SNAPSHOT · 21 Sep, 18:08 UTC

Ratings · Graded on the public record

Who manages DAO treasuries and token liquidity — and how openly

A DAO treasury or liquidity manager runs assets it does not own, under a mandate from a DAO or a token issuer: diversifying a treasury, deploying protocol-owned liquidity, making markets or managing AMM positions. We grade each firm A to F on whether the mandate is public, the money disclosed, and the work reported.

WORKS INSIDE THE CLIENT’S SAFE

ENS sets its permissions by on-chain vote, it never takes custody, and its Endowment reports have been public every month since 2023.

MANAGER CANNOT WITHDRAW

Gauntlet runs the strategies as guardian, but the vault owner sets what it may do, and it cannot move funds out.

LEAST ON THE RECORD
The market makers

None of the five publishes an amount under management or any public reporting on what it does with a client’s tokens.

What we did · September 21, 2026We looked for each firm’s mandates in DAO governance records and in announcements from both sides, read what each discloses about the money it manages and who can move it, and checked what anyone outside the client can see — then graded three blocks from A to F, with every fact sourced and dated.

Researched by the ChainWatch Daily ratings deskLast checked How this is fundedSomething wrong? Tell us

Custody is the dividing line

The six firms graded highest never take custody of what they manage: they act through Safes and vaults whose permissions are limited and visible on-chain.

The market makers work the other way round. Wintermute, GSR and Kairon Labs describe deals in which the client lends them its tokens; Flowdesk keeps client assets in its own Fireblocks workspace, and in 2025 a TreasureDAO contributor proposed withdrawing an idle $785,000 held there. That is how market making usually works, and it is legitimate. It also means the client is trusting a company, and none of the five publishes what it does with the tokens.

The ranking

#NAMEGRADE
1
AAeraVault protocol · run by Gauntlet
A
Xai announced the mandate itself, the manager cannot withdraw funds, and vault documentation and dashboards are public.
Theo Walsh
MandateA
MoneyA
Transp.A

Strengths

  • A client announced the mandate itself
  • The manager cannot withdraw funds; its allowed actions are set by the vault owner
  • Vault documentation and dashboards are public

Weaknesses

  • Aera is the vault; the manager is Gauntlet, so the two are graded together
  • Its focus has moved towards earn vaults for fintechs and institutions
  • No result found for the Arbitrum treasury proposal
Read the full review & evidence →
2
KkpkTreasury manager · formerly karpatkey
A
ENS sets its permissions by on-chain vote, it never takes custody, and anyone can read its monthly Endowment reports.
Priya Nair
MandateA
MoneyA
Transp.A

Strengths

  • Its permissions over the ENS Endowment are set by on-chain DAO votes
  • Non-custodial: it acts inside the client’s own Safe, with restricted roles
  • Monthly public reports for the ENS Endowment since 2023

Weaknesses

  • Two firm-wide size figures on its own site, $8bn+ and $300m+, measure different things and are not labelled clearly
  • Public reporting found for ENS only; most listed clients are named by kpk alone
  • Its Aave treasury role appears to have ended in December 2024, while Aave stays on its client list
Read the full review & evidence →
3
TTokenLogicTreasury manager
A
Aave DAO has renewed its mandate by vote, and it moves treasury funds through contracts that deny it custody; Aave is its only documented client.
Ethan Brooks
MandateA
MoneyB
Transp.A

Strengths

  • Mandate renewed by Aave DAO vote, with terms and payment published
  • Moves treasury funds through Steward contracts, without taking custody
  • Maintains a public dashboard of Aave’s finances

Weaknesses

  • One documented client: Aave
  • No current figure for the treasury it works on; the last one we found is from November 2023
  • A delegate objected in 2026 that the team preparing financial proposals would also maintain the tool that validates governance proposals
Read the full review & evidence →
4
SSteakhouse FinancialTreasury manager & vault curator
A
Compound’s treasury committee chose it to curate up to about $12M, its vaults leave the DAO holding the shares, and it publishes monthly Arbitrum reports.
Theo Walsh
MandateB
MoneyA
Transp.A

Strengths

  • Non-custodial vaults with separated curator, sentinel and guardian roles
  • Monthly public reports for programmes it runs for DAOs
  • Discloses total assets supplied to its vaults

Weaknesses

  • DAO treasury work is now a side line next to vault curation for fintechs and exchanges
  • No completed vote found for the Compound appointment
  • The $4.34B figure is undated
Read the full review & evidence →
5
AAvantgarde FinanceTreasury manager
B
Nexus Mutual documents Avantgarde executing its vault allocations; the firm’s homepage gives two different treasury totals and links to no public reports.
Ethan Brooks
MandateA
MoneyB
Transp.B

