How DAOs Work: Three Governance Models and a Legal Landscape Still Taking Shape

Decentralized Autonomous Organizations (DAOs) operate through software-enforced rules, with governance models like delegation, automatic execution, and multisig signing defining how decisions are made and executed. The legal status of DAOs is rapidly evolving, with states like Utah introducing legislation to provide limited liability frameworks.

Omar Haddad

March 31, 2026 · Updated October 20, 2026 · 6 min read

A decentralized autonomous organization is, at its core, an organization whose rules are enforced by software rather than by a manager or a board. Ethereum's own educational documentation describes DAOs as communities that coordinate through smart contracts—self-executing code that holds funds and carries out decisions—so that no single party can act unilaterally. What distinguishes one DAO from another is not the blockchain underneath but the governance model layered on top: who votes, how votes translate into action, and who ultimately presses the button that moves the treasury. Three patterns dominate—delegation, automatic transaction execution, and multisig signing—and each embodies a different trade-off between participation, speed, and trust.

How DAO Governance Models Work

The first question any DAO must answer is how proposals become decisions. Ethereum's documentation frames the choice as a set of considerations around "how voting and proposals work," and the models that emerge resolve those considerations in visibly different ways. The distinction matters in practice because it determines whether a DAO behaves like a direct democracy, a representative one, or something closer to a board-controlled trust.

Delegation is, in Ethereum's framing, "the DAO version of representative democracy." Token holders who do not want to follow every proposal can delegate their votes to members who nominate themselves and commit to stewarding the protocol and staying informed. The ENS DAO is the canonical example: ENS token holders can delegate their votes to engaged community members who represent them. The model buys informed decision-making at the price of concentrating influence in a professional class of delegates.

Automatic transaction governance removes the human execution step entirely. In many DAOs, transactions execute automatically once a quorum of members votes affirmative. Nouns DAO illustrates the pattern with a twist: a transaction is automatically executed if a quorum is met and a majority votes affirmative—as long as it is not vetoed by the founders. That veto clause is the tell. Pure automation maximizes trustlessness, but Nouns retains a founder check, showing how hybrid designs hedge against automated outcomes the founding team considers harmful.

Multisig governance takes the opposite approach to trust. Even a DAO with thousands of voting members may keep its funds in a shared wallet controlled by five to twenty active, trusted community members—typically people whose public identities are known to the community. After a community vote, these signers execute the will of the members. The model does not eliminate trust; it localizes it in a small, accountable group. Notably, the multisig pattern is primarily a mechanism for fund custody and execution rather than a complete governance system, and the available descriptions of it do not attach it to a single named DAO the way delegation attaches to ENS or automation to Nouns.

Comparing DAO Governance Models

The table below categorizes the three governance structures by their decision mechanism and execution method, mapping each to its documented example. Reading across the rows shows the central trade-off: as execution becomes more automated, fewer trusted humans stand between a vote and a completed transaction.

Three documented DAO governance models and their operational differences
Governance Model	Decision Mechanism	Execution Method	Documented Example	Primary Trade-off
Delegation	Token holders delegate voting power to self-nominated representatives who commit to staying informed	Representatives vote on proposals; execution follows the delegate-driven vote	ENS DAO, where holders delegate votes to engaged community members	Informed, specialized decision-making versus concentrated influence in a delegate class
Automatic transaction governance	Proposals pass when a quorum votes and a majority votes affirmative	Transactions execute automatically once thresholds are met	Nouns DAO, where approved transactions execute automatically unless vetoed by the founders	Trustless, immediate execution versus exposure to automated outcomes—mitigated in Nouns by a founder veto
Multisig governance	Community votes signal intent, but a small group of 5–20 trusted, publicly known signers must approve transactions	Signers execute the community's will after a vote	Described generally by Ethereum's documentation as a fund-custody pattern; no single named DAO is attached to it in the available accounts	Accountable human control over funds versus reliance on a small trusted group rather than pure code

Three documented DAO governance models and their operational differences

Governance Model

Decision Mechanism

Execution Method

Documented Example

Primary Trade-off

Delegation

Token holders delegate voting power to self-nominated representatives who commit to staying informed

Representatives vote on proposals; execution follows the delegate-driven vote

ENS DAO, where holders delegate votes to engaged community members

Informed, specialized decision-making versus concentrated influence in a delegate class

Automatic transaction governance

Proposals pass when a quorum votes and a majority votes affirmative

Transactions execute automatically once thresholds are met

Nouns DAO, where approved transactions execute automatically unless vetoed by the founders

Trustless, immediate execution versus exposure to automated outcomes—mitigated in Nouns by a founder veto

Multisig governance

Community votes signal intent, but a small group of 5–20 trusted, publicly known signers must approve transactions

Signers execute the community's will after a vote

Described generally by Ethereum's documentation as a fund-custody pattern; no single named DAO is attached to it in the available accounts

Accountable human control over funds versus reliance on a small trusted group rather than pure code

The comparison also clarifies what "decentralized" means in each case. Delegation decentralizes the choice of who decides; automatic execution decentralizes the act of deciding itself; multisig centralizes only the final signature while leaving the vote open to everyone. A reader evaluating any specific DAO can ask which of these three boxes its treasury and proposal process actually fall into—and many organizations combine more than one.

Key Provisions of Utah's Decentralized Autonomous Organizations Act

The details of Utah's law matter as much as its existence. Attorneys Matthew Berlin and Gus N. Paras, writing for ArentFox Schiff, report that the Utah Decentralized Autonomous Organizations Act was signed into law on March 13, 2023, and took effect on January 1, 2024. Their account places Utah in a specific lineage: while Wyoming, Tennessee, and Vermont had already enacted legislation enabling DAOs to be formed as limited liability companies, Utah was the first state to classify DAOs as a distinct legal entity—the limited liability decentralized autonomous organization, or LLD.

The distinction between "an LLC you can form as a DAO" and "a DAO as its own entity type" is more than semantic. Under the LLD framework, the law firm reports, members of an LLD are liable only for their on-chain contributions to the organization—a boundary that, if it holds in practice, sharply limits the general-partnership exposure Chainalysis describes as a risk elsewhere.

Qualification is not automatic, and the requirements reveal the legislature's priorities. According to the ArentFox Schiff analysis, the Utah DAO Act requires each LLD to provide the Utah Division of Corporations with evidence that its software code "has undergone quality assurance," and that it maintains a graphical user interface through which all transactions originating from, or addressed to, the LLD's smart contracts can be monitored. In other words, Utah trades legal recognition for verifiability: a DAO seeking LLD status must be able to show that its code has been tested and that its activity can be observed. That is a governance requirement as much as a technical one, and it connects directly back to the models above—an automatically executing DAO like Nouns, or a multisig treasury, must be able to surface its transactions to a human-readable interface to qualify.