Rollup Sequencing Economics: Fees, MEV, and Trade-offs

Updated on Sep 19, 2026

11 min read

Rollup sequencing economics explains who gets paid when a Layer 2 accepts a transaction, which costs the operator must cover, and how transaction ordering creates additional value. This matters to rollup operators, application developers, and users comparing networks: a low advertised transaction fee does not necessarily mean the system has low costs or a sustainable business model. The Ethereum scaling documentation places rollups in the broader scaling architecture, while this guide focuses on the economic role of the sequencer that turns user order flow into batches, fees, and settlement obligations.

What Is Rollup Sequencing Economics?

Rollup sequencing economics is the study of the incentives and costs around accepting, ordering, executing, and publishing Layer 2 transactions. A sequencer is the component that receives transactions, builds a canonical order, produces a fast local result, and submits data or commitments to a settlement and data-availability layer. The OP Stack transaction flow documentation and Arbitrum’s sequencer deep dive show how two widely used rollup designs expose this flow in different ways.

The economic system has several participants:

Users and applications pay for execution, data publication, and sometimes additional protocol services.
Sequencers operate the ordering and execution path and may collect fees or retain ordering value.
Settlement and data-availability layers charge for publishing the information needed to verify or reconstruct the rollup.
Provers, validators, and infrastructure providers supply correctness checks, uptime, and recovery services.

The key distinction is between gross user fees and operator profit. A rollup may pass most of a charge through to Ethereum blobspace, while its margin is the residual after execution, data, proving, and infrastructure costs.

The Problem / Why It Exists

A rollup moves execution away from the base layer, but it does not remove the need for a block producer. Transactions still need a deterministic order, a current state, a fee policy, and a mechanism for publishing enough data to make the result verifiable. Someone must run that service and pay for the resources it consumes.

The market has two tensions: users want predictable fees, while operators must fund settlement, monitoring, security, and downtime risk; and ordering can create value that becomes private rent rather than a transparent user discount.

The sequencing choice determines how those tensions are distributed:

A centralized sequencer can provide fast confirmations and simple fee estimation, but it concentrates revenue, censorship power, and operational risk.
A decentralized sequencer set can share ordering responsibility, but coordination, consensus, and communication add cost and latency.
A shared sequencer can amortize infrastructure across rollups and coordinate cross-chain activity, but it creates dependencies between otherwise independent systems.
Based sequencing can use the base layer’s proposer and builder path, reducing a rollup-specific trust dependency while giving up some local control.

The design can look cheap at low demand yet become fragile when data prices rise, volume falls, or the operator subsidizes fees to maintain adoption.

How It Works / Architecture

A simplified rollup transaction passes through four economic stages:

user submits transaction
        |
        v
sequencer accepts, orders, and executes it
        |
        +--> charges an L2 execution fee
        |
        v
batcher compresses transactions and publishes data
        |
        +--> pays settlement and data-availability costs
        |
        v
proof or challenge system confirms the state transition

The user-facing fee is often a combination of several components:

user fee
  = L2 execution gas × L2 gas price
  + estimated L1 data-publication fee
  + protocol or operator fee
  + optional priority or service fee

The exact formula depends on the rollup. Some networks expose one combined quote, while others expose separate execution and data estimates. The data component can change when the rollup’s compression ratio, batch interval, or underlying blob market changes. EIP-4844 defines the blob transaction mechanism that many rollups use for temporary data publication, but it does not define a universal retail fee formula for every rollup.

The operator’s economic result is different:

sequencer margin
  = execution and protocol revenue
  + retained ordering value
  - L1 settlement and data costs
  - proving and verification costs
  - infrastructure and operations costs
  - rebates, incentives, and failure costs

Retained ordering value includes priority revenue and, depending on the design, MEV captured by the sequencer. It is not automatically profit because private order-flow integrations, risk controls, incentives, and refunds also cost money.

