In February 2026, Ethereum co-founder Vitalik Buterin publicly declared in a lengthy post that the roadmap set five years prior—one that positioned Layer 2 as Ethereum’s primary scaling solution—had "become obsolete." This statement wasn’t just a technical aside; it was corroborated by a structural reversal in on-chain data.
According to Token Terminal, the number of monthly active addresses on Ethereum Layer 2 networks plummeted from around 58.4 million in mid-2025 to about 30 million by February 2026—a nearly 50% drop. Meanwhile, active addresses on Ethereum mainnet doubled from roughly 7 million to 15 million. Users are migrating back from L2 to L1—a trend that runs counter to the mainstream expectation of the past three years that "L2s will handle the vast majority of Ethereum transactions."
This reversal in user ratio coincided with mainnet gas fees hitting historic lows. The Fusaka upgrade, which introduced PeerDAS (Peer Data Availability Sampling) technology, and the increase in blob target capacity from 6 to 14 (with a max of 21), have multiplied Ethereum mainnet’s transaction throughput compared to the early post-merge days. With the L1 gas limit raised to 60 million units—and plans to further increase it to 100 million or even 200 million—the very necessity of L2 as a "cheaper, faster" scaling tool is now fundamentally in question.
What’s Driving Users Back to Mainnet?
The most visible reason for this shift in user behavior is the convergence of gas fees, but the deeper drivers span technology, economics, and security.
On the technology front, Ethereum mainnet’s progressive scaling upgrades have significantly reduced L1 transaction costs. After the Dencun upgrade, the data availability costs for L2s posting to Ethereum dropped by over 90%. Ironically, this upgrade—intended to benefit L2s—undermined their competitive edge. When L1 gas fees are nearly as low as L2, users no longer have a strong incentive to migrate for cost savings.
Economically, L2 token value capture mechanisms have shown fundamental flaws. In 2025, total industry-wide L2 revenue plunged 53% year-over-year to about $129 million, with most of that revenue going to centralized sequencer operators. Token holders received virtually nothing. Major L2 tokens like ARB and OP are primarily used for governance voting, with no staking rewards or burn mechanisms, earning them the market label of "valueless governance assets." When L2 tokens fail to capture the consensus premium of network operations, users lose motivation to hold them.
Security is even more critical. Vitalik pinpointed the core issue: An EVM chain capable of 10,000 TPS but connected to L1 via a single multisig bridge hasn’t truly scaled Ethereum—it’s just a trust-based, standalone platform. According to L2beat, among the top 20 rollup projects, only one has reached Stage 2 (fully trustless), while 12 remain at Stage 0, heavily reliant on multisigs and auxiliary controls. When users realize their funds’ safety ultimately depends on a handful of private key holders rather than Ethereum’s cryptographic guarantees, returning to L1 becomes the rational choice.
What Are the Costs of This Structural Shift?
L2s have been downgraded from "official Ethereum shards" to "specialized plugins," and this structural transformation imposes asymmetric costs across the ecosystem.
For L2 projects, the cost is most direct. Leading L2 tokens have crashed over 90% from their all-time highs, shrinking the sector’s total market cap to about $7.95 billion. Top projects like Arbitrum and Optimism saw their tokens drop 15% to 30% in January 2026 alone. More importantly, the funding window is closing—previous L2 valuations were built on the narrative of "inheriting Ethereum’s security." With Vitalik himself refuting this, primary market valuation logic is being rewritten.
For the Ethereum ecosystem, the cost comes in the form of fragmented liquidity and distracted developer attention. Over the past five years, L2s have splintered Ethereum into dozens of isolated islands, forcing users to bear bridging risks and fees when moving between L2s. L2 teams tend to build their own tokenomics and ecosystems rather than reinforcing L1. This "warlord fragmentation" has reduced Ethereum mainnet to a mere settlement layer, diluting the ecosystem’s collective strength.
For users, the cost is increased cognitive load. Average users struggle to distinguish between L2 security stages or to judge which L2s truly inherit Ethereum’s security. Vitalik’s "trust spectrum" concept—from Stage 0 centralized multisigs to Stage 2 fully trustless—helps clarify differences but requires users to have substantial technical literacy to make safe choices.
What Does This Mean for the Crypto Industry Landscape?
This structural adjustment is reshaping the balance of power within Ethereum and spilling over into the broader crypto industry.
First, Ethereum mainnet has reestablished itself as the primary locus of value capture. Over the past three years, L2 tokens siphoned off capital and attention that might otherwise have flowed into ETH. Now, with L1 scaling and L2 token value capture mechanisms failing, capital is reassessing ETH’s security and scarcity as a base asset. Vitalik’s proposal for "native rollup precompiles"—allowing Ethereum to directly verify ZK-EVM proofs—further cements L1’s role as the ultimate verification layer.
Second, the L2 sector is undergoing brutal consolidation. A 21Shares report at the end of 2025 noted that out of more than 50 L2s, Base, Arbitrum, and Optimism handled nearly 90% of transaction volume, with Base alone accounting for over 60%. Smaller rollups saw activity drop 61%, with some projects like Kinto shutting down and Blast’s TVL collapsing by 97%. The industry is shifting from "a hundred flowers bloom" to a winner-takes-all integration phase.
Third, competition is polarizing between monolithic chains and L1-native scaling. Some capital is flowing to high-performance monolithic chains like Solana, which has demonstrated million-TPS potential in tests with its Firedancer client. Other capital is returning to Ethereum mainnet, seeking the certainty premium of security. This polarization means L2 projects must choose between "deep integration with Ethereum" or "fully independent development" as their strategic direction.
How Might This Evolve in the Future?
Vitalik hasn’t written off L2s entirely; instead, he’s charted a new course—from "scaling tools" to "specialized plugins."
