Following extensive research and community discussion, the Core Team is sharing its view of CoW DAO’s evolved thinking on value distribution mechanisms for the COW token. This is based on our understanding, research, and close collaboration with Aragon team under the RFP : CoW Value Distribution Mechanism and Aragon’s proposal.
1. Current Context
CoW Protocol has achieved product-market fit as the largest DEX aggregator by volume, with more than $200B in volume over more than 12M trades on CoW products. Since 2024, when revenue generation started, the DAO generated $41.9M across products like CoW Swap and MEV Blocker, increasing in tandem with market share growth - in general, showing a growing and healthy business.
For generating this traction, CoW DAO has a significant cost element regarding solver incentives - a key differentiating factor to competitors, allowing for better-than-market execution.
This incentive program, together with development costs in this expansion phase, leaves the DAO with positive cashflow, but not sufficient to meaningfully lead to circulating supply reduction, through added value distribution mechanisms.
2.Definition of circulating supply
While Circulating Supply definition is not universally standardized, we agree with Token Terminal’s definition: Circulating supply measures the number of tokens that are currently available for trading on the open market. It excludes tokens that are locked, unvested, or held in the project’s own treasury.
Nonetheless, there are relevant aspects under the CoW Ecosystem, that should be considered:
Note: The above definitions of circulating and non-circulating supply are according to current definition of circulating supply.
- The CoW DAO Safe holds tokens that are long term held - which can only be deployed under governance proposals. Holds ~357.3M COW, with 47.3M COW currently vesting under CIP-83 to the Core Team, over a 4y period. Non-committed COW amount to ~310M COW.
- The CoW Foundation and related corporate entities holds several purpose specific wallets, with different time horizons, mainly:
2.1. CoW Treasury - Holds ~59M COW, including idle COW earmarked for potential future fundraising, OTC or capital deals and Liquidity provision positions.
2.2. CoW Buyback wallet - Holds ~22M COW from market buybacks, and receives longer term buybacks which is used to top up the Payouts safe (Solver rewards budget).
2.3. Payouts safe - Holds ~4M COW as liquid solver rewards budget. - Solver Bonds hold a significant amount of idle COW expected to be used as collateral / capital at risk in case of solver misbehaviour.
3.1. Full bonds - 5 bonds accounting for 7.5M COW: CoW, Gnosis, Rizzolver, Fractal, TSolver
3.2. Reduced bonds - 5 bonds accounting for 2.5M COW
CoW Circulating Supply is currently calculated under GitHub - cowprotocol/cow-token-supply-api · GitHub, which accounts for total supply minus
- The vCOW unvested part (which is now fully vested);
- The CoW DAO Managed Treasury (only Mainnet)
As such, we suggest an update to the circulating supply methodology to the CoWmunity to the following:
CoW Circulating supply as calculated in the public Github [link] measures the number of tokens that are currently available for trading on the open market. It excludes tokens that are locked (under staking mechanisms), unvested, held in solver bonds or in the project’s treasury in any network.
3. The current buyback mechanism - 2y review & emission behaviour
Prior buyback update: COW token buyback - An update on 1 year of execution
Official buyback query: https://dune.com/queries/5273980/8661207
COW buyback start date: 2024-04-17
Figure 1 - CoW Buybacks per quarter
Left axis denotes amounts in the period in COW. Right axis denotes cumulative amounts in COW
| Solver emissions since start date | 66,588,732 |
|---|---|
| Tokens buybacks since start date | 78,616,859 |
| Net emissions | -12,028,127 |
Table 1 - Solver mechanism emissions from start of buyback period(calculated on May 8th 2026)
The Buyback operation has migrated from the Treasury team to the Core Team as per KPK’s communication on Nov 2025, keeping the same target of buying 120% of solver emissions from the open market, ensuring the Protocol operations are not inflationary.
However, the DAO still performs other activities which lead to COW emissions, namely:
- Core Team grant - CIP-83: Renewing Team Grant Allocation with Performance-Linked Milestones
- Grants DAO program - CIP-82: CoW DAO Grants 2026 Renewal
Note: the emissions related to the TGE (vCOW related) have been fully vested in Feb, 12th 2026, marking the end of a significant emission schedule for investors, the Core Team initial allocation, advisors, etc.
For more detailed simulation on the forecasted emission schedule, which is dependent on the two emission sources above and on a percentage of CoW DAO revenue, we refer to Aragon’s modelling here.
4. What we can do different in terms of token initiatives
Our approach centers on improving the fundamental demand/supply equation for COW through two parallel tracks:
- Supply side: How to decrease circulating supply.
- Demand side: How to increase token demand.
This equation is distinct from protocol fundamentals. CoW Protocol can be trading billions in volume and improving revenue rates without that translating into token attention. Improving the demand/supply equation requires deliberate mechanism design, not just protocol growth. Nonetheless all initiatives need to be funded by the protocol’s “Net Profit” (Fees minus Solver Incentives and OPEX) adjusted for any capital expenses (Free Cash Flow).
4.1. Conclusions from Benchmarking
Key Principles:
- Avoid dilution. Unchecked emissions are effectively fundraising in the market at the expense of existing holders.
