What Is a Sequencer?
A sequencer is the component in a Layer 2 blockchain network responsible for ordering transactions before they are settled on the underlying Layer 1 chain. It receives user transactions, determines the order in which they are executed, batches them together, and submits the resulting data to the underlying Layer 1 for settlement.
The sequencer is not what most people picture when they think of blockchain infrastructure. It is not the chain itself. It is not the node network. In most rollups today, it is implemented as a single service sitting between the user and the chain, making decisions in real time about which transactions execute in what order.
In most Layer 2 networks today, including Base, Optimism, Arbitrum, and others, the sequencer is operated by a single entity: the team that built the network.
Why Layer 2 Networks Have Sequencers
Layer 1 blockchains like Ethereum process transactions slowly and at significant cost. Layer 2 networks solve this by moving most transaction execution off the main chain. Users transact on the Layer 2, which processes thousands of transactions quickly and cheaply, then periodically submits a compressed summary of all those transactions back to Ethereum for final settlement.
The sequencer is the component that makes this fast and cheap. It receives user transactions, processes them in milliseconds, and produces confirmations almost immediately. Without a sequencer, every transaction would have to wait for the slower, more expensive Layer 1 confirmation cycle.
The trade-off is centralisation. Running an efficient sequencer requires speed, reliability, and coordination that a decentralised node network currently cannot match. As a result, most Layer 2 teams operate a single sequencer under their own control.
What the Sequencer Controls
The sequencer has meaningful power over the transaction experience on any Layer 2 network. Its specific capabilities include:
Transaction ordering: The sequencer decides which transactions execute first. This has significant implications for Maximal Extractable Value (MEV), the practice of extracting profit by inserting or reordering transactions around others. A centralised sequencer creates a single entity with complete visibility into the transaction queue.
Inclusion and exclusion: In most current implementations, the sequencer decides which transactions to include in each batch. A sequencer that refuses to include certain transactions, or transactions from certain addresses, creates effective censorship at the infrastructure layer.
Liveness: If the sequencer goes offline, normal transaction submission typically stops until it recovers or users fall back to any available Layer 1 escape mechanisms. Most production rollups include some form of forced inclusion or escape hatch, even if it is slower and rarely used in practice.
When Sequencers Fail
In June 2026, Base, the Layer 2 network built by Coinbase, experienced two separate block production halts within 24 hours.
The cause was a bug in the sequencer's block-building logic. A failed transaction was not properly cleaned up. Stale execution data from that transaction leaked into the next transaction path. The sequencer produced an invalid state transition. The network could not continue processing transactions and halted entirely for approximately 116 minutes on June 25, then halted again for around 20 minutes the following day when a recovery attempt exposed a second bug in the sequencer node synchronisation process.
No funds were lost. The issue was patched and the network resumed. But the incident made visible something that most Base users had never thought about: the network they were transacting on could stop entirely because one company's system had a software bug. The Nexus Report covered the Base sequencer outage in detail.
The blockchain was secure. The sequencer was not redundant. Those are different properties, and users who assumed the first implied the second discovered they were wrong when the network sat still for nearly two hours.
Execution Versus Settlement
On a rollup, transaction execution and transaction settlement are separate processes. The sequencer executes transactions quickly on the Layer 2 and provides users with near-instant confirmations. Settlement occurs later, when the batch is posted to the underlying Layer 1 and secured by its consensus.
Understanding this separation explains why a Layer 2 can inherit Ethereum's settlement security while still relying on its own sequencer for day-to-day operation. The security of the final settled state comes from Ethereum. The speed and liveness of the execution layer comes from the sequencer. These are different properties, provided by different components, and they can fail independently.
The Sequencer and the Trust Assumption
When a user submits a transaction to a Layer 2 network, they are placing trust in several systems simultaneously. They trust the smart contracts to execute as written. They trust the cryptographic proofs or fraud proofs that secure settlement back to Layer 1. And they trust the sequencer to include their transaction, order it fairly, and remain online.
That last trust assumption is rarely made explicit. Nobody tells a new Base user that their ability to transact depends on a single machine operated by Coinbase. The assumption is embedded in the architecture and invisible until it fails. As The Nexus documented in its analysis of Q2 2026's exploit wave: "A protocol cannot be understood only by reading its contracts. You have to understand how it is operated, how it connects to other systems, who can upgrade it, who can pause it, who holds the keys."
This is the nature of most infrastructure trust assumptions: they are not hidden intentionally, but they are not visible without deliberate examination. A user who checks whether Base is decentralised will find documentation about Ethereum settlement and fraud proof security. They may not find a clear statement that the sequencer is centralised and operated by one company.
Sequencer Decentralisation
The major Layer 2 networks are aware of the sequencer centralisation problem and have published roadmaps toward decentralisation. The proposed solutions vary:
Multiple sequencers: Rather than one sequencer, a network of sequencers competes to produce the next batch. This removes the single point of failure but introduces coordination complexity and potential for new forms of MEV.
Shared sequencing: Multiple Layer 2 networks share a common decentralised sequencer network. This allows for cross-rollup atomicity (transactions that span multiple chains executing together) while distributing the sequencer function.
Based sequencing: The Layer 1 validators sequence Layer 2 transactions directly, inheriting the decentralisation of the base chain. This removes the separate sequencer entirely but may introduce latency.
Although research, pilots, and early deployments exist, no fully decentralised sequencing model has yet replaced the operational simplicity of a single sequencer across the largest Layer 2 ecosystems. The Ethereum Foundation's current mandate, published June 22, 2026 by interim Co-Executive Director Bastian Aue, explicitly identifies sequencer centralisation and MEV extraction as core structural security concerns for the network. Ethlabs, an independent research organisation launched by five former Ethereum Foundation researchers on the same day, is focused in part on this problem.
The direction of travel is clear. The timeline is not.
Sequencers Versus Layer 1 Validators
On a Layer 1 blockchain like Ethereum or PulseChain, there is no sequencer. Validators collectively determine transaction ordering through the network's consensus process, rather than through a single privileged operator.
PulseChain operates 44,901 active validators as of mid-2026. No single validator controls block production. If one validator goes offline, another produces the next block. The liveness property of the network does not depend on any single machine or operator.
This architectural difference is not relevant to every use case. Layer 2 networks with centralised sequencers offer significantly faster and cheaper transactions for most everyday activity, and the security of their asset settlement is ultimately backed by Ethereum. But the liveness property, the guarantee that you will be able to transact when you need to, is qualitatively different.
Understanding which you are using, and what that implies about who controls the infrastructure you depend on, is part of the trust assumption analysis that applies to every system in the stack.
Key Questions for Evaluating Any Layer 2
- Who operates the sequencer? Is it a single entity or a decentralised network?
- What happens if the sequencer goes offline? Is there a fallback?
- Does the sequencer have the power to censor transactions or exclude addresses?
- Is there a published decentralisation roadmap? What stage is it at?
- How is transaction ordering governed? Is MEV extracted, and by whom?
These questions do not have a single right answer. Centralised sequencers exist because they provide genuine performance benefits. The relevant judgment is whether those benefits are worth the liveness and censorship-resistance trade-offs for your specific use case.
Related reading: