Are L2s eroding the value of L1? Is Ethereum losing its global composability?
During the peak of the rollup narrative, concerns like these spread throughout the Ethereum community. The prevailing logic was straightforward: if transactions on Mainnet were becoming increasingly expensive, more applications, assets, and users should move to independent L2s.
But in 2026, that trend began to shift.
Synthetix deployed its new perpetual futures exchange back on Ethereum Mainnet. ENS abandoned the long-planned Namechain. Projects such as Trueo and Superseed also chose L1 for their next phase. For a growing number of teams, maintaining an independent L2 no longer appeared to be the optimal choice.
Some L2s even began scaling back or shutting down. In early October, Blast, which once held more than $2 billion in TVL, announced that it would wind down operations. Abstract then announced plans to close its network in December.
Meanwhile, the Ethereum Foundation reorganized protocol development around three strategic priorities—Scale, Improve UX, and Harden the L1—pushing Mainnet performance into a new phase of engineering and delivery.
So how did Ethereum move from “everything can be a rollup” to a growing return to Mainnet?
1. From the L2 Migration Wave to a Mainnet Return
Over the past few years, high gas fees on Ethereum Mainnet made launching a dedicated L2—or deploying a protocol’s core business on an existing rollup—a standard move for many leading projects.
More recently, however, the onchain ecosystem has begun moving in the opposite direction. Projects across different sectors are reassessing the choice between L1 and L2 based on infrastructure costs, asset security, and access to ecosystem liquidity.
ENS is one of the clearest examples.
In 2024, ENS Labs planned a dedicated ZK rollup called Namechain for ENSv2, aiming to reduce the cost of registering and managing names.
But in February 2026, ENS co-founder Nick Johnson announced that development of Namechain would stop and that ENSv2 would instead be deployed directly on Ethereum Mainnet. His explanation was simple: Ethereum L1 had scaled far faster than the team expected two years earlier.
According to ENS Labs, the average gas cost of registering an ENS name fell by roughly 99% over the previous year—from nearly $5 to less than $0.05—as Ethereum's gas limit increased.
Maintaining an independent L2, meanwhile, would have required additional nodes and infrastructure, cross-chain resolution, bridging, security audits, and ongoing operational work.
If a service can already run on Mainnet at minimal cost, maintaining a separate rollup may only make its architecture more complicated.
A similar shift is taking place in DeFi.
In June 2025, Synthetix—which had previously been an active supporter of Ethereum’s L2 ecosystem—published “L2s Are Dead, Long Live Ethereum.” The article argued that the existing L2 scaling model had introduced problems such as fragmented liquidity and unreliable infrastructure.
In December of the same year, Synthetix launched its next-generation perpetual futures exchange on Ethereum Mainnet.
More projects followed in the second half of 2026:
- Trueo: In September, Trueo announced that it would move its primary deployment from Base to Ethereum L1. The project cited Mainnet’s stronger integration potential, network effects, and neutrality as important foundations for building a more credible prediction market and outcome oracle.
- ShapeShift: On October 1, ShapeShift moved its rFOX staking program from Arbitrum to Ethereum Mainnet. The reason was practical: most FOX was already held on Mainnet, so requiring users to bridge their tokens simply to participate in staking added unnecessary friction.
- Superseed: In August, Superseed shut down its L2 and announced plans to move the ecosystem's core products to Ethereum Mainnet. Instead of relying on L2 gas revenue to repay loans, the project would pursue the same product goal through application-layer yield strategies.
- Nillion: In October, Nillion launched Blacklight L1 and began moving node operations from Nillion L2 to Ethereum Mainnet, with plans to retire the existing L2 network.
These cases do not mean that every protocol is leaving L2. They do, however, show that more teams are choosing infrastructure based on their actual requirements for security, liquidity, and execution efficiency.
The operating pressure facing some independent L2s has also become harder to ignore.
On October 2, Blast announced that it would wind down operations because the cost of maintaining the network had exceeded the revenue generated by the L2, with no credible path to economic sustainability.
Abstract subsequently announced that it would shut down its network on December 15. Although the chain had integrated more than 144 applications and attracted over 400,000 users, thin liquidity, a limited DeFi ecosystem, and weak institutional participation made an independently operated consumer chain difficult to sustain.
These cases demonstrate an important point: owning a chain does not automatically create a sustainable business model.
