen if the price of LDO falls below a certain ratio to ETH). I am, however, unclear on the 50% distribution and the cap.

Why would ‘only’ 50% of treasury inflows from staking over $40m be used for buybacks? With the $40m revenue threshold, I believe the operational costs are covered; 100% of inflows above this $40m wouldn’t make sense as there need to be funds retained for growth and general treasury use, but 50% seems low (especially given the $40m revenue requirement). A better way to implement distribution may be to use a sliding scale whereby for a set period of time (5y? 10y?), the distributed increases - this would allow the DAO to retain a higher in Y1, slightly less in Y2 and so on (with the final never hitting 100).

I do not think a cap should be implemented, especially given the thresholds that would be in place. I believe a cap would reduce the buybacks beyond the requirements already in place and I believe the frequency requirement(s) are sufficient so as to avoid smashing the buys.

Finally, whilst I agree with the ETH > $3000 requirement, is it best to refer to mcap?

2 Likes
11

If $40M is already allocated for operational expenses, then adding another rule that only 50% of the remaining revenue will be used for buybacks doesn’t make much sense. The $40M threshold already covers operational needs, so the funds beyond that should be considered “surplus.”

Limiting buybacks to just 50% of that surplus — and capping them at only $10M — feels unnecessarily restrictive. If the protocol’s revenue truly exceeds $40M, then the DAO should be able to allocate a higher portion of that excess directly toward buybacks without such tight limits.

1 Like
12

This is an interesting idea and to address @kaaniko ‘s point, what might be interesting is a sliding scale as @ZK0T suggests, but rather than time-based, have it tapering out to 100% towards a revenue limit, as a way of reflecting or approximating the ‘relative’ opportunity cost to a new token holder. i.e. the faster the surplus accrual the more likely it becomes that a token holder will prefer to have control over the alternatives.

We’ve thought about this extensively and generally agree with the perspective regarding LDO price - such a threshold could certainly be implemented at a later stage. Our own view is that choosing the right threshold for relative valuation is harder and more arbitrary than for revenue and ETH price which are reflective of the ability to invest in maintenance and growth and the level of the cycle. Of course, the ETH and revenue thresholds are arbitrary to a degree as well but have closer first-order relevance to developing initiatives that will help the DAO grow.

LDO price also has a relationship as a reflection of the incremental dilution that a new LDO issued would represent to invest in a given grant but drawing the right threshold is much more difficult / finding the right threshold.

1 Like
13

instead of ETH > 3000 condition

I would suggest Buy when ETH<3000, buy more when ETH<2000
Go all in when ETH<1000

15

What happens to the tokens after the buyback? Are they burned or airdropped to token holders?

16

This is based on assumption that growing the treasury is strictly worse than buying the token, which I’m unconvinced is true. Treasury itself can be pretty valuable for having financial stability and runway, engaging in acquisition opportunities, backstopping the risk, and allocating capital to unforeseen growth opportunities.

The proposal here is compromise between people who want full buybacks and people who want full treasury growth which are more of publicly silent type but not a small share of token holders.

2 Likes
17

Buybacks have become a structural necessity in today’s token economy.

As the cycle matures, bull market ending, retail leaving, tokens falling, we need to realizing that if we don’t give holders a reason to keep holding now, they may lose them forever.

Hard truth.

But we can see this model can have an immediate impact, even before a formal vote:
According to Nansen, this week’s top LDO buyers include leaderboard wallets, whales, and even market makers,.. most of them only buying. Top 100 LDO wallets added +$316K, with zero sell side activity.

The thresholds of ETH > $3000 and revenue > $40M are solid anti cyclical guards, so DAO only deploys surplus during periods of strength.

However, i think tying the trigger solely to ETH price doesn’t capture LDO’s relative valuation. LDO could remain cheap even when ETH rallies.

I’d suggest adding valuation based triggers, can consider the LDO/ETH ratio or a simple onchain P/E so buybacks kick in when LDO is truly undervalued relative to protocol performance, not just when ETH pumps.

1 Like
18

The idea behind this proposal is clear. It tries to reduce the amount of LDO in the market while also making it easier for people to trade LDO without big price swings. I like that it only turns on when the protocol is doing well, so it does not put pressure on the treasury during slow periods.

Using an LP instead of only buying back LDO is a bit more involved, but it could help build deeper liquidity over time without needing constant attention. That part feels useful.

My only questions are around the limits and the triggers. They look fine for a start, but they may need a review later as things change. It would also help to know what happens if something goes wrong and how quickly the system can be paused.