Comparing sequencing models

Feature	Centralized sequencer	Decentralized sequencer set	Shared sequencer	Based sequencing
Primary orderer	One operator	Coordinated operator set	Common service for several rollups	Base-layer proposer and builder path
Confirmation latency	Usually lowest	Higher coordination overhead	Low to moderate	Tied to base-layer timing
Fee-market control	High local control	Shared by protocol rules	Shared policy or governance	More constrained by L1
Revenue allocation	Concentrated	Split among operators or protocol	Shared service and participating chains	Routed through L1-aligned actors
Main cost risk	Operator and L1 publication costs	Coordination and redundant infrastructure	Service dependency and governance	Higher latency or reduced customization
Main trust concern	Censorship and single-operator failure	Set membership and collusion	Cross-rollup dependency	Builder and base-layer concentration
Best fit	Fast, vertically integrated rollups	Systems prioritizing operator diversity	Interoperable rollup ecosystems	Rollups prioritizing L1-aligned ordering

No row is universally superior. The economically appropriate model depends on whether the rollup values latency, sovereignty, cross-chain composability, censorship resistance, or the ability to customize its fee and ordering policy.

Components / Key Concepts

L2 execution fees

The execution fee pays for computation on the rollup. The sequencer can use a base fee, a priority fee, or a policy that resembles a first-price auction. EIP-1559 is useful background for understanding algorithmic base fees and tips, but an L2 may implement its own parameters and fee policy rather than copying Ethereum exactly.

Execution fees are not always the largest cost in a rollup. Cheap computation can coexist with expensive data publication or proving. An operator should therefore measure cost per transaction and cost per published byte, not only the displayed gas price.

Data-publication costs

The batcher compresses transactions and publishes data or commitments to the settlement layer. A fuller batch can reduce fixed settlement overhead per transaction, but waiting longer increases confirmation latency and may expose the operator to a changing blob or calldata market. Blobspace economics explains why the data fee can move independently from ordinary execution gas.

Compression creates a trade-off rather than a free saving. More compression can lower publication costs, but it may increase CPU requirements, implementation complexity, and recovery time. A production batcher needs a publication deadline as well as a target batch size.

Sequencer revenue and rent

A centralized sequencer may collect the difference between what users pay and what it spends on settlement and operations. That spread funds the service, but it can also become an opaque rent. Governance must decide whether revenue is retained by the operator, shared with users, used to subsidize fees, or directed to a protocol treasury.

Revenue accounting should keep these quantities separate:

user charges collected;
Layer 1 settlement and data costs;
proving and verification costs;
infrastructure and staffing costs;
rebates or incentives;
net protocol revenue.

Combining them into a single “sequencer fee” makes it difficult to identify whether a price increase comes from network demand, worse compression, higher margins, or a measurement error.

MEV and order flow

Transaction ordering can create value through arbitrage, liquidations, backrunning, and other state-dependent opportunities. A sequencer with exclusive access to the rollup’s order flow can capture some of that value directly or sell preferential inclusion through a private market. The MEV systems explainer covers the broader ordering problem.

MEV is not automatically harmful: efficient arbitrage can keep markets aligned, and reliable liquidation can protect lending protocols. The concern is who receives the value and whether users understand the policy. A rollup can route MEV to a treasury, rebate it, use encrypted or private order flow, or expose it to a competitive auction. Each approach changes the revenue and fairness profile.

Liveness and censorship costs

The sequencer’s most visible responsibility is inclusion. An outage can stop new transactions even when the settlement layer remains healthy. A forced-inclusion mechanism or fallback path reduces censorship risk, but it may be slower and more expensive than the normal route.

Availability also has an economic value. A rollup that promises fast confirmations must budget for redundant RPC endpoints, failover sequencing, data recovery, and incident response. If those costs are ignored, the apparent margin is overstated.

Real-World Use Cases

Fee design for consumer applications

Payments, gaming, and high-frequency applications are sensitive to small fees and confirmation delays. A rollup may choose to subsidize execution or smooth fees even when its own data costs fluctuate. The subsidy can be rational during growth, but it should be tracked as a deliberate acquisition or ecosystem expense rather than presented as the protocol’s natural cost.