- Token demand must be structural, not mercenary.
2.1. If demand is derived from the token having a use - it scales with protocol expansion.
2.2.If demand is derived from distributions - it is mercenary and will leave when incentives drop.
Drawing on Aragon’s benchmarking here and internal studies, we tangentially agree with:
- VE tokenomics appear to work for some liquidity-hub projects (Curve/Convex, Velodrome) but do not reliably correlate with token price (see CRV, BAL), as they can lead to over-emission.
- Narrative drives price above anything else. Of the protocols benchmarked, the three (Hype,GMX,Sky) with clearly successful outcomes share one trait: a simple, legible, revenue-funded mechanism executed at meaningful scale.
We propose to the community, three initiatives:
-
Properly enforce the “HODL rule” for solver rewards - this is based on an independent CIP on solver bond rules (to be shared).
1.1. As per current rules, solvers are mandated to keep 25% of the received rewards - this is a soft rule, given that the entirety of the rewards is transferred to the solver.
1.2. Move from 25% to 20% and direct that amount to the solver bond, ensuring:
1.2.1. The capital accumulation by the solver;
1.2.2. An increase in bond size leads to more solvers being under an independent full bond, thus reducing the risk for the DAO;
1.2.3. Exposing solvers to the success of the DAO and reducing arbitrary selling pressure. -
Reduce the overhang of non-circulating supply
2.1. Under the proposed definition, all CoW Foundation held tokens are non circulating supply, meaning that bought back tokens, tokens held by the Managed Treasury and the DAO safe are non-circulating.
2.2. Perform a trial until Dec 2026, where for each token emitted by the DAO’s operation (for solver rewards, team compensation or grants), the equivalent amount is burnt from the DAO Safe.
2.3. This would lead to a 1:1 burn of future optionality. If the protocol grows significantly, significant burns are performed, reducing the potential for further token expansions in the future, until all COW is distributed.
2.4.An estimated 60M-85M CoW would be permanently removed from future usage and reduce the total token supply.
2.5. This would allow for narrative and message signaling, without compromising the current optionality at the operational level. -
Implement a flexible buyback mechanism
3.1. Grant a mandate to the Core Team to make buybacks flexible according to price movements of COW and ETH, and overall Gross Margin / Fee Rates - to be assessed on a quarterly basis.
3.2. Allow extra buybacks (up to 100% of weekly revenue) to be executed at:
i. Below $0.20 COW prices;
ii. Above $3,000 ETH prices;
iii. When Weekly Gross Margin (Protocol fees - solver rewards) is above an average $500,000 for over 4 consecutive weeks
iv. When Fee Rate (Protocol Fees / Volume) is above 5 bps for over 4 consecutive weeks
3.3. Allow for a reduction of buybacks up to 100% of solver rewards when any combination of 3 out of 4 of the conditions above is not met.
Initiatives that we shortlisted but didn’t consider appropriate to implement:
- Blindly increasing buybacks to cover full emissions
- This would require increasing buybacks to cover Grants DAO emissions and Team related emissions.
- Given the majority of said emissions being vested, and the selling rate of grantees and team members in itself being either small or deferred in time, this would lead the DAO to upfront capital investment to reduce COW circulation in the expectation of future selling pressure from new tokenholders.
- We are of the opinion that the DAO benefits more from allocating the extra capital / FCF to growth and expansion initiatives than growing the revenue engine.
- Given the bear market, this would lead the DAO to fund new growth initiatives in future years from treasury reserves instead of revenue generation, putting the DAO at risk of being forced in the future to sell significant number of tokens in a “desperate” capital situation.
- Community funded-bonds - Allow stakers to provide their COW to solvers, improving automation and decentralization of the whitelisting process
- While this would give a direct incentive to stakers and add token utility it would concentrate solver execution risk on the staker side (even if solvers had to contribute a % of capital)
- It would expose the protocol to bad actors who could deposit tainted/stolen assets as bonds, steal contract buffers, and thereby launder their funds
- It would lead to a revenue cut for solvers if implemented similarly to 1inch’s model, causing drawdowns as identified in Aragon’s study
- Change solver rewards to payment in native token
- We keep the same opinion on the mechanism as here, in essence:
- We see paying in COW as incentive aligning - as Solvers benefit from CoW’s future growth.
- Current buyback has had positive PnL on the buyback operation.
- Would lead to the reduction of buybacks as supply expansion would be lower.
- Given that not all solvers sell the entirety of their rewards (from internal tracking of some solver wallets, opposed to the sampling performed by Aragon), this would likely lead to a net decrease in buyback volume.
- We keep the same opinion on the mechanism as here, in essence:
- Surplus top up - Allocate a governance specified % (e.g. 50%) of net profit of the DAO to a trading incentive program.
- This mechanism would be overly complicated versus Aragon’s proposed Buyback-and-distribute.
- We are not certain of the utility impact of such an initiative, and think that the implementation of such a mechanism would outweigh the benefits.
We intend for this post to open a discussion, prior to proceeding with a feedback based CIP that would also take into account Aragon’s study and suggested Buyback-and-Distribute above a hurdle of FCF ($10m).