As Ethereum continues to lower the cost of publishing L2 data while improving Mainnet performance, attracting users through low gas fees and token incentives alone is becoming less likely to create a lasting competitive advantage.
2. Reframing L1 and L2: Keep State on Mainnet, Move Computation Elsewhere
Why are projects becoming less determined to maintain all-purpose L2s?
Beyond operating costs lies a more fundamental issue: as applications and liquidity spread across different networks, what happens to one of Ethereum’s most valuable properties—global composability?
The original rollup-centric roadmap established a clear division of labor. L1 would provide security, data availability, and final settlement, while L2s would deliver greater scale and lower gas fees.
When Mainnet transactions regularly cost tens of dollars, this appeared to be the only practical answer.
The reality has proved more complicated.
As of November 2026, ETHBEAT—formerly L2BEAT—tracked dozens of different L2s. Only a handful had reached Stage 2, and the value they secured remained small relative to the largest rollups. Major networks such as Base, Arbitrum One, and OP Mainnet were still classified as Stage 1.
Progress toward Stage 2 has been much slower than expected. Ethereum L1, meanwhile, has scaled increasingly quickly.
As Vitalik Buterin summarized it, the original vision for L2s and their role in Ethereum no longer makes sense in its previous form.
Some L2s may also choose to remain at Stage 1 because of regulatory requirements, business models, or a desire to retain greater control over customers.
That does not necessarily mean they have made the wrong choice. For institutional applications with specific regulatory obligations, retaining administrative and upgrade authority may be an intentional part of the product design.
However, a network that continues to depend on a small group of entities for critical upgrades or state decisions should not automatically be treated as a fully decentralized shard of Ethereum Mainnet.
At the application level, one practical compromise is to keep assets and critical state on L1 while moving high-frequency or computationally intensive work offchain or into other execution environments.
Synthetix’s new perpetual futures exchange is a representative example.
It uses a hybrid architecture in which user assets remain in contracts on Ethereum L1, while order matching is handled by a high-performance offchain central limit order book (CLOB). Trade results are then settled back to Mainnet in batches.
Users do not need to bridge large amounts of capital to another chain to obtain faster execution. At the same time, they can access an order-matching experience closer to that of a centralized exchange.
This architecture does not eliminate every trust assumption. Some margin management still depends on the offchain order book, and the matching system itself must remain subject to scrutiny.
But custody and final withdrawals remain secured by Mainnet contracts. Users do not have to entrust the full security of their funds to the sequencer and bridge of an independent L2.
This points to another way of thinking about onchain finance: not every computation has to happen onchain. The more important question is which parts of a system require the trust guarantees that only onchain rules can provide.
Over the longer term, Vitalik’s work on native rollups may offer a more complete solution.
Further reading: “Making Sense of Vitalik’s L2 Rethink: Moving Past Fragmentation and Refocusing on Native Rollups.”
In simple terms, native rollups would use Ethereum’s protocol-native ZK-EVM verification capabilities, allowing Mainnet to verify EVM execution performed by rollups directly. This could reduce dependence on external security councils and create the conditions for more trustworthy interoperability and synchronous composability between rollups.
Initiatives such as the Ethereum Economic Zone (EEZ) are also exploring tighter synchronous interaction between rollups and Mainnet.
These technologies remain at different stages of research, development, and verification. But they point in the same direction: users should not be forced to accept fragmented assets, liquidity, and trust models simply because they use different execution environments.
L2s will not disappear. They will continue to offer important advantages for high-frequency trading, gaming, privacy-preserving computation, and applications that require greater control over their execution environments.
What needs to change is the assumption that scaling should remain the sole defining purpose of L2s.
3. When Both L1 and L2 Are Cheap, How Does ETH Capture Value?
The changes discussed so far concern infrastructure and application architecture. But they ultimately lead to a more practical question: if Ethereum Mainnet becomes more capable and more applications return to L1, will ETH’s value proposition change as well?
Over the past few years, the crypto market’s valuation of ETH has relied heavily on the fee-burning mechanism introduced by EIP-1559.
Under this mechanism, the base fee paid by users is burned, while the priority fee is paid to validators. When transaction demand rises and blockspace becomes scarce, the base fee generally increases, causing more ETH to be burned.
This became a key foundation of the “ultrasound money” narrative. If network activity continued to grow, fee burn could offset or even exceed new ETH issuance, making the asset net deflationary.
But this is also where the problem lies.
