DAO

What a DAO is operationally, and what that leaves open

The operational description is short. Token holders propose changes, vote onchain, and a contract executes the result. Treasury spending, parameter changes and contract upgrades run through that pipe instead of through a person with signing authority. Nothing in that description says anything about legal form, and that silence is where most of the trouble starts.

Three things get bundled under the one label and they come apart under pressure. There is the governance process, which is code. There is the entity, which is a filing or the absence of one. And there is the control map, which is the list of every key and role that can move funds or change contracts without a vote. A project can have a working governance process, no entity at all, and a core team holding every upgrade key. That combination is common, and it is not what people picture when they hear the word.

Wyoming: the DAO as a limited liability company

Wyoming's 2021 DAO Supplement, codified at Wyo. Stat. Ann. sections 17-31-101 through 116, does something narrower than the headlines suggested. It did not create a new kind of legal person. It makes a DAO a limited liability company whose articles of organization state that it is a DAO, and it lets an existing Wyoming LLC convert into one. The registered name has to contain DAO, LAO or DAO LLC, so the entity announces itself on the public register.⁴

Two provisions matter more than the rest. A Wyoming DAO LLC can be member managed or algorithmically managed, and if it is algorithmically managed the underlying smart contract has to be upgradeable rather than fully immutable. Founders usually read immutability as the property a statute would want. Here it is disqualifying. Second, unless the articles or the operating agreement say otherwise, members owe no fiduciary duty beyond the implied covenant of good faith and fair dealing, and they keep the ordinary limited liability an LLC carries.⁴

That last clause is the entire reason to file. Everything else is paperwork around it.

The Marshall Islands route, and what it shares with Wyoming

The Republic of the Marshall Islands passed its DAO Act as P.L. 2022-50, commencing 25 November 2022 and amended by P.L. 2023-83. The structure is the same idea in a different jurisdiction. A DAO there is defined as a resident domestic limited liability company organised under the Act, formed and managed under the pre-existing Limited Liability Company Act 1996, with a registered name that has to include DAO LLC. The statute names MIDAO Directory Services, Inc. of Majuro as the registered agent.⁵

Put the two statutes side by side and the pattern is obvious. Neither invented an entity. Both took the LLC, which already carries limited liability and decades of commercial case law, and bolted DAO specific rules onto it. That is the good news. The protection you are buying is the protection an LLC has always offered, argued and tested in ordinary litigation, rather than a novel shield nobody has defended yet.

Ooki DAO: what the case actually settled

On 22 September 2022 the CFTC charged Ooki DAO, successor to bZeroX, LLC, with illegally offering leveraged and margined retail commodity transactions, failing to register as a futures commission merchant, and failing to implement a Bank Secrecy Act customer identification programme. The agency's stated theory was that the founders transferred control to the DAO in order to propagate the notion that they could no longer be held responsible for wrongdoing.¹

Judge William H. Orrick of the Northern District of California entered default judgment against Ooki DAO on 8 June 2023, in case number 3:22-cv-05416-WHO.² Before that, in December 2022, the same court authorised service of process on the DAO through its own online discussion forum and the help chat box on its website.³

If you were waiting for a ruling on whether a regulator can reach an unincorporated DAO as a thing, that is the ruling. It can be named, it can be served in ways that did not exist five years ago, and judgment can be entered against it.

And what the case left open

This is where commentary overreaches, so it is worth being exact. Ooki was a default judgment. The DAO did not appear and defend, so the court accepted the CFTC's allegations as true for the purpose of granting relief.² It did not adversarially litigate the outer boundary of individual token holder liability, and a district court default is not appellate precedent.

So the settled part is this: a DAO can be named as a defendant, served through its own forum, and have judgment entered against it as an entity.² The contested part is what happens to an individual holder who did nothing but vote. No appellate court has resolved it. The theories circulating on both sides are arguments, not holdings, and anyone presenting either one as settled law is telling you something the record does not support.

In our view, reading that silence as safety is the mistake founders make. The registered wrappers exist precisely because the unregistered answer is unknown, and an unknown answer that you have priced at zero is still a position you have taken.

Participation: one measured number and a lot of folklore

You will read constantly that most DAOs run under ten percent governance participation. We went looking for the study behind it. There is not one. What exists is Falk, Pathan, Rigas and Tsoukalas's 2024 study, which pulled onchain voting data across 130 Aave proposals and 141 Compound proposals and measured the average share of total token supply cast at 3.2 percent for Aave and 7.7 percent for Compound.⁶ Named protocols, stated method, stated denominator, published and dated.

That is the honest floor of what is known, and it is worth saying plainly what it does not cover. Two protocols is not most DAOs. The denominator is total token supply, which is the harshest of the available denominators and produces lower figures than circulating or delegated supply would. Governance architectures that are not EVM token voting get counted differently again and land nowhere near these numbers. Comparable measured participation data across DAOs does not exist.

Saying so is more useful to a founder than another round number. If you are calibrating a quorum, a proposal threshold or a delegate programme, the only turnout that governs your design is your own. Instrument every vote from the first proposal onward and within a year you will hold the only dataset that applies to you.

What we settle before a client calls it a DAO

Four questions, answered in this order. Which entity, in which jurisdiction, files what. Who counts as a member, and whether becoming one requires anything beyond buying a token on an exchange. What the governance contract can execute without a human. And which keys and roles sit outside the vote entirely, with the blast radius of each written next to it.

Reversed, the pattern is one we see often: a launched token, a forum full of proposals, a treasury multisig held by three founders, and a lawyer engaged in month nine to work out what was created. The wrapper is cheap at the start and expensive later, which is an unusual property for a legal cost and the reason we push it to the front of the schedule.

The wrapper also has to be worth wrapping. A DAO with no revenue underneath it is a governance process allocating a countdown, and no entity choice fixes that. The token coordinates who decides how value gets used. The business is what produces the value.

One caution covering all of the above. Whether a specific structure limits member liability in a specific jurisdiction is a fact specific question for your counsel. This page is reference material for design work. It is not legal advice, and nothing here is a recommendation to buy, sell or hold any asset.