DAOs as Legal Persons

June 8, 2026

To date, the question whether a token holder who voted in a protocol governance proposal was a general partner of her co-voters had a real and unpleasant answer. The leading authority, Sarcuni v. bZx DAO, 664 F. Supp. 3d 1100 (S.D. Cal. 2023), held that members of a DAO could function as a general partnership under California law because they had associated to carry on a business for profit. Joint and several liability followed. The CFTC took the same theory to its logical end in CFTC v. Ooki DAO, No. 2:22-cv-05416 (N.D. Cal. 2023), serving the DAO via chatbot in its own help forum and obtaining a default judgment against it as an unincorporated association. The cumulative message to anyone who had ever submitted a YES vote on a Compound or Uniswap parameter change was bleak: under existing state and federal default rules, you were potentially a co-principal in whatever your protocol did next.

The architecture frame established in Post #1 treated CLARITY’s reroute around Howey as a four-cornered structure: the trichotomy of network token, ancillary asset, and digital commodity; the §4B disclosure-only regime; the §104 coordinated-control test; and the decentralized governance system as a separate legal person. This post is on that fourth corner. Section 2(5) does an outsized amount of doctrinal work in roughly twenty-five lines. It overrides Sarcuni. It overrides Ooki DAO. It overrides the Restatement (Second) of Torts § 876 (acting in concert) theory as applied to governance participation. And it does so without inventing a new federal entity form, by accommodating the state-law wrappers that practitioners have spent four years building.

What §2(5) actually says

The definition has four moving parts. Subparagraph (A) defines a decentralized governance system as a “transparent, rules-based system permitting persons to form consensus or reach agreement in the development, provision, publication, maintenance, or administration of the distributed ledger system, in which participation is not limited to, or under the control of, any person or group of persons under common control.” Four elements: transparency, rules-based operation, the formation of consensus or agreement around development or operation of the ledger, and open participation outside the control of a common-control group.

The transparency requirement does light work but matters at the margins. A DGS whose rules cannot be observed or whose voting outcomes cannot be verified will not qualify. The rules-based element is the more substantive constraint: discretionary governance, where a small group can act outside any predefined rule, fails the test. The open-participation requirement is where the bill draws the practical line between something that looks like a DAO and something that looks like a private club whose token holders happen to vote.

Subparagraph (B) is the doctrinal payoff. With respect to a DGS, “the decentralized governance system and any persons participating in the decentralized governance system shall be treated as separate persons unless those persons are under common control or acting pursuant to an agreement to act in concert.” This is direct legislative overruling of the Sarcuni default rule. Mere participation, including voting, no longer triggers the general-partnership inference. Joint and several liability does not attach by default. The DGS has its own legal personality (within the four corners of this Act), and the participants are separate from it.

Subparagraph (D) reinforces this in slightly different language: “a decentralized governance system shall not be deemed to be a person or a group of persons acting under common control.” Where (B) addresses the relationship between the DGS and its participants, (D) addresses the DGS itself, vis-à-vis the rest of the statute. The two provisions are reconcilable. The DGS is a “person” for the purpose of being separate from its participants. It is not a “person or group of persons under common control” for the purpose of any regulatory test that turns on that phrase. Wherever “person” or “common control” is the operative trigger in the rest of the bill, the DGS slides outside the test. This is the mechanism by which DGS treatment flows through to §4B network-token classification, §104 coordinated control, §301 trading-protocol status, and the §2(13) related-person definition. Cross-references in section three.

Two things subparagraph (B) does not do are worth noting. First, it does not override common-control or in-concert findings where they actually exist. If a group of insiders coordinates governance votes via off-chain agreement, the carve-out collapses. Subparagraph (B) protects only the structural inference, not the conspiracy. Second, it does nothing to override liability for one’s own conduct. A core developer who pushes malicious code, or a multisig signer who absconds with treasury funds, remains liable for what she did. What §2(5)(B) eliminates is the imputed liability for what someone else did via the protocol. That is enough to make the difference between a workable governance role and an uninsurable one.

