Lido DAO exposes token flaw and forces a budget review
Lido’s DAO is debating whether its 2027 grant budget should guide automated purchases of the LDO governance token.
The Oct. 7 discussion puts two choices on the table: link NEST’s revenue threshold to approved spending, or periodically clear the buyback engine’s negative spending counter.
NEST had made no purchases funded by its allocator when checked on Oct. 13. Genuine cost cuts could leave more income available for buying, while a reset would only remove past constraints.
Governance contributor Kuzmich proposed the budget link on Oct. 7. In an Oct. 9 response, Jenya_K supported considering it in the 2027 grant request, but said the amount must come from bottom-up planning.
Kuzmich’s Oct. 11 follow-up corrected the treatment of staking rebates and tested profits without assumed treasury income.
What the counter measures
Under Lido’s implemented specification, NEST compares measured staking revenue with a daily baseline of $109,589, approximately $40 million a year. Half the difference feeds a running spending counter.
Revenue below the baseline subtracts from that counter, and stronger revenue must repair the shortfall before purchases can start. A treasury can hold assets while the counter remains negative, and a positive counter would not by itself establish accounting profit.
The latest captured checkpoint on Oct. 13 recorded a counter of minus $564,230.96 and a skipped allocation.
Separately, the allocator held about 41.12 stETH when checked on Oct. 13. Its complete indexed transfer history showed one 41-stETH top-up on Aug. 28 and no outgoing trans
Any future spending must satisfy sufficient counter capacity and funded stETH. It also faces a $50,000 daily cap, a $10 million cap over fixed 365-day windows, and a $1,000 allocation minimum.
Lido’s separate discretionary stETH-to-LDO acquisitions are a different program, and the community reconstruction identifies an August trade as a route test.
NEST does not currently capture monthly Community Staking Module rebates or Earn income. Jenya_K said CSM rebates are already included in financial reporting, so adding them again to reported net revenue would double-count them.
Lido’s unaudited accounts for the first half show $15.7 million in net staking revenue plus $230,000 from Earn, producing $15.9 million in net DAO revenue. Foundations expenses of $14.3 million left an operating surplus of $1.6 million, and a $6 million Kelp-related one-off turned the total result into a $4.4 million loss.
The report separately places $1.4 million of treasury income and a $60.9 million asset-price effect in its treasury bridge, outside the profit-and-loss statement. Jenya_K also warned that treasury operations can include losses.
The 2026 grant request presents a $43.8 million baseline, and the first-half report references $41 million and projects full-year expenses of $37.7 million. The cited materials do not explicitly reconcile the two baseline labels, and none is interchangeable with NEST’s $40 million revenue threshold.
The budget model needs an expense correction
Kuzmich’s Oct. 11 model uses 1,000 ETH price paths over 12 months from Oct. 1, beginning at $2,679.84 with 61% annual volatility and no expected arithmetic price drift. Its purchase frequencies are not measured probabilities of what will happen next.
In the published Oct. 1 simulations under the current rule, mean purchases were about $2.3 million, but the median was zero. Stronger ETH paths lift the average even when more than half the simulations produce no purchases.
Lower budget-linked thresholds produce larger purchases in that model. However, its profit comparison contains a consequential mismatch: the forum says actual expenses match the approved budget, while the code scales a forecast expense curve by each budget divided by $43.8 million.
That curve implies about $39.8 million of expenses across the modeled year before scaling, and only $27.3 million in the row labeled $30 million.
The table below holds the published revenue paths and purchase rules constant, but sets actual expenses equal to each threshold. Profit excludes treasury income and is measured before purchases.
| Annual threshold and actual expenses | Mean modeled purchases | Median modeled purchases | Mean modeled profit before purchases |
|---|---|---|---|
| $40 million | $2.32 million | $0 | −$1.92 million |
| $35 million | $3.35 million | $1.26 million | $3.08 million |
| $30 million | $4.71 million | $3.18 million | $8.08 million |
The expense correction leaves the purchase amounts unchanged because this version of NEST buys against the revenue threshold, without checking actual expenses.
At $30 million in actual expenses, mean modeled profit excluding treasury income falls from the published $10.7 million to $8 million. Among paths with purchases, the fraction ending in a loss rises from 12.6% to 22.2%.
A separate CryptoSlate sensitivity starting Oct. 13, assuming continued funding and using the refreshed counter and explorer ETH quote but unchanged September fee volumes and first-half deduction and Earn assumptions, produced a $2.1 million median for the $30 million case.
The economic direction remains clear within these assumptions: real spending cuts improve the modeled profit balance. Lowering the contract’s threshold alone permits additional buying even when the DAO’s broader operations finish at a loss.
A reset loosens the spending brake
An annual reset would discard a negative counter at a calendar-year boundary while retaining positive capacity.
CryptoSlate’s reset comparison corrects a period overlap in the published script: whole-month purchase totals count September 2028 in two adjacent 365-day profit periods.
That recalculation gives average spending of about $2.6 million per modeled 365-day period with carry-forward and $2.8 million with annual resets, resulting in a roughly 8% gain. Quarterly resets produce about $2.9 million.
With treasury income excluded, periods ending in a loss account for about 19.5% of purchase-making periods under carry-forward, compared with 30.0% under annual resets. Their shares of purchase dollars are much smaller, about 3.5% and 4.4%, respectively.
The result supports a limited conclusion: a reset cannot make the DAO more profitable, and its modeled spending effect is smaller than the budget changes.
The existing allocator has no direct reset function. Its Voting-controlled threshold setters preserve the accumulated counter and apply changes prospectively, so a recurring reset needs a different implementation or governance arrangement.
The 2027 grant process would need to set a feasible spending budget and its corresponding threshold, account for additional revenue without double-counting, and fund any permitted purchases.
The treasury-only design returns purchased LDO tokens to the DAO, with no automatic burn or holder payout. Jenya_K reported no specific plans for the acquired LDO, burns remain a discussion proposal, and liquidity mode requires a separate DAO decision.
Lido’s report treats token acquisitions as asset swaps or purchases without an immediate profit-and-loss effect, with LDO excluded from headline treasury value. More buying can change the treasury’s asset mix without establishing stronger earnings, guaranteed price support, or better ETH staking returns.
That leaves the grant request with a concrete economic test: whether net income improves after actual costs, and whether the acquisition policy preserves those gains.