Why
A self-description changed, and self-descriptions are where revenue models announce themselves early. Use Ethereum cheaply sells transactions and earns per transaction. Run your chain here sells space and earns rent. The second is a better business — recurring, less elastic, harder to leave — and it is a different relationship with whoever builds on top. Tenants and users are not the same thing, and the shift is visible in the sentence before it is visible in the accounts.
The convergence is what makes it worth a card rather than a note. headless-exchange-engines describes an interop project selling matching, clearing and settlement so others can run venues on it. This is an L2 selling settlement space so others can run chains on it. Different starting points, one destination: sell the place, not the tool. When two competitive positions arrive at the same model independently, it is usually because the tool has commoditised and the place has not.
Which reframes a decision this project actually has to make. Which chain do we launch on looks like a technical question — throughput, fees, tooling, ecosystem. Under a rent model it is a commercial negotiation, and the terms are the ones any tenant should ask about: who controls the sequencer, how are fees split, and what does leaving cost. Those three determine margin more than gas prices do, and none of them appears in a benchmark.
The exit cost is the term to price first, because it is the one nobody markets. Migration is quoted in engineering weeks, which understates it — liquidity does not migrate with the code, integrations have to be re-agreed, and users have to be told to move. A host that is cheap to join and expensive to leave has priced the second half into the first, and the honest way to compare two offers is to compare them at the moment you want to go.
How it works
The lease term sheet
| Term | The question | Why it is margin, not architecture |
|---|---|---|
| Sequencer control | Who orders transactions, and can they censor or reorder? | Ordering rights are MEV rights; whoever holds them holds a claim on your flow |
| Fee split | What share goes to the host, and who can change it? | A split that the host can revise unilaterally is not a price, it is an option |
| Exit cost | Engineering weeks, plus liquidity that will not follow, plus integrations to re-agree | The only term that grows the longer you stay |
| Host failure | What happens to open positions if the host stops operating? | For a settlement product this is not a business risk, it is a correctness question |
Two pitches, one destination
| Was selling | Now selling | Earns from | |
|---|---|---|---|
| L2 / Offchain Labs | Cheap blockspace on Ethereum | Somewhere to run your own chain | Rent on settlement space |
| Interop / LayerZero ATLAS | Messaging between chains | An engine to run your own venue | The venues built on it |
Both moved from selling a tool to selling a place, which usually means the tool stopped being scarce. For a tenant the practical consequence is identical either way: the thing being negotiated is no longer performance, it is control.
The row to price first
Exit. Everything else is quoted openly because it helps the host compete; exit cost is the one term that only matters after you have committed. Estimate it honestly and it changes the ranking more often than fees do: a host that is 30% cheaper per transaction and locks liquidity behind a bridge nobody else supports is not cheaper. cross-border-rail-interop reached the same conclusion about payment corridors — governance, not throughput, decides which rail wins — and a lease is governance written as a contract.