Morpho CEO Vault Classification Proposal Draws Fire From Aave Founder Kulechov

Morpho CEO Paul Frambot proposed today that onchain vaults be divided into two regulatory categories: "non-custodial" and "discretionary." By nightfall, the DeFi community had pushed back hard.

Frambot's framework was explicitly designed with regulators in mind, offering a clean binary that policymakers could use to assess custody risk in decentralized vaults. Non-custodial vaults, under his proposed taxonomy, would operate under conditions that remove meaningful manager control. Discretionary vaults, by contrast, would acknowledge that a human or organizational actor retains authority over capital allocation. The intent is to give regulators a workable vocabulary for a product category that currently has none.

Aave founder Stani Kulechov was unconvinced. He called the split "self-serving," a pointed accusation that implies Morpho's architecture would fare better under Frambot's classification than competing protocols. Kulechov's criticism landed quickly and publicly, framing the proposal less as neutral infrastructure thinking and more as competitive positioning dressed in regulatory language. Aave runs its own vault-adjacent products and has a direct stake in how such categories are drawn.

The more technical objection came from vault builders themselves, who argued that the proposal's core mechanism is flawed. A timelock is a smart contract feature that delays the execution of certain privileged actions, giving users a window to exit before changes take effect. Frambot's framework appears to lean on timelocks as a marker of non-custodial status, but builders countered that a delayed action is still a manager's action. The manager retains the ability to queue changes, to influence outcomes, and to exercise discretion over the vault's direction. Calling that arrangement "non-custodial" overstates the protection users actually receive.

That technical gap matters enormously in a regulatory context. If a classification framework incorrectly labels discretionary vaults as non-custodial, it could mislead regulators into applying lighter-touch oversight to products that carry meaningful custody risk. Misclassification has historically invited enforcement actions in adjacent product categories, including staking services and pooled lending arrangements, where regulators have argued that user funds remained under the effective control of a centralized party regardless of the smart contract wrapper around them.

The speed of the pushback, all arriving on the same day as the proposal, reflects how sensitive vault governance has become as DeFi protocols navigate an increasingly assertive regulatory environment. The debate exposes a structural tension that the industry has not resolved. Protocols want regulatory clarity, but they disagree sharply on what that clarity should look like, and those disagreements are rarely free of competitive interest. A classification standard that one protocol helps design will almost inevitably reflect the design choices that protocol has already made.

Whether Frambot refines the proposal in response to today's criticism or defends the existing framework, the exchange has surfaced a real problem: DeFi vaults occupy a genuinely ambiguous regulatory space, and the industry's internal disagreements make it harder, not easier, for outside observers to draw principled distinctions. A fragmented set of competing classification standards, each championed by a different protocol, would likely produce more regulatory confusion than a single imperfect framework. The vault builders who rejected the timelock logic are right that the mechanics are more nuanced than a binary allows. The question now is whether anyone in the space can produce a framework that is both technically honest and politically acceptable to the protocols that would have to live under it.