The Wrapper Accommodation and the Centralized-Management Line

The bill could have stopped at subparagraph (B). The interesting policy choice, and the one most relevant for transactional practice, is subparagraph (C). It states that the term “decentralized governance system” includes “a legal entity, including a decentralized unincorporated nonprofit association or other entity created pursuant to State law, used to implement the rules-based system described in subparagraph (A), provided that the legal entity does not operate pursuant to centralized management.” The reference to the decentralized unincorporated nonprofit association is unambiguous: this is the Wyoming DUNA (Wyo. Stat. §§ 17-32-101 et seq.). The phrase “other entity created pursuant to State law” sweeps in the Wyoming DAO LLC (Wyo. Stat. §§ 17-31-101 et seq.), the Vermont Blockchain-Based LLC, and any subsequent state framework Congress wants to leave room for. The Marshall Islands DAO LLC and the Cayman foundation, although not “created pursuant to State law” in the U.S. sense, satisfy the broader structural test if they otherwise meet (A).

Wrapping the protocol does not, by itself, forfeit DGS status. That is the central drafting choice. The competing model, where any incorporated entity is presumptively centralized, would have pushed structures offshore and incentivized the same legal-form arbitrage that drove the Marshall Islands and Cayman frameworks in the first place. CLARITY says instead: pick the wrapper that fits the operational reality, and let the substantive test (centralized management or not) do the work.

The centralized-management exclusion is the limit. A wrapper that “operates pursuant to centralized management” is not a DGS, regardless of what it calls itself. The bill does not define centralized management directly, but the second sentence of (C) tells us what it is not: “the delegation of ministerial or administrative authority at the direction of the participants in a decentralized governance system shall not be construed to be centralized management.” Read against subparagraph (A)’s open-participation requirement, the picture comes into focus. Centralized management means decision-making authority concentrated in a person or common-control group, exercised outside the rules-based system. Ministerial delegation, by contrast, is the execution of decisions already made by participants pursuant to the transparent rules. A foundation that pays bills on behalf of the DAO does not exercise centralized management. A foundation board that decides which proposals to fund without participant authorization does.

This distinction maps onto current structuring practice with surprising fidelity. The “service-provider” foundation model, in which the entity exists to file taxes, sign contracts, hold IP, and execute participant-approved disbursements, sits comfortably inside the ministerial-delegation safe harbor. The “steward” foundation model, in which the entity retains discretionary authority over treasury, protocol upgrades, or grants, sits outside it. The practitioner question is whether a given wrapper looks like the first or the second, and many existing structures contain both elements. Section four returns to this.

Two interpretive points before moving on. The phrase “at the direction of the participants” implies a chain of authority running from participants down to the wrapper, not the reverse. A multisig that can act on its own initiative is not acting “at the direction of” anyone. A multisig that executes transactions pre-approved by an on-chain vote is. Second, the cybersecurity carve-out at §104(b)(3)(C), which permits pre-defined, rules-based emergency measures by an incident response or security council, suggests that Congress contemplated some narrow discretionary scope being preserved without forfeiting DGS status. But that carve-out is drawn tightly: the response must be limited to a “specific and documented cybersecurity incident or imminent threat,” the rules and procedures must be publicly disclosed in advance, and no single person can hold unilateral control. It is not a license for general discretionary action.

How DGS Personhood Flows Through the Rest of the Bill

The §2(5) definition is the trunk. The carve-outs scattered through the rest of the bill are the branches, and they work consistently. The first occurs in the related-person definition at §2(13)(B), which expressly excludes a DGS from related-person status. This matters for the §104 disposition restrictions, which apply only to sales by related persons. A DGS-controlled treasury that sells units back into the market is not making a sale by a related person, and the holding-period and volume caps in §104(c) do not apply.

The next is in the network token definition itself. Section 4B(a)(7)(B) excludes from network-token status any investment contract conferring certain disqualifying financial rights, including “an entitlement to, or a reasonable expectation of, an interest, dividend, or other payment, or direct or indirect transfer of value, from a person (other than a decentralized governance system)” (§4B(a)(7)(B)(ii)(III)), and an express or implied financial interest provided by a person other than a DGS (§4B(a)(7)(B)(ii)(IV)). The parenthetical carve-out is doing serious work. Without it, any token holder who could expect payments from the DGS, whether through fee-switch distributions, staking rewards funded out of protocol revenue, or treasury grants, would be holding something with the cash-flow profile of an investment contract under standard Howey analysis. The “(other than a decentralized governance system)” parenthetical converts those flows into non-disqualifying features. The rule of construction at §4B(a)(7)(C)(ii) adds that the “granting of economic interests or voting capabilities with respect to a distributed ledger system or its decentralized governance system” does not disqualify either. Read together, these provisions mean that the DGS can pay, distribute, and confer governance rights without dragging the token into the securities perimeter.

