SEC Commissioner Clarifies Securities Law Reach for Crypto Vaults and Lending Protocols
According to Deloitte’s Accounting Research Tool, SEC Commissioner Hester Peirce has stated that placing securities-related activity on-chain does not remove it from the reach of federal securities laws.

Her comments specifically address crypto vaults and on-chain lending strategies—two design patterns where protocol architecture can obscure, but not necessarily alter, the regulated activity beneath it.
For builders and reviewers, the operative issue is not whether a product uses smart contracts, automated execution, or self-custody interfaces. It is whether the system is conducting securities-related activity, and which party in the stack is responsible for the relevant function.
The compliance boundary is functional, not infrastructural
The statement is a useful warning against a familiar category error: treating the blockchain settlement layer as a legal perimeter. A vault may package deposits, allocation rules, strategy execution, and withdrawals into a single contract interface; an on-chain lending strategy may similarly automate collateral management and lending flows. Neither architectural compression, on its own, determines regulatory treatment.
That matters because “vault” is an overloaded term. In protocol design, it can describe anything from a narrowly scoped smart-contract container to a broader strategy layer that directs assets into other activities. The same ambiguity exists in lending: the front end, contract logic, governance process, and entities surrounding deployment may each perform materially different roles.
Peirce’s stated position, as summarized by Deloitte, is that an on-chain implementation does not create an exemption where securities laws otherwise apply. The practical implication is straightforward: teams should map the activity before they map the interface.
A vault review should follow the asset and control flow
Security audits usually begin with attack vectors: privileged functions, upgradeability, oracle dependencies, withdrawal paths, and liquidation logic. Those remain essential. But vault and lending teams now need a parallel review of the operational model surrounding the code.
The first question is what the vault or strategy actually does with deposited assets. The second is who defines, modifies, or executes those rules. The third is where discretionary control, asset pooling, or external dependencies enter the system. A protocol can be technically decentralized in selected components while still presenting identifiable control surfaces elsewhere in the stack.
That distinction is also relevant to documentation. Product language that calls a system “non-custodial,” “automated,” or “permissionless” may describe a technical property, but it does not substitute for a functional account of the activity. Advanced users should therefore look beyond the transaction path and ask what a strategy is designed to accomplish, who can alter its parameters, and how the lending or vault logic is presented to participants.
What to track next
The available summary indicates that Peirce addressed evolving definitions and regulatory implications, rather than announcing a protocol-specific finding. That leaves important implementation questions open: how particular vault structures will be characterized, how responsibility is allocated among participants, and which facts become decisive in a given deployment.
For protocol teams, the near-term response should be disciplined evidence collection rather than reactive relabeling. Preserve a clear record of system roles, control pathways, upgrade authorities, strategy logic, and the relationship between any on-chain mechanism and its surrounding service layer. That record supports both security review and legal analysis.
The longer-term signal is that regulatory scrutiny is moving closer to application architecture. As vaults and lending strategies become more composable, the challenge is not merely avoiding state bloat or reducing smart-contract risk. It is ensuring that composability does not make the relevant activity—and the accountability attached to it—harder to identify.