Overall, the direction feels solid and worth exploring further.

19

As a general response, the idea is indeed to launch and iterate on the parameters. It’s easier when we agree on a common framework (generalized buybacks directed to increasing the liquidity and utility of LDO) and we can have meaningful data-informed discussions regarding the specific parameters later on.

Appreciate all the comments and replies to the thread.

2 Likes
20

Find this idea very appealing and experimented early on with what those triggers could look like. We shouldn’t discard necessarily forever just that, for the time being, in our view it felt too complicated to introduce at this stage

2 Likes
21

Hi everyone,

This is my first attempt at conducting quantitative analysis, and I’d appreciate feedback and discussion from the community. The goal of this work was to understand the economic implications of the proposed treasury strategy to buy back LDO and pair it with wstETH in a Uniswap V2-style pool. To do that, I ran Monte Carlo simulations to estimate the expected impermanent loss (IL) and the structural cost of providing liquidity under current market conditions.

This post focuses on IL only. A follow-up post will incorporate trading fee modeling and net results.

1. Methodology

I used daily LDO and wstETH price data from May 15, 2023 (the post-withdrawals period which was the last significant event for LDO on DefiLlama) through today. From this dataset, I calculated log-returns, volatilities, correlations, and average drift. Using these observations, I ran 20,000 Monte Carlo simulations with a standard geometric Brownian motion model, applying a Uniswap V2 50/50 AMM structure to estimate how a liquidity position would evolve over a 365-day period.

This model should not be interpreted as a prediction, but rather as a reasonable baseline for understanding the expected IL profile given current market characteristics.

2. Market Conditions Since Withdrawals

The simulation is based on the following observed post-2023 market regime:
LDO has shown very high volatility (around 110% annualized) and significantly negative drift (–45% annualized), consistent with its decline from ~$2.12 on May 15, 2023 to ~$0.67 today.
wstETH, by contrast, has had a more moderate volatility (around 63% annualized) and a positive drift of +22%.

Despite the dramatically different long-term trends of the two assets, their daily returns are fairly tightly correlated (0.77). This means the two assets often move in the same direction day-to-day, even though LDO has underperformed heavily over the full period.

3. Impermanent Loss Results

Across the 20,000 simulated price paths, the results consistently show that LPing LDO/wstETH leads to a meaningful structural cost. The average IL over one year was approximately –10%. The median path produced about –6.3% IL, while more extreme downside scenarios reached –30% or worse.

In simple terms:
under current conditions, providing liquidity produces about 10% less value than simply holding an equal-value LDO + wstETH portfolio.

These results are calculated before fees and represent the baseline structural effect of IL, driven mainly by LDO’s high volatility and negative drift.

4. Treasury Implications

Whether this IL is acceptable or not depends entirely on the DAO’s strategic goals.

If the treasury’s priority is to maximize its financial value, LPing is clearly disadvantageous: IL introduces a persistent drag, and the volatility dynamics of LDO strongly bias the position toward losing value relative to passive holding.

If the goal is deflationary pressure, buying LDO and reducing the circulating supply, LPing is also counterproductive. Because liquidity provision continually rebalances the pool, the DAO does not actually remove LDO from circulation: half the position remains in wstETH, and the AMM often sells LDO when its price declines, increasing the amount of LDO in the pool over time.

If the priority is improving market liquidity, however, IL functions as a form of market-making expense. LPing deepens liquidity, reduces slippage, and improves execution for traders. In that context, IL is one of the operational costs of maintaining healthy markets.

Finally, if the goal is to create structural buy-side pressure or to tighten the economic relationship between LDO and stETH, LPing can be acceptable. A buyback-and-LP strategy stabilizes liquidity, increases the presence of LDO–stETH pairing, and supports a more robust market structure. In this framing, IL acts as a strategic cost rather than a loss.

5. Interpretation and Next Steps

The main takeaway is that IL is meaningful in this pair,about a 10% expected drag over a year,but whether that matters depends on what the DAO wants to achieve. A treasury optimization strategy should avoid LPing, while a liquidity or market-structure strategy may justify it.

I am currently extending the model to include fee revenue based on historical volume-to-TVL behavior, along with sensitivity analyses across volatility, correlation, and drift assumptions. I will publish those results in a follow-up post so we can evaluate the net cost or benefit of LPing under different market regimes.

Looking forward to hearing your thoughts and suggestions as I refine the model further.

2 Likes