The §104 coordinated-control architecture has two DGS carve-outs. The “distributed ledger control person” defined in §104(a)(3) is “any person or group of persons under common control, other than a decentralized governance system,” that has the unilateral authority to control or materially alter the ledger system. Then §104(b)(3)(B) supplies the safe harbor: a DGS is not a person or common-control group for purposes of the coordinated-control test, and a ledger system is not precluded from being found not-under-coordinated-control “solely based on a functional, administrative, clerical, or ministerial action of a decentralized governance system, including any such action taken by a person acting on behalf of and at the direction of that decentralized governance system.” The ministerial-delegation language from §2(5)(C) reappears here, anchored this time to a specific operational test. The structural choice is consistent: rule-based DGS conduct, including downstream ministerial execution by service providers, does not break decentralization.

Finally, §301(a)(2)(B) supplies the analogous rule for “non-decentralized finance trading protocols.” For purposes of determining whether a protocol is non-DeFi (and therefore subject to traditional intermediary regulation), “a decentralized governance system, solely by virtue of the operation of the decentralized governance system, shall not be considered to be a person or a group of persons under common control or acting pursuant to an agreement to act in concert.” The DeFi-trading test asks whether some common-control person has the authority to alter the protocol, censor users, or operate it outside of pre-encoded rules. DGS operation alone does not satisfy any of those.

The cumulative effect is that DGS treatment is not a one-off accommodation. It is a coordinated decision running through five distinct regulatory choke points: related-person status (§2(13)(B)), network-token classification (§4B(a)(7)), distributed-ledger-control-person status (§104(a)(3)), coordinated control (§104(b)(3)(B)), and DeFi-trading-protocol status (§301(a)(2)(B)). Wherever the bill could have caught a DGS in a person-and-common-control test, it doesn’t. That consistency is what makes §2(5) a structural pillar rather than a definitional curiosity.

Structuring guidance

For anyone structuring (or restructuring?) a DAO under CLARITY, the operative questions reduce to three.

First, does the protocol’s governance qualify as a DGS in the first place? That requires transparency, rules-based operation, formation of consensus around development or operation of the ledger, and participation not limited to or controlled by a common-control group. A multisig of seven anonymous signers with no public charter does not qualify. A token-weighted voting system with on-chain rules, public discussion, and dispersed token ownership ordinarily does. The harder cases sit in between: small councils with formal procedures, hybrid governance models with off-chain ratification, and rolling delegated voting structures all require closer reading against the (A) elements.

Second, does any wrapper preserve DGS status? Wyoming DUNAs and DAO LLCs, Vermont BBLLCs, and offshore analogs all qualify provided they do not operate by centralized management. The drafting question for counsel is whether the operating agreement, association charter, or articles vest discretionary authority in a manager class or council, or whether they confine the entity’s role to executing participant-approved actions. The bill rewards the latter. Existing structures that contain both elements (DAO-controlled treasury plus a discretionary foundation council, for example) should be reviewed against the centralized-management exclusion before any post-enactment certification under §104(d). The wrapper accommodation is generous, but it is conditional, and the condition is structural rather than nominal.

Third, how do off-chain operational arrangements interact with the centralized-management line? A service provider executing payments at participant direction is fine. A service provider with discretionary spending authority is not. A multisig acting on pre-approved on-chain outcomes is fine. A multisig with independent veto authority is not. The cybersecurity-council carve-out at §104(b)(3)(C) creates a narrow exception for predefined emergency response, but it should not be relied on outside its terms.

Section 2(5) is the most consequential twenty-five lines of CLARITY for any practitioner who has worked on DAO structuring under existing law. The Sarcuni general-partnership default is gone. The Ooki unincorporated-association default is gone. The doctrinal cost of governance participation has been reduced from “potential joint and several liability” to “ordinary individual liability for one’s own conduct.” That is a meaningful shift, and the rest of the bill’s DGS carve-outs are not gestures. They are the structural mechanism by which Congress carries DGS personhood through every place in the statute where the form would otherwise have collapsed back into the general-partnership default that Sarcuni and Ooki had built.

Written by David Lopez Kurtz