5 years later, Vitalik personally overturned the future roadmap he set for Ethereum back in the day! On February 3, 2026, he posted on X, stating: The original vision of treating Layer 2 as 'branded sharding' to solve scalability is no longer valid. This statement detonated like a bomb within the Ethereum community, effectively declaring the end of the mainstream narrative from the past 5 years, and the L2 camp is facing its greatest legitimacy crisis.
Looking back at 2020, Ethereum was crippled by high Gas fees: DeFi Summer + NFT craze, transfers often costing dozens or even hundreds of dollars, while Solana mocked 'Ethereum is dead' with its low fees and high TPS. Vitalik introduced the 'Rollup-centric' approach: L2 handles massive transactions, with the mainnet serving as a secure settlement layer, and L2 expands infinitely like 'branded sharding'. After the Dencun upgrade, L2 fees plummeted by 90%, and everyone thought a savior had arrived.
But reality is harsh: Most L2 projects are progressing slowly in terms of decentralization, with giants like Arbitrum, Optimism, Base, and Starknet still stuck in Stage 1, facing centralized sequencers, essentially serving as centralized databases disguised as blockchains. Project teams are raising huge amounts of capital (Arbitrum is valued at 1.2 billion USD, zkSync at 458 million USD), but due to MEV income, regulatory demands, and commercial pressures, they are unwilling to give up control. Vitalik criticized: 'Using a multi-signature bridge to connect L1's 10,000 TPS EVM is not scaling Ethereum!'
More crucially, Ethereum L1 has become stronger! From 2025 to 2026, the Gas Limit surged from 30 million to 200 million (the Glamsterdam upgrade introduced perfect parallel processing), and average transaction fees dropped to 0.44 USD or even 0.01 USD, nearly as low as Solana. Daily transaction volume on L1 broke a new high of 1.87 million transactions; why would users still endure the risks of L2 cross-chain bridges and fragmented liquidity?
Vitalik offered a new path for L2: Don't just focus on scalability, but differentiate in ways that L1 cannot achieve—privacy protection, specific application optimization (gaming, AI, social), ultra-low latency, and non-financial scenarios. L2 should become 'functional plugins' instead of the main forces for scaling. This is a return of power: The Ethereum mainnet reclaims sovereignty, and if L2 lacks unique value, it will become transitional.
This is not betrayal, it is growth. When the L1 Gas Limit hits 200 million by the end of the year, and transaction fees are just a few cents, users will vote with their feet. The leading L2s (Base relying on Coinbase traffic, Arbitrum/Optimism speeding up to Stage 2) may survive, while other ghost towns will go GG. The Layer 2 era ends, and Ethereum enters a new phase of 'L1 dominance + L2 specialization'!
