#以太坊L2如何发展？ L2 is Cooling Off? Vitalik Changes Tune After Reversal, the Deadlock of Ethereum L2's Fate

On February 3, 2026, Ethereum co-founder Vitalik Buterin changed his long-held view that Layer-2 is the primary way to scale Ethereum, stating that this approach "no longer makes sense."
Vitalik's remarks clearly put L2 at a crossroads—what does the future hold for L2?

1. L2 at the Crossroads
Layer-2 aims to participate in "Ethereum scaling" by creating block space fully protected by the Ethereum mainnet. In this block space, all transactions are valid, censorship-resistant, and final, but many Layer-2 networks have failed to meet this standard.
"If you create a 10,000 TPS EVM, but its connection to L1 is via a multi-signature bridge, then you're not really scaling Ethereum," Vitalik believes. The original "rollup-centric roadmap" no longer reflects the evolution of the ecosystem. This shift comes as Ethereum's own Layer-1 scaling projects continue to advance, reducing the reliance on rollup block space capacity at the base layer.
According to core market data, the total value of Ethereum rollups has decreased by 13.2% year-over-year, down to $40.3 billion. This marks a significant retreat from the near $50 billion peak in mid-2025.
However, on the other hand, L2 transaction activity is surging. Rollups currently process about 3,470 user operations per second (UOPS), a substantial increase since early 2025. The rising activity indicates that rollups remain central to execution and low-cost transactions. But the decline in their safeguarding value suggests users and developers are increasingly viewing them as execution layers rather than repositories for large funds.
Vitalik's strategic shift is a direct response to Ethereum's own scaling progress. Ethereum's L2 strategy transformation depends on a technical upgrade: transitioning the underlying validation mechanism to zero-knowledge proofs. This is not a minor patch but a structural overhaul, promising to handle 10,000 transactions per second without sacrificing decentralization. The key is shifting from validators re-executing each transaction to only verifying concise transaction data. However, the risks are also significant. The upgrade cycle is complex and has not been extensively tested at scale. Therefore, this transition will take years, at least until 2027, introducing execution risks and market uncertainties.
Disagreements still exist within the developer community regarding the optimal architecture, which will further intensify friction.
As the underlying network expands, the original mission of L2 "scaling" is no longer necessary.

Ethereum's own progress has shifted the balance of power:
Improvements like PeerDAS and blob have increased throughput, while the introduction of ZK-EVM verification mechanisms provides a way to scale directly on Layer 1. These measures reduce dependence on external networks. Today, the Ethereum ecosystem is at a crossroads—L2 must prove its value beyond just temporarily solving high fees and network congestion.

2. The Future of L2
Vitalik states that Layer-2 networks including Arbitrum, Optimism, Base, and Starknet should shift from focusing solely on scalability to specializing in specific areas, such as privacy, identity, finance, social applications, and artificial intelligence. Vitalik's advice is clear: L2 networks should now compete in features like privacy, efficiency for specific applications, and non-financial use cases. This shift opens the door to specialized networks—some focusing on high-throughput gaming, some on confidential DeFi, and others on identity or supply chain tracking. The goal is to build a more diverse, application-driven L2 ecosystem rather than a single, unified scaling layer.

All current debates on this issue are market participants' early pricing. Whether L2 will truly live or die remains to be seen. Factors such as whether ZK verification can be stable and scalable, whether L1 scaling truly reduces marginal costs, and whether applications are willing to pay will all influence Ethereum's future ecosystem.
Vitalik is not denying Layer-2 itself but is rejecting the era of "L2 surviving solely on Ethereum narrative."