Strengths

  • The client keeps custody; Avantgarde manages as a delegate on the client’s Enzyme vault
  • A client, Nexus Mutual, voted on its proposal and documents the relationship itself

Weaknesses

  • Two different treasury totals on the same homepage, both undated
  • No regular public reports found
  • Homepage logos do not distinguish treasury clients from other relationships
Read the full review & evidence →
6
RRe7 CapitalVault curator · Re7 Labs
B
Its vaults are public and non-custodial, but no DAO has voted it a treasury, and it absorbed Stream Finance bad debt in 2025.
Priya Nair
MandateC
MoneyA
Transp.A

Strengths

  • Curates without holding deposits; a guardian can veto changes
  • Every vault is public on-chain
  • Value managed is tracked independently

Weaknesses

  • No DAO treasury mandate on the record; issuer roles are curation of public vaults
  • Took about $27.4M of bad debt in the November 2025 Stream Finance collapse
  • The firm-wide “$1B+ Deployed” is undated
Read the full review & evidence →
7
AArrakis FinanceOn-chain market maker for token issuers
B
The issuer owns each vault and can withdraw at any time, but Arrakis publishes no reports on what it does inside them.
Theo Walsh
MandateB
MoneyA
Transp.D

Strengths

  • The issuer owns the vault and can withdraw at any time
  • A DAO treasury holds Arrakis-managed vaults
  • Value managed is tracked independently

Weaknesses

  • No public dashboard or reports on what it does in a client’s vault
  • Two DAO proposals naming it were not enacted
  • An Arrakis vault lost 2.94 WETH to an exploit in August 2026; whether users were made whole is not reported
Read the full review & evidence →
8
SSteer ProtocolAutomated liquidity manager
B
Its clients appear only as testimonials on its own site, and it does not document what its strategists can do with vault funds.
Ethan Brooks
MandateC
MoneyB
Transp.B

Strengths

  • Non-custodial vaults
  • Public vault listing
  • Value managed is tracked independently

Weaknesses

  • Clients named only in testimonials on its own site
  • Strategist and keeper permissions are not documented
  • Vault value has roughly halved since September 2025
Read the full review & evidence →
9
IICHIAutomated liquidity manager
B
Its vaults are non-custodial with no privileged rebalance function, but its only DAO case study does not name the DAO.
Theo Walsh
MandateD
MoneyA
Transp.B

Strengths

  • Non-custodial vaults with no privileged rebalance function
  • Value managed is tracked independently

Weaknesses

  • Its DAO case study does not name the DAO
  • Vault value has fallen from $35.4M to $8.1M in a year
  • A 2022 bad-debt episode in an ICHI-run lending pool sent its token down about 90%
Read the full review & evidence →
10
KKairon LabsMarket maker
C
Elastos’s council voted to appoint it in 2022, but where borrowed client tokens are held is not public, and reports go to clients only.
Priya Nair
MandateA
MoneyD
Transp.C

Strengths

  • A DAO council voted to appoint it
  • Weekly liquidity dashboards for clients

Weaknesses

  • No amount under management, and custody of borrowed tokens not described
  • Its own figures vary: founding year 2017 or 2018, 300 or 500+ clients
  • Reports go to clients only
Read the full review & evidence →
11
GGamma StrategiesAutomated liquidity manager
C
No DAO mandate is on the record, its vault value has fallen 94% since January 2025, and Gamma estimated 25–40% recoveries after its 2024 exploit.
Priya Nair
MandateD
MoneyB
Transp.B

Strengths

  • Non-custodial vaults with open withdrawals
  • Public vault listing, with rebalances visible on-chain

Weaknesses

  • No DAO mandate over treasury assets on the record
  • An exploit in January 2024 took an estimated $3.4M–$6.2M; Gamma estimated 25–40% recoveries
  • Vault value has fallen about 94% since January 2025
Read the full review & evidence →
12
KKeyrockMarket maker
C
Radix confirms the engagement, while loan terms are private and reports go only to the client, so the public record ends at the announcement.
Priya Nair
MandateB
MoneyD
Transp.C

Strengths

  • A client-side announcement confirms the Radix engagement
  • Trading statistics for clients on demand

Weaknesses

  • No amount managed or volume disclosed
  • Loan terms for client tokens are not public
  • Its Morpho vaults hold its own idle USDC, so they say nothing about client mandates
Read the full review & evidence →
13
FFlowdeskMarket maker
C
SG-FORGE names it as market maker on its own site; Flowdesk holds client assets itself, segregated, and publishes nothing about client reporting.
Ethan Brooks
MandateA
MoneyD
Transp.F

