Ethereum now sells two different resources with two different prices. Ordinary execution is paid for in gas, while the data rollups publish uses a separate market entirely.
Rollups need to publish data, not execute it
A rollup processes transactions elsewhere and posts enough information for anyone to reconstruct and verify what happened. That posting is the security anchor.
The base chain does not need to execute those transactions. It needs only to guarantee the data was available for a period long enough for anyone to check it.
Storing that data permanently in ordinary transaction fields was expensive and unnecessary, since verification only requires availability for a limited window.
Competing for one resource hurt both users
When rollup data occupied the same space as regular transactions, the two bid against each other. Heavy rollup activity raised costs for ordinary users and vice versa.
This coupling meant scaling success made the base layer more expensive, which is the opposite of the intended outcome.
Separating the resources lets each find its own price. Congestion in one no longer transmits directly into the other.
Blobs are temporary by design
Data posted as a blob is carried by consensus nodes for a defined period and then discarded. The chain retains a commitment proving what was posted.
Anyone needing the data long term must store it themselves, which rollup operators and third-party services do. The base layer guarantees availability, not permanence.
This deliberate impermanence keeps node storage requirements bounded even as rollup throughput grows substantially.
The fee market targets an average
Blob pricing uses a mechanism similar to the base fee for gas, adjusting up when usage exceeds a target and down when it falls short.
Because the target sits below the maximum, there is normally spare capacity and the price stays near its floor. Costs for rollups become very low most of the time.
When demand exceeds the target consistently, the price rises quickly. The mechanism is designed to respond sharply rather than to allow sustained congestion.
The economics push work off the base layer
Cheap data changes what is worth building. Applications that could not justify base layer costs become viable on rollups posting into this market.
It also changes where fee revenue accrues, since activity moving to rollups pays blob fees rather than execution gas. The two are not equivalent in magnitude.
How that revenue balance evolves as blob demand grows is an open question with real consequences for the network's long-run economics.