The first evolutionary path is moving up the "trust spectrum." Vitalik insists that L2s managing Ethereum assets must at least reach Stage 1 security, where smart contracts have limited governance powers rather than relying solely on multisigs. For L2 projects aiming for long-term viability, progressing to Stage 2 (fully trustless) is key to building competitive moats. Native rollup precompiles are seen as critical infrastructure, enabling Ethereum to directly verify proofs and keep pace with protocol upgrades.
The second path is deep specialization in vertical domains. Vitalik encourages L2s to explore new value propositions "beyond scalability," such as privacy VMs, application-specific optimizations, and dedicated architectures for non-financial use cases (social, identity, AI). For example, on-chain identity and payment infrastructure designed for AI agents (like the x402 protocol and ERC-8004) is creating a closed technical loop. These applications don’t need to compete with L1 for general-purpose compute but instead offer unique features that L1 cannot easily provide.
The third path is complementing, rather than replacing, L1. Some industry observers suggest that L2s shouldn’t compete for transaction execution rights but should instead serve as liquidity gateways and user onboarding channels for L1. As L1 handles core asset settlement and high-value transactions, L2s can focus on high-frequency, low-value, latency-sensitive applications, forming a clearly defined symbiotic relationship.
What Are the Potential Risks and Limitations?
This transformation brings multiple risks that could affect Ethereum’s smooth evolution.
Technical risks center on scaling ZK proof verification. While native rollup precompiles can theoretically align L2 and L1 security, the upgrade cycles are complex and untested at scale. Ethereum’s transition to zero-knowledge proofs as its base verification mechanism will take years—at least until 2027—posing execution risks and market uncertainty. The developer community is still divided on the best architectural path, and technical friction is inevitable.
Economic risks involve the sustainability of L2 business models. After the Dencun and Fusaka upgrades, L2 gas arbitrage profits have been severely squeezed, with total industry revenue down 53% year-over-year in 2025. Without new revenue streams (such as MEV distribution after sequencer decentralization or application-layer service fees), many projects may exit the market due to unsustainable operating costs. 21Shares predicts that most L2s won’t survive 2026—a forecast that’s rapidly becoming reality.
Governance risks relate to the power struggle between the Ethereum Foundation and L2 interest groups. Some L2 teams have made it clear that, due to regulatory requirements, they may never move beyond Stage 1 security, as this would mean relinquishing ultimate network control. This creates tension with Ethereum’s core principles of permissionlessness and trustlessness. If L2 projects retain centralized control, the ecosystem risks long-term fragmentation into "independent kingdoms flying the Ethereum flag."
Conclusion
The collapse in Ethereum L2 user share is not a temporary market blip but a clear signal of structural transformation. As mainnet gas fees hit record lows, L2 security progress lags, and token value capture mechanisms fail, the original mission of L2 as a "scaling tool" is coming to an end.
Vitalik’s roadmap revision essentially rejects the past five years’ model of "multi-billion dollar valuations based solely on the scaling narrative." Ethereum is shifting from a "rollup-centric" approach to a new architecture where "L1 scaling is core and L2s serve as specialized complements." In this model, L2s can no longer rely on the brand narrative of "inheriting Ethereum security" but must prove their worth through unique value propositions like privacy, application-specific optimization, or AI agent infrastructure.
For industry participants, this marks a fundamental shift in evaluation standards—no longer asking "How high is this L2’s TPS?" but instead, "What can this L2 do that L1 cannot?" The narrative era is ending; the era of productivity is beginning.
FAQ
Q1: What’s the core reason for the collapse in Ethereum L2 user share?
The main reasons users are moving from L2 back to L1 are threefold: First, the Fusaka upgrade has massively boosted Ethereum mainnet’s capacity, bringing gas fees down to near-L2 levels. Second, most L2s remain stuck at Stage 0 or Stage 1 security, relying on centralized sequencers and multisig bridges, and have failed to truly inherit Ethereum’s security. Third, L2 tokens lack effective value capture mechanisms, offering no staking rewards or burn features to benefit holders.
Q2: What is Vitalik’s latest stance on L2s?
In February 2026, Vitalik publicly rejected the old "rollup-centric" roadmap, arguing that L2s should no longer be seen as "branded shards" of Ethereum. He introduced the idea of L2s as existing along a "trust spectrum," from Stage 0 (centralized multisig) to Stage 2 (fully trustless). L2s must justify their existence by providing unique value—such as privacy, application-specific optimization, or non-financial use cases—that L1 cannot deliver.
Q3: Do L2 projects still have room to survive?
Yes, but their survival will focus on two directions: One, moving up the trust spectrum to reach Stage 1 or even Stage 2 security and deeply integrating with Ethereum; and two, specializing in vertical domains, such as privacy VMs, AI agent infrastructure, or game-specific chains that L1 cannot efficiently support. General-purpose L2s relying solely on the "cheap and fast" narrative will face the greatest survival pressure.
Q4: What are the next steps for Ethereum mainnet scaling?
Ethereum mainnet plans to further increase blob target capacity to 48 by June 2026, with gas limits targeted for 100 million or even 200 million units. The Glamsterdam upgrade will focus on reducing MEV-related manipulation, stabilizing gas fees, and laying the groundwork for future scaling. The long-term goal is for L1 to independently handle massive transaction volumes while enabling deep interoperability with L2s via native rollup precompiles.
Q5: What does this mean for regular users?
With transaction costs now similar, regular users can choose the more secure Ethereum mainnet for asset operations. For scenarios requiring L2, users should pay attention to the L2’s security stage (Stage 0/1/2) and trust assumptions, prioritizing projects that have reached Stage 1 or above and are aligned with Ethereum’s security, and avoiding storing assets on centralized L2s that rely on multisig bridges.
