Hi all,
NEST has been live since 10 August (launch thread) and has not bought a single LDO yet. Its budget counter stood at −$542,451 at the end of September and keeps falling on most days. I spent the last few weeks building a full model of the mechanism to understand whether this can be fixed by tuning parameters. Parameter tuning alone does not solve it. The model points instead to the budget and to how profit is defined as the main levers. Below is what I propose, followed by the reasoning. The full report, model and data are linked at the end.
What I propose
Step 1. Now, through votes, no contract changes
1. Cut the Foundations budget for 2027 materially and, in the same vote, set the NEST threshold equal to the approved budget. I am asking for around $30M, about 30% below the current $43.8M. In the model, buybacks start to grow meaningfully from a cut of about 20% (roughly $35M), and a 30% cut roughly doubles them compared with today. The approved Foundations budget would serve as the operating baseline, not as a claim that it captures every DAO outflow; material off-budget spending, such as the $6.06M Kelp-related costs, would need to be reflected separately.
2. Change the threshold only together with the budget. Any increase in spending, during the year or in the next budget, comes as a separate proposal that explains what the money is for and what it delivers, and the threshold rises with it.
3. Keep the surplus share at 50% until a true-up against actual expenses is in place (item 8 below).
4. Publish budget requests with a breakdown by main line items and an estimate of NEST buybacks at that level of spending, and publish quarterly expense reports faster with the same breakdown.
5. Give NEST its own Easy Track top-up program, separate from the limit used for expense payments.
6. Do not use NEST-bought LDO for contributor payouts. Otherwise distributions eat the buybacks. Whether the tokens are burned or locked is a separate decision for the DAO.
Step 2. In parallel, in development
7. Connect additional DAO income that NEST does not currently capture. Treasury income and Earn appear to add roughly $3.3M a year at the H1 2026 run rate. CSM rebates may add further income, but their accounting treatment relative to reported net staking revenue should be confirmed first. The threshold would also need adjusting for referral payouts, which NEST counts before deduction.
Step 3. Later, if the DAO decides
8. A new allocator with a quarterly true-up against actual expenses. It would keep accruing 50% of the daily surplus, then replace that with the actual quarterly result once the report is out: 75% of a quarterly profit goes to buybacks, and a quarterly loss goes into the deficit in full. This is the point where the share of real profit going to buybacks can rise above 50%.
Why
NEST is not buying because staking revenue is below the $40M reserve. With the current module set, the threshold is covered only when ETH is above roughly $2,668, and every weaker day adds to the deficit. The four days between activation and connecting the revenue source added $147,658 to it, but that is a secondary factor.
The accumulated deficit makes this worse. Every dollar of it has to be earned back before NEST can buy anything. At a constant ETH price of $2,700, roughly where it is now, the model has the first purchase 838 days away, about 2.3 years. At $2,680 it is about 6.3 years, and below roughly $2,668 the deficit only grows. Across random ETH price paths, NEST makes at least one purchase in the next 12 months in only about half of the scenarios. ETH rallying would fix this, but that is not something the DAO controls. The one lever it does control, without weakening the protection the mechanism is built on, is expenses. With the threshold at a $30M budget and ETH at $2,700, the same deficit would be paid off in about 40 days.
Tuning the numbers in the current contract does not solve this. A lower threshold or a higher share buys more, but it also raises the share of purchases made in years the DAO ends at a loss. A $30M threshold with a 100% share at current expenses gives about $6.9M of buybacks a year while the DAO’s profit before buybacks is only about $1.3M, and more than half of those purchases land in loss-making years. That is spending the treasury, not distributing profit.
The only sustainable way to increase buybacks is more income or lower expenses. But today savings never reach buybacks, because the threshold is a fixed $40M unrelated to the DAO’s actual budget. Cut expenses by 10%, 20% or 30% and NEST still buys about $2.15M a year in the model. Once the threshold equals the approved budget, that changes:
| Expenses | Buybacks, threshold = budget | Share of purchases in loss-making years | DAO profit before buybacks |
|---|---|---|---|
| as now ($43.8M) | $1.58M | 5% | about $1M |
| −10% | $2.25M | 6% | about $5M |
| −20% | $3.14M | 8% | about $9M |
| −30% (about $30M) | $4.28M | 5% | about $13M |
A 10% cut barely changes buybacks; the effect becomes meaningful from about 20%. Buybacks grow here because there is real profit left after expenses, not because NEST splits the surplus more aggressively.
The 50% share matters for the same reason. The current contract sees only the approved budget, not actual spending, so a 100% share leaves no buffer if expenses run over or revenue falls during the year. At a $30M budget, 100% would give about $5.63M of buybacks but more than double the share of purchases in loss-making years. With a true-up in place, the report suggests 75% of actual profit rather than 100%.
On top of that, the allocator currently holds about 41 stETH, roughly two days of purchases, so even a positive budget would run dry quickly without a top-up rule.
How I checked this
I reproduced the allocator logic from the verified contracts and matched it against 32 on-chain budget recalculations to the dollar. I rebuilt the treasury’s staking fee series since January 2024 from rebase events and Chainlink prices at each block and reconciled it with Lido’s financial reports. On that series I ran the current rule and dozens of alternatives on history and 12 months forward across thousands of ETH price paths, plus stress scenarios with expense overruns, one-off losses and late reports. The report goes through 19 questions in detail.
Why now
Last year the budget request was posted in early December and voted on mid-month. If the threshold follows the budget, the 2027 request is the natural place to agree on it, ideally in the same vote. I would rather discuss this before the request is drafted than after.
Questions for the team
1. Why does NEST not count CSM rebates and other treasury income, and are revenue sources planned for them?
2. Are CSM rebates included in reported net staking revenue, and how much were referral payouts in H1 2026?
3. How is NEST meant to be topped up, and why does it share an Easy Track limit with expense payments?
4. What is planned for the purchased LDO, and will the liquidity mode be used?
5. Was it expected at launch that the mechanism would build a deficit for four days with no revenue?
6. Could quarterly expense reports come out faster and with a breakdown by main line items?
7. Which vote and which role would change the threshold parameter?
Materials
Full report: lido-nest-analysis/NEST_report.md at main · 0xkzmch/lido-nest-analysis · GitHub
Model, scripts and data: GitHub - 0xkzmch/lido-nest-analysis · GitHub
All figures are model scenarios under the revenue, expense and market assumptions described in the report, not forecasts. The model does not try to predict the effect of buybacks on the LDO price, and the revenue base reflects the current set of staking modules. Happy to walk through any part of it and to rerun the numbers with other assumptions.