DeFi and liquidation markets

DeFi applications pay for timely inclusion because a delayed swap or liquidation can lose value. A sequencer can prioritize high-fee transactions, use application-specific lanes, or introduce a private order-flow channel. Those choices affect both user prices and the distribution of MEV.

Multi-rollup ecosystems

An ecosystem with many app-specific rollups can share sequencing infrastructure to reduce duplicated operations and coordinate cross-chain actions. Shared sequencers describes the architecture; the economic question is how the service charges participating chains and how it handles outages, cross-chain ordering, and operator rewards.

L1-aligned rollups

Based rollups move sequencing closer to the base-layer proposer and builder market. This can reduce the number of rollup-specific services that must be trusted, but the trade-off is less control over preconfirmations, ordering policy, and revenue routing. Based rollups explains the architecture in more detail.

Getting Started / Practical Guide

The first practical task is to measure the fee components separately. Set an RPC endpoint for a development or production network you are authorized to query:

if [ -z "$ROLLUP_RPC_URL" ]; then
  printf '%s\n' 'Set ROLLUP_RPC_URL to a rollup JSON-RPC endpoint' >&2
  exit 1
fi

curl --fail --silent --show-error "$ROLLUP_RPC_URL" \
  -H 'content-type: application/json' \
  --data '{"jsonrpc":"2.0","method":"eth_gasPrice","params":[],"id":1}'

The result is an execution-price signal, not a complete fee quote. For a transaction estimate, provide the actual contract call and inspect the rollup’s documented fee fields:

curl --fail --silent --show-error "$ROLLUP_RPC_URL" \
  -H 'content-type: application/json' \
  --data '{"jsonrpc":"2.0","method":"eth_estimateGas","params":[{"from":"0x0000000000000000000000000000000000000001","to":"0x0000000000000000000000000000000000000002","value":"0x0"}],"id":2}'

Do not treat a placeholder address estimate as a deployable transaction. Replace both addresses and add the calldata for the application you are testing. For production accounting, record:

the fee quote returned to the user;
the L2 gas used and effective execution price;
the batch identifier and publication transaction;
the number of bytes or blobs published;
the settlement-layer fee and inclusion block;
the proving, infrastructure, and rebate costs attributable to the batch.

A basic daily report can then calculate:

cost per L2 transaction =
  (L1 publication cost + proving cost + operations cost)
  / transactions settled

net sequencing margin =
  user charges - attributable costs - rebates

Monitor queue age, batch size, compression ratio, fee-estimation error, publication delay, failed submissions, and forced-inclusion usage. A rising queue age with a stable gas price may indicate a batcher or RPC problem rather than demand. A rising publication cost with stable transaction volume may indicate weaker compression or a higher data-availability fee. These distinctions lead to different operational responses.

Common Misconceptions

“The user fee is the sequencer’s profit.”

No. A substantial part of the charge may pay for Layer 1 data publication, settlement execution, proving, and infrastructure. Profit is the residual after those costs and any rebates.

“A lower L2 gas price means the rollup is cheaper.”

Not necessarily. A rollup can advertise inexpensive execution while passing through a large data fee, charging a separate operator fee, or subsidizing the difference. Compare the complete fee quote and the operator’s cost per settled transaction.

“MEV is only a user-experience problem.”

MEV also determines how sequencing revenue is allocated. An operator that captures ordering value may fund the protocol or retain it privately. The policy affects market fairness, application incentives, and the long-term distribution of rollup revenue.

“Decentralizing the sequencer removes all economic risk.”

It changes the risk profile. A sequencer set can reduce single-operator censorship, but members need coordination, availability, rewards, slashing or replacement rules, and a way to agree on ordering. Those mechanisms have real operating costs.

Changelog

Initial publication; review the fee model whenever the rollup changes its sequencing, data-publication, or revenue policy.