On February 3, 2026, Vitalik Buterin published a lengthy article on X and the Ethereum community, with the core point being just one sentence: The roadmap established five years ago, which viewed L2 as the main scaling solution for Ethereum, has become obsolete.
When others say this, it sounds pessimistic; when Vitalik says it, it is a verdict.
Meanwhile, the market has given its own judgment - mainstream L2 tokens have plummeted over 90% from historical highs, Bitcoin's market share is close to 60%, and altcoins are collectively bleeding.
A brutal question lies before everyone: aside from a few cryptocurrencies like BTC and ETH that have already launched ETFs, do the tens of thousands of altcoins have any way out?
01
Vitalik's L2 "backlash"
For a long time, the valuation of L2 has been built on a core commitment— they can "inherit Ethereum's security".
However, by 2026, the reality is that the vast majority of leading L2s are still stuck in "stage 1" or even "stage 0", relying on centralized sorters and multi-signature bridges.
Vitalik's criticism is straightforward: an EVM chain with a processing capacity of 10,000 TPS, if its connection to L1 relies on a multi-signature bridge for adjustment, has not truly scaled Ethereum in a meaningful way, but merely established an independent platform based on trust.
In other words, most L2s are not extensions of Ethereum, but rather independent kingdoms with Ethereum's name.
Another key factor leading to the downgrading of L2 strategies is the evolution of Ethereum itself. The Fusaka upgrade activated in December 2025 introduced PeerDAS (peer data availability sampling), allowing validators to confirm availability by randomly sampling part of the Blob data. Through gradual BPO upgrades, the target capacity of the mainnet Blob increased from 6 to 14 (maximum 21), with plans to further increase to 48 by June 2026, significantly improving transaction processing capacity compared to the early stages of merging.
The gas limit for L1 has been raised to 60 million units, with future plans to further increase it to 100 million or even 200 million units; the Ethereum mainnet can now handle a large number of transactions that previously had to be outsourced to L2, and costs are still within a reasonable range.
L2 has been downgraded from being "Ethereum's scaling crutch" to a "specialized plugin". The new framework proposed by Vitalik is a "trust spectrum"—L2 is no longer Ethereum's "official shard", but must prove its existence by providing unique value that L1 cannot offer, such as privacy protection, ultra-low latency, or specific application optimization, rather than relying solely on cheap gas fees.
The era when "cheap and fast" could support a valuation of tens of billions is over.
02
The institutional decline of altcoins
If Vitalik's statement is the needle that bursts the L2 bubble, then the ETF is the pump that drains the liquidity of altcoins.
After the approval of Bitcoin and Ethereum spot ETFs in the US in 2024, institutional funds have surged into an extremely narrow channel. By the end of 2025, the assets under management for Bitcoin ETFs reached approximately $120 billion (with IBIT alone reaching $68 billion), while Ethereum ETFs climbed to approximately $18 billion.
Hedge funds, pensions, and family offices have gained secure exposure without the need to manage private keys. However, this influx of liquidity is exclusive—due to compliance and auditing requirements, institutional funds can hardly touch altcoins outside the top ten by market cap.
This is the "pump effect": after allocating core assets, institutions tend to choose public chains (like Solana and Chainlink) with clear technical barriers and compliance paths, even when pursuing higher risk returns, rather than spreading investments across tens of thousands of application layer tokens.
On the other end of the secondary market, the "star altcoins" launched in 2024 are undergoing collective valuation corrections. Most projects were driven up to tens of billions or even 100 billion in fully diluted valuation (FDV) during seed and private rounds by VCs, but on TGE, only about 12% of the circulation was released on average. The second quarter of 2026 is set to welcome a peak in large-scale token unlocks, with selling pressure looming.
More critically, there is a lack of development activity. Data shows that the share of so-called "blue-chip" projects with fewer than 10 commits per month on GitHub surged in 2025—without real developers, without a business model, only a token slowly approaching zero.
The plight of L2 tokens is particularly stark. Although L2 networks handled about 95% of the ecosystem's transactions in 2025, the prices of native tokens did not reflect this level of activity at all.
The reason is simple: after the two upgrades of Dencun and Fusaka, the data availability costs paid by L2 to Ethereum have decreased by more than 90%. User fees have dropped, but L2 can no longer profit from gas price differences. In 2025, the total revenue of the entire L2 industry plummeted by 53% year-on-year, down to approximately $129 million, with most of the revenue going to centralized sorter operators, leaving token holders with nothing.
The core use of tokens like ARB and OP remains limited to governance voting, with no staking rewards and no burn mechanism; the market has accurately labeled them as "worthless governance assets".
As long as sorters are still operated centrally by project parties, L2 tokens cannot serve as underlying secure collateral like ETH. Tokens cannot capture the consensus premium from network operations and naturally become worthless.
03
Survivor Game
The narrative of altcoins has overall collapsed, but not all tracks are dying. According to analysis by JPMorgan, the cryptocurrency market recorded an inflow of approximately $130 billion in 2025, while the capital inflow in 2026 is expected to be more led by institutional investors rather than the previous retail and corporate treasury activities.
The AI agent economy is forming a technological closed loop. The core narrative of 2026 is no longer the marketing slogan of "AI + blockchain", but the real landing of AI agents in autonomous trading and resource procurement.
The x402 protocol (released by Coinbase) allows AI agents to directly use stablecoins to pay for API services, computing power, and data costs through the HTTP 402 status code; ERC-8004 provides AI agents with on-chain identity and credit standards, together forming an autonomous trading infrastructure that requires no human intervention.
Decentralized computing projects like Render (RNDR) and Akash (AKT) increased AI inference capabilities in 2025, with tokens becoming "hard currency" for AI model training and execution—this demand supported by physical infrastructure provides real price support.
RWA tokenization has extended from government bonds to private credit and non-standard assets. BlackRock's tokenized fund BUIDL peaked at nearly $2.9 billion in 2025, and Chainlink's CCIP cross-chain interoperability protocol has covered over 11,000 banks globally through integration with SWIFT, becoming the de facto standard connecting traditional finance with the blockchain settlement layer, with its staking mechanism providing node operators with around 7% returns, outperforming most pure application tokens in this cycle.
The differentiated competition of high-performance public chains offers the market another vision. Solana's Firedancer client (launched on the mainnet in December 2025) demonstrated a transaction processing potential of millions per second during testing, with over 20% of validators migrating, establishing a moat in micropayments, high-frequency trading, and consumer applications. Sui attracted a large number of Asian game developers by utilizing parallel transaction processing and object-oriented architecture, with daily bridging inflows once exceeding Ethereum.
The common feature of these projects is that token value is driven by "machine demand" or "real cash flow", rather than retail speculation.
04
Summary
Vitalik's "negation" of L2 strategies essentially does not announce the end of L2, but rather denies the previous rough model that relied solely on scaling narratives to support token value.
The cryptocurrency market in 2026 is undergoing a monetization of cognition. As predicted by JPMorgan, this is no longer a question of "whether to start a bull market", but rather "whether one can survive in the restructuring of institutionalization and productivity".
BTC, ETH, SOL, and XRP are consolidating their monopoly positions through ETFs and compliance frameworks. For the remaining tens of thousands of altcoins, if they cannot establish a solid developer ecosystem and real cash flow by 2026, they will be completely marginalized by institutionalized currents.
Only those projects that adapt to AI agents, compliant RWA, and ultra-high-performance computing needs from the underlying architecture may find their own survival space under the radiance of Bitcoin.
The era of narratives is over, and the era of productivity has begun.
For everyone still in this market, the real question is just one: does the coin in your hand have any actual users?