A hyperstructure is a crypto protocol that can run forever without anyone maintaining it, without a way to turn it off, and without collecting a fee for itself. That definition is deliberately extreme, and the interesting work is in deciding how much of it you actually want.
Where the idea came from
The argument, made by Jacob Horne in 2022, was about what tokens are for. If a protocol has an owner, that owner can change the rules — raise fees, censor users, switch off a feature that other protocols depend on. Every integration therefore includes a trust assumption about a team.
Remove the owner and the protocol becomes infrastructure: something other people build on the way they build on HTTP, because the terms cannot change under them. The properties in the original definition are free, unstoppable, expansive, permissionless, and credibly neutral. Expansive is the one that matters most and is discussed least: if a protocol takes a fee on every call, other protocols have no incentive to build on it, so it stays small.
Why it is harder than it sounds
Credible neutrality is a social property as much as a technical one. A contract with no admin key but a foundation that can rally token holders to fork it away is not neutral in the sense that matters. Neither is one whose upgrade path is a multisig controlled by the original team.
Unstoppability also cuts against the thing protocols need most in their first year: the ability to fix a bug. A truly immutable contract with a flaw is a permanently broken one, and the DAO's decision to fork Ethereum rather than accept the outcome is the standing counterexample. Most projects resolve this by starting upgradeable and planning to renounce, which usually means the renunciation never quite happens.
Fees are the third tension. A protocol with no protocol-level fee has to fund development some other way — a token, a grant, a foundation — and all three introduce a dependency the hyperstructure framing was meant to remove.
What it gets right anyway
The durable insight is about integration risk, not ideology. When a protocol's rules cannot change, a dependent protocol's security argument gets simpler, and that is worth a lot even if the protocol is not fully ownerless.
The second insight is that fee-taking is a scalability choice. Uniswap succeeded partly because it takes a small cut of a swap rather than charging for the interface, and because anyone could deploy the contracts. Protocols that tax every call create a niche for a free clone.
The third is that a protocol can be infrastructural in effect without being immutable in code — because it has ossified in practice. A network that has run unchanged for years, where changing anything would break irreversible integrations, behaves like a hyperstructure even if an admin key technically exists.
How to judge the claim
Ask four things. Does the contract have an admin key, and can it change user-facing rules? Can it be paused, and by whom? Does the protocol take a fee, and if not, who pays for development? And is the neutrality credible — would a fork of it be available if the current maintainers went the wrong way?
A protocol that scores well on all four is genuinely close to infrastructure. One that scores well on none should stop using the word, because the claim is load-bearing for the trust argument of everyone building on top of it.
Next: the governance designs that try to make ossification a process rather than an accident. Join our Discord.
References
Next, read Composable Finance 101. More in DeFi Markets.