Strengths

  • A regulated client names it on the client’s own website
  • Authorised under MiCA and licensed by VARA in Dubai
  • Client assets are segregated from the firm’s own

Weaknesses

  • Client assets are in Flowdesk’s custody, not the client’s
  • No amount under management disclosed
  • Nothing published on how, or whether, it reports to clients
Read the full review & evidence →
14
GGSRMarket maker
D
It proposed its token-loan terms to Stake DAO in public in 2022, with no vote recorded, and publishes no reporting on client tokens.
Ethan Brooks
MandateB
MoneyD
Transp.F

Strengths

  • Its deal terms were proposed in public, on the DAO’s own forum
  • Operating since 2013

Weaknesses

  • Client tokens are lent to GSR outright
  • No confirmed vote on its Stake DAO proposal
  • No public reporting on client mandates
Read the full review & evidence →
15
WWintermuteMarket maker
D
Its DAO market-making proposals are public, the tokens are lent to it outright, and we found no public reporting on what it does with them.
Theo Walsh
MandateB
MoneyD
Transp.F

Strengths

  • Deal terms with DAOs have been proposed in public, on governance forums
  • Liquidity across more than 60 venues

Weaknesses

  • Client tokens are lent to Wintermute outright, with an option to pay cash instead of returning them
  • No public reporting on what it does with a client’s tokens
  • Lost about $160M from its DeFi operations in a September 2022 hack; it said it stayed solvent and OTC was unaffected
Read the full review & evidence →

Open any row for the verdict and the quick read, or a firm’s name for every fact behind its grades, with sources. Last checked September 21, 2026. How we grade.

HOW WE GRADE THIS — THREE BLOCKS, A TO F+
BlockWhat we checkACF
MandatemandateIs there public confirmation of the firm’s authority over the client’s assets?A DAO vote, or an agreement named by both sidesThe client is named only by the managerNo clients are named
MoneymoneyIs the amount under management disclosed, and is it described how the firm reaches the assets?Amount disclosed; management is non-custodial with restricted permissionsAmount disclosed; access is not describedNothing is disclosed
TransparencytransparencyCan the client and the public see what is done with the assets?Regular public reports, or an on-chain dashboardReports go to the client onlyNo reporting
  • B and D are given when the condition for the grade above is partly met.
  • For sorting, A = 4, B = 3, C = 2, D = 1, F = 0. The overall grade is the average of the three blocks, rounded to the nearest letter, halves up. Only the letter is shown.
  • Two firms with the same average are ordered by their mandate grade, then alphabetically.
  • Under every firm is a one-sentence verdict, written from the facts on its card and signed by a member of the desk.

A fact we could not source is recorded as not published and graded as missing, never assumed — read the methodology.

How to choose a DAO treasury or liquidity manager

Start with custody, then check the public record. A good DAO treasury manager works inside a Safe or vault the DAO owns, under permissions the DAO has voted on, and reports in public on what it did with the money. Of the 15 firms in this rating, four meet that closely enough to grade A overall: Aera, kpk, TokenLogic and Steakhouse Financial. At the other end, none of the five market makers publishes an amount under management or any public reporting on what it does with a client’s tokens.

Treasury managers, vault protocols and market makers are different products

Treasury managers act on assets that never leave the client. kpk holds “strictly limited permissions within client-owned Safes”; TokenLogic moves Aave’s funds through Finance Steward contracts that stop it taking custody; Avantgarde Finance is a delegated manager on an Enzyme vault the client owns.

Vault protocols and curators keep capital in contracts and let the manager rebalance within limits. Aera’s guardians cannot withdraw funds; an Arrakis vault belongs to the token issuer, who can withdraw at any time; ICHI vaults have no privileged rebalance function.

Market makers usually take the tokens. Wintermute, GSR and Kairon Labs describe deals in which the client lends them its tokens, and Flowdesk keeps client assets in its own Fireblocks workspace. That is the right tool for exchange liquidity, but it is a different relationship: the DAO is trusting a company rather than granting a permission.

What counts as a real mandate

The strongest evidence is the client’s own governance. ENS DAO puts kpk’s permissions over its Endowment to on-chain votes; Aave governance has renewed TokenLogic by name; Nexus Mutual members voted for NMPIP 196, staking 6,624 WETH through Kiln on Avantgarde’s proposal; and Elastos’s Cyber Republic council approved Kairon Labs as market maker, 11 of its 12 members in favour.

A proposal with no recorded result is weaker. Wintermute’s 2021 offer to Index Coop and GSR’s 2022 offer to Stake DAO are both public, and neither page records a vote. Weakest of all is a client named only by the firm: Steer Protocol’s clients appear as testimonials on its own site, and ICHI’s DAO case study does not name the DAO.