As upgrades such as Dencun and Fusaka reduce the cost of publishing L2 data to Mainnet, while Ethereum continues to raise the L1 gas limit, transaction costs and fee revenue may decline.
Even if total ecosystem activity continues to grow, ETH burn will not necessarily rise at the same rate.
Relying on higher gas fees to support ETH’s valuation could also force ordinary users and real applications to bear greater costs, ultimately weakening Ethereum’s long-term competitiveness.
Vitalik Buterin has addressed ETH’s value proposition directly. His position is clear: rather than treating ETH primarily as an asset whose value comes from fee burn, Ethereum should devote more attention to ETH as a store of value.
Vitalik did not reject EIP-1559 or propose eliminating fee burn.
On the contrary, he argued that the mechanism remains useful and should be retained. But even if fee burn offsets less than 10% of new ETH issuance over the long term, holders should not be disappointed, because fee burn is not ETH's only source of value.
Why?
First, the periods of highest ETH burn were driven largely by speculative activity within the crypto industry.
The 2022 Otherdeed NFT mint by Yuga Labs, for example, generated roughly $180 million in fee burn in a single day. But this was not external demand in any meaningful sense. If the crypto market contracted and users had less capital available for speculation, both this activity and its associated fee revenue could disappear.
Historical burn figures therefore do not provide a reliable long-term valuation floor for ETH.
Second, if Ethereum had to generate enough fee revenue to justify a market capitalization of hundreds of billions of dollars, it might be forced to operate more like a large commercial corporation—continually increasing development spending, changing product direction, and intervening more aggressively through governance.
That would not be fully compatible with Ethereum’s long-standing goals of decentralization and resistance to control.
More importantly, deliberately keeping gas fees high to increase ETH burn could exclude the very users who most need access to privacy, security, and open financial services.
So if fee burn is not the primary foundation, what should support ETH’s long-term value?
Vitalik’s answer is Ethereum’s genuinely distinctive advantage: decentralization and the properties he summarizes as CROPS—Censorship and Capture Resistance, Open Source, Privacy, and Security.
Together, these properties point to a more practical source of value: stability—the network’s ability to continue operating as the external environment changes.
In an era of rapidly advancing AI and continuous change across traditional finance and digital infrastructure, a financial system that can remain neutral, reliable, and difficult for external powers to control may be far more valuable than a temporary period of high fee revenue.
From this perspective, strengthening L1 reinforces ETH’s long-term value proposition in three ways:
- Core state and settlement strengthen Ethereum’s most important network effect. Computation can increasingly become a commodity, and execution environments can be replicated. But a shared settlement layer containing large amounts of capital, mature financial protocols, and deep liquidity cannot be rebuilt quickly. As core financial activity converges on Mainnet, ETH’s role as the native gas asset, collateral, and settlement asset may become stronger.
- Mainnet scaling allows ETH to move beyond dependence on expensive individual transactions. Over the past two years, Ethereum has worked to resolve a long-standing tension: how to increase throughput without making the hardware needed to run a node prohibitively expensive. If the relevant technologies mature, Ethereum can evolve from a low-frequency, high-cost settlement network into a higher-throughput, lower-cost global financial infrastructure.
- Staking and economic security create a broader foundation for ETH demand. ETH is the native resource required to use the network, the economic collateral securing PoS consensus, and an important collateral and liquidity asset across DeFi and onchain finance. This provides a more complete explanation of ETH’s long-term value than fee burn alone.
Final Thoughts
If Ethereum’s development over the past few years is viewed as a large-scale experiment in scaling, the changes of 2026 suggest that the ecosystem is beginning to search for a new balance.
L2s will not—and should not—disappear. They remain an important part of Ethereum’s ability to support new applications, differentiated execution environments, and more users.
At the same time, L1 cannot remain indefinitely confined to the role of data availability and final settlement.
Ethereum must strengthen Mainnet execution, reinforce its position as a center of liquidity, preserve censorship resistance, and deepen economic security so that the most important assets and financial state can remain in a sufficiently trustworthy environment.
This reflects the philosophy behind Lean Ethereum: remove unnecessary redundancy and friction while bringing security, censorship resistance, and state consensus back to the center of the protocol.
Further reading: “One Year Later, Lean Ethereum Returns: What Is Ethereum Trying to Prove?”
As the most valuable state and liquidity reconverge on Mainnet, Ethereum may enter its next decade with greater performance, a more unified user experience, and stronger economic foundations.