Ethereum does not necessarily need to process every transaction itself to remain the main settlement layer for crypto. That may sound like a contradiction, but it sits at the heart of the network’s current scaling strategy.
Rollups move much of the transaction execution away from Ethereum’s base layer, process transactions in batches, and then post the information needed to check the resulting state back to Ethereum. The aim is not to replace Ethereum, but to give it a narrower — and potentially more important job.
That raises a bigger question than whether Layer 2 networks are cheaper. If more users and applications move to rollups, could the Ethereum mainnet actually become more valuable because it does less?
Ethereum is increasingly being designed around that idea.
Ethereum rollups change what the mainnet is expected to do
A blockchain becomes difficult to scale when every participant is expected to process the same growing volume of activity. Adding more transactions is not simply a matter of making computers faster; the network must also remain accessible enough for people to verify it and run nodes.
Rollups approach the problem differently. A Layer 2 executes many transactions outside the Ethereum mainnet, bundles them together and sends the relevant transaction data and state updates back to Ethereum. This spreads the cost of interacting with the base layer across a much larger batch of activity.
For a user, the result can be a cheaper transaction. For Ethereum, the more interesting change is architectural: execution can happen elsewhere while the base layer remains responsible for settling the final state and making the data needed for verification available.
That is why Ethereum’s roadmap treats rollups as a central part of scaling rather than as a temporary workaround. The roadmap has even moved away from the older idea of scaling mainly through separate shard chains, as Layer 2 rollups developed faster than expected.
The less obvious piece of the puzzle is data
It is tempting to think that rollups solve Ethereum’s scaling problem simply by taking transactions “off-chain”. That is only half the story.
A rollup still needs to publish enough information for others to verify what happened. Data availability is therefore a critical part of the design. Ethereum needs a way to handle that information without forcing every rollup transaction to carry the same cost as a normal mainnet transaction.
This is where blobs matter.
Introduced with the Dencun upgrade, blobs provide temporary data storage designed specifically to make rollup data cheaper to publish. Unlike ordinary blockchain data that remains permanently accessible on the execution layer, blob data is retained for a limited period and can be removed from the network later.
That distinction is easy to overlook, but it changes the economics of Layer 2s. Ethereum is not simply asking rollups to process more transactions. It is also changing the infrastructure underneath them so that publishing the information required for verification becomes more efficient.
The result is a useful way of thinking about the Ethereum ecosystem: the base layer does not have to carry every piece of work forever. It needs to provide the guarantees that other layers cannot easily reproduce.
Optimistic and ZK-Rollups solve the verification problem differently
“Rollup” is an umbrella term, not a single technology.
Optimistic Rollups generally assume that the state update submitted to Ethereum is correct and provide a mechanism for challenging an invalid result. ZK-Rollups, by contrast, use cryptographic proofs to demonstrate that a batch was processed correctly.
For a beginner, the practical distinction is less important than the shared principle. Both approaches try to move execution away from Ethereum while retaining a connection to the base layer for verification and settlement.
The trade-offs are significant. Different designs can differ in proof systems, operating requirements, confirmation processes and the degree of decentralisation they currently achieve.
This is why it would be misleading to frame Ethereum’s future as a race in which one rollup technology must eventually win. Arbitrum, Optimism, Starknet and other networks can be understood as different implementations of a broader strategy: create more execution capacity without asking Ethereum’s base layer to do all of the work itself.
The real paradox: Ethereum could become more important by processing less
This is where the rollup-centric model becomes more interesting than a simple “faster Ethereum” story.
Suppose activity increasingly takes place on Layer 2 networks. At first glance, that sounds like users are moving away from Ethereum. But if those networks continue to rely on Ethereum for settlement and data availability, the relationship is closer to an infrastructure network and its dependent systems.
The user may interact almost entirely with a Layer 2, while that Layer 2 still depends on Ethereum for guarantees that matter underneath the interface.
In other words, there is a difference between people using Ethereum directly and applications relying on Ethereum indirectly.
That distinction could become increasingly important. A successful Ethereum ecosystem may not be one where everyone pays a fee on the mainnet. It may be one where many applications and specialised networks can operate at scale while still using Ethereum as their common settlement foundation.
But that model introduces a question the “rollups are the future” argument sometimes glosses over.
A Layer 2 can inherit important security properties from Ethereum without automatically inheriting every property of Ethereum.
The weak points may sit between the layers
A rollup is not magically decentralised because Ethereum is decentralised.
Many Layer 2 systems have relied on sequencers — components responsible for ordering transactions and coordinating their execution. When these systems are concentrated, they can create additional operational risks, including the possibility of censorship or dependence on a particular operator.
There are other layers of trust to consider as well: bridges, proving systems, governance arrangements and the infrastructure used to operate the network.
That matters because the security question is no longer simply “Is Ethereum secure?” It becomes “How much of Ethereum’s security does this particular Layer 2 actually inherit, and where does additional trust enter the system?”
For users, that is a much more useful question than treating every Ethereum Layer 2 as interchangeable.
The long-term bet is on a network of specialised layers
The most interesting consequence of Ethereum rollups may therefore be a change in what we mean by “Ethereum”.
Instead of picturing a single blockchain trying to accommodate every payment, application and smart contract on the same execution layer, imagine a common base supporting many specialised environments.
One network might be optimised for trading. Another could focus on gaming or consumer applications. Others might prioritise privacy, payments or a particular type of financial infrastructure.
The point is not that every rollup will succeed. It is that Ethereum can potentially support a much broader range of activity without requiring every application to compete for the same block space.
That is also why data availability remains so important to Ethereum’s roadmap. Improvements such as PeerDAS are designed to make it possible to handle more rollup data while keeping the network practical to verify and operate.
The challenge is to expand this capacity without making the base layer so demanding that ordinary users can no longer participate in securing it.
That is the tension running through the entire strategy: Ethereum wants to become a platform for much more activity, but it cannot solve scale by simply pushing all of the costs onto the people who run the underlying network.
So, could rollups be the future of Ethereum?
Possibly — but not because they make Ethereum disappear.
Their importance lies in changing the division of labour. Layer 2 networks can handle increasing amounts of execution, while Ethereum concentrates on settlement, security and data availability.
That is a more ambitious idea than simply making transactions cheaper. It suggests that Ethereum’s success may eventually be measured less by how much it executes itself and more by how much economic activity can safely settle around it.
The paradox is that the strongest version of Ethereum may be the one users see least often: a base layer doing fewer things directly, while quietly providing the foundation for many more.