A rollup fee looks cheap. A user pays a small gas number and treats that number as the whole bill. The sequencer pocket is the first story. The base-layer receipt is not.
Cheap gas can be real. It can also be a spread: users pay the operator, the operator pays the layer underneath to publish data, and the gap is called revenue. I read a batch the way a shipping invoice is read. The sticker on the box is not the dock charge.
- Cheap gas is the first reading
- Two ledgers, one user
- A single operator and the line
- Sheet I kept for a rollup called RIVER
- Cheap gas stays cheap only in a window
- Split the two bills before you praise the fee
Cheap gas is the first reading
The problem is easy to name. A rollup advertises a fee that looks like a fraction of the base layer. The first reading treats that fraction as proof the system costs less to run. The user sees one number. The operator sees two.
Is the first reading always false? No. If data posted underneath is cheap that hour, and the sequencer takes a thin spread, the user number and the system number can sit close together. The first reading fails when the posted-data bill spikes, when the operator is the only door into the queue, or when “revenue” counts user fees and forgets the settlement invoice.
My claim is narrow. A sequencer is a sorter. It takes transactions, orders them, and later publishes enough data that someone else can check the state. The fee it charges is not automatically profit. Profit, if the word is even useful here, is what remains after the base layer is paid. Mixing those two lines makes a fat margin out of a pass-through.
Opinion, not a law for every design: a dashboard that shows only what users paid is a shop window. The batch on the base layer is the storeroom.
Two ledgers, one user
Users pay the rollup. The rollup pays the base layer to land data. Those are different ledgers. Compression, blobs, or calldata changes the size of the second bill. It does not erase it.
Question worth keeping on the page: who sets the user price, and who cannot avoid the data price? The operator can raise or lower what the wallet shows. The base layer sets what a byte costs that hour. A sequencer that underprices users still owes the dock.
A short model, with a limit. Call U the fees collected from users in a window. Call D the data-and-settlement cost paid underneath in the same window. The spread is U minus D. If U is 12 units and D is 9, the spread is 3. That is structure. It is not a print, and it is not a promise the next window looks the same. The limit: D moves with congestion under the rollup. A quiet afternoon and a crowded one do not share a D.
Exception: some designs share the sequencer or rotate it. Some send a share of U to a public treasury. The two-ledger split still holds. The name on the leftover changes.
A single operator and the line
Many live rollups start with one sequencer. One address orders the line. That is simple to run. It is also a single door. If the door stops, users wait, or they wait on a slower escape hatch that is written in the contract and rarely felt until it is needed.
A single door is not a theft charge. It is concentration. Censorship, reordering, and downtime sit in that concentration. I have opened a public explorer, found a batch transaction on the base layer, and matched its timestamp to a stretch of empty rollup blocks. The gap was an operational fact. It was not a price target.
Hybrid talk is common: decentralize the sequencer later. Later is a calendar word. Until that date, the economics are the economics of one sorter plus a data invoice.
Sheet I kept for a rollup called RIVER
I keep a worksheet for a made-up rollup I call RIVER. In one quiet window the sheet says users paid 10 units. Data posted underneath cost 7. Spread: 3. In a crowded window users still paid 10, because the wallet fee had been left flat. Data cost 14. Spread: minus 4. Same sticker. Different dock.
Was I looking at a live book? No. RIVER is a page. The experience was lining U and D on one row and watching the sign flip when only D moved. That is the kind of mistake a “cheap gas” headline hides. A reader can repeat the row on any public batch without taking a position in a token.
Condition I left in the margin: if RIVER compresses well, D falls and the same U looks fat. If the base layer is quiet, D falls again. Neither case turns U into a complete cost. Both cases change the leftover.
Cheap gas stays cheap only in a window
The first reading works when D is small next to U, when the operator does not take a wide spread, and when an escape path exists that a user can actually use without a research desk. It works in a window. It is not a personality trait of the chain.
It fails when the wallet number is treated as the system number. It fails when one sorter is praised as a market because the fee looks low. Low for whom? The user, the operator, or the layer that published the bytes?
I do not treat a negative spread as a command to leave. I treat it as a reason to keep the two bills on separate lines. A week of negative spread can be a choice to grow users. A year of hidden D is a different object.
Split the two bills before you praise the fee
The solution that survives the conditions above is a short split, not a slogan.
Write down U for a window you can actually observe. Write down D from the batches posted underneath in that same window. Name the operator. Note whether the door is one address or many. Keep token emissions that pay the sorter on a third line; those are not user fees. If a document will not show D, the cheap-gas sentence is not ready to stand.
If two dashboards disagree on U, say so and stop before the leftover becomes certain.
The spare thought on the desk is small. Cheap gas can be true for a user in a quiet hour. The dock charge is still on the other ledger. I split the bills first. I do not call the sticker profit because the first reading said the house was free.
The articles on this site are not investment recommendations or financial advice. They are structural analysis based on on-chain data and project documents.
Comments
Post a Comment