Commentary · Fraud & asset recovery · Technology disputes
The $285M Drift exploit: can victims hold the Foundation liable?
- On 1 April 2026, Drift Protocol lost $285 million of user assets. It was a governance failure, not a code exploit: North Korean state-sponsored actors manipulated the Security Council into "blind-signing" delayed authorisations.
- Although marketed as non-custodial, Drift's Foundation held upgrade keys that could rewrite the rules protecting the vaults, so it was the single point of failure.
- Victims' strongest target is arguably the Foundation (in tort and fiduciary duty), not the opaque terms of use. Then come the contractual interface entity, and the off-ramps where stolen funds land.
What happened
Drift is a decentralised exchange on Solana specialising in perpetual swaps. The attackers spent months building trust, posing as high-frequency traders and technical contributors. They then exploited durable nonces, a Solana feature that lets transactions be signed in advance and executed later, to get Security Council members to sign delayed authorisations disguised as routine maintenance. Blind-signing in this way is like signing a blank cheque.
Once triggered, those "master key" signatures let the attackers whitelist a worthless token as valid collateral, bypass the protocol's safety logic, and drain the vaults. The safe didn't break; the people holding the keys were tricked into handing them over.
A hierarchy of control behind the "decentralised" label
- Participant layer: users and liquidity providers connect wallets and deposit assets.
- Asset layer: deposits sit in program-derived addresses, programmatically locked "safes" without private keys. On paper, this is the non-custodial layer.
- Core protocol: the smart contract "accountant" that decides when vaults open and close, and runs trades, margin and liquidations.
- Governance layer: the Drift Foundation and its Security Council hold the upgrade keys, with the power to rewrite the core protocol and so authorise the vaults to release funds to anyone.
Layers 2 and 3 are automated, but layer 4 is human. Because the Foundation could overwrite the rules protecting the vaults, it arguably had custody of every participant's assets.
Why the terms of use are the wrong first target
The usual instinct is to look for the platform's contractual or tortious duty to keep deposits safe. But terms of use are stacked against victims, almost always include arbitration clauses and class action waivers, and here don't even identify the operating entity. The Drift Labs and Drift Foundation entities aren't readily a matter of public record. Operating through obscured entities lets foundations enjoy the benefits of a centralised business while sheltering behind the language of decentralisation.
The Foundation has no contract with victims, but it had real control over their assets, and it has assets of its own to enforce against. Targeting it avoids a long on-chain chase after state-sponsored actors, and sidesteps many of the hurdles in the terms of use.
A three-tier recovery strategy
Tier 1: the Drift Foundation and Security Council
- By keeping master upgrade keys without a timelock or rigorous verification (as the durable nonce failure shows), the Foundation arguably assumed a duty of care to keep user assets secure. Claims would focus on negligent management of the Security Council multisig, and on fiduciary duties the Foundation and council members arguably owed users.
- Proceedings against the Foundation and the interfacing entity as persons unknown, supported by Norwich Pharmacal or other disclosure orders against key personnel, founders or centralised providers such as hosts, and by worldwide freezing orders. Freezing a foundation's treasury can paralyse its operating capital and bring about a settlement before value dissipates.
Tier 2: the contractual interface
- Argue an express or implied term that only users can withdraw their funds, and that a third-party withdrawal authorised by governance signature is a fundamental breach.
- Look to professional liability insurance or corporate reserves held by the interface entity.
Tier 3: hackers, intermediaries and off-ramps
- Where stolen assets are traced to a centralised exchange and victims move fast, seek proprietary or freezing injunctions, asserting a proprietary interest held on constructive trust.
- An exchange that receives funds after a public exploit notice may face claims for dishonest assistance or knowing receipt.
Drift has since taken steps towards making victims whole, including US$148 million in funding from Tether and partners. As DeFi moves towards "managed DeFi", the line between software providers and fiduciary custodians is blurring, and the law needs to catch up with the architecture.
Frequently asked questions
Can victims of a DeFi exploit sue the protocol's foundation?
What if the entities behind a DEX aren't publicly identified?
This is a theoretical framework for recovery and general information only, not legal advice. No ill-will is intended towards any party.