Why market makers grade lowest

Because the standard deal is a token loan with an option attached. Under Index Coop’s IIP-83, Wintermute would borrow 100,000 INDEX for a year and then either return it or pay a fixed $5M or $7.5M; GSR proposed borrowing 1,250,000 SDT, about 3% of Stake DAO’s supply, on similar terms. The tokens sit with the market maker for the length of the deal, none of the five discloses how much it manages, and none publishes what it does with a client’s tokens. A DAO hiring one should write the reporting into the contract, because it will not appear otherwise.

How much money these firms manage

Where the assets sit in public contracts, the size can be checked. Value in each firm’s vaults on 21 September 2026:

  • Aera: $249.1M
  • Re7 Labs: $90.5M
  • Arrakis: $60.1M
  • Steer: $21.9M
  • ICHI: $8.1M
  • Gamma: $3.3M, down from $58.8M at the start of 2025

The ENS Endowment that kpk manages stood at $67.7M on 30 June 2026. Totals the firms publish themselves are harder to use: kpk’s site shows “$8,000,000,000+” deployed on one page and “$300m+” on another, Avantgarde’s homepage gives both “more than $60M” and “more than $90M”, and Re7 Capital’s “$1B+ Deployed” carries no date.

What the grades do not tell you

They grade what is on the record, not how well the money was run. Returns, fees against results and execution quality are not graded, because almost none of it is published. Losses are recorded on each firm’s page with their sources instead: an exploit at Gamma Strategies in January 2024, after which Gamma estimated 25–40% recoveries; about $27.4M of bad debt in Re7 Labs vaults after the Stream Finance collapse in November 2025; and about $160M taken from Wintermute’s DeFi operations in 2022.

Questions to ask before you hire one

  • Will the assets stay in a Safe or vault the DAO owns — and if not, who holds the keys?
  • Which permissions does the manager get, and can the DAO revoke them without the manager’s help?
  • Will the appointment go to a DAO vote, with the fee, term and termination notice in the proposal?
  • Will reports be public, how often, and where will they be posted?
  • For a market maker: are the tokens lent, at what repayment terms, and what reporting is in the contract?

What changed since last time

  • 2026-09-21Category published: fifteen firms graded, one held back.Treasury and liquidity managers publish almost nothing to measure, so this rating grades the public record instead — the mandate, the money and the reporting — rather than inventing a 0–10 score.

Questions

What does a DAO treasury manager do?+

It runs assets it does not own, under a mandate from a DAO or a token issuer: diversifying a treasury out of the project’s own token, placing stablecoins in lending or staking, deploying protocol-owned liquidity, or managing liquidity positions on exchanges. The better arrangements leave the assets in the DAO’s own Safe or vault and give the manager only the permissions the DAO voted for — kpk operates that way for the ENS Endowment, and Aera’s vaults stop the manager withdrawing funds at all.

What is the difference between a treasury manager and a market maker?+

Custody, mostly. The treasury managers and vault protocols at the top of this ranking work inside wallets or vaults the client owns. The market makers work differently: Wintermute, GSR and Kairon Labs describe deals in which the client lends them its tokens, often with the option to pay cash instead of returning them, and Flowdesk holds client assets in its own Fireblocks workspace. None of the five market makers here publishes an amount under management or any public reporting on client tokens.

Why letter grades instead of a score out of ten?+

Because these firms publish almost nothing we can measure. Our other ratings score measurable things — a fee, a spread, a reserve ratio — on 0 to 10. Here the evidence is a public record: a vote, an amount, a report. Grading that as “7.4/10” would invent a precision the evidence does not have, so each firm gets A to F on three blocks, and the overall grade is their average, computed rather than chosen.

How can I check a manager’s mandate myself?+

Look for the client’s side of the record. A real mandate usually leaves a trail in the DAO’s own governance — a Snapshot or Tally vote, or a forum proposal naming the firm, the amount and the fee — or in an announcement the client itself publishes. A client named only on the manager’s own website, or a case study that does not name the client at all, tells you much less. Every mandate we cite links to the page we read and the date we read it.

Can a firm pay to be included or to improve its grade?+

No firm on this page paid anything. The site can mark an entry as a sponsored inclusion, and if it ever does, the label appears in the table and on the firm’s page and its website link is marked as sponsored — but the grade and the position come from the same rubric and the same arithmetic as every other firm. Payment can change a label; it cannot change a grade.

How this is funded

It is not. There are no affiliate links on this page or anywhere on this site, no paid placements, and no sponsored positions. Nobody in this table can buy a place in it or influence a grade — and none of them paid us anything. The full policy.