The Ethereum L2 Shakeout Has Arrived, And Most Chains Will Not Survive

Three years after the rollup thesis became the defining bet in blockchain infrastructure, the evidence is in: most Ethereum Layer 2 chains are dying slowly, and a handful of application-specific or brand-backed networks are quietly absorbing every marginal user and dollar. The bifurcation is no longer a forecast. It is a structural fact visible in fee revenue, total value locked, and daily active address data across the entire L2 ecosystem.

The problem is not that scaling failed. It is that scaling succeeded so completely that commoditization set in faster than anyone predicted. As of May 30, the top five L2 networks by transaction volume account for more than 80% of all rollup activity while the remaining 50-plus chains in the long tail fight over scraps, according to data tracked by L2Beat. The Ethereum L2 shakeout is not coming. It is already here.

The Scale Of The Problem

The Ethereum Layer 2 ecosystem grew from roughly 10 live rollups in early 2023 to more than 80 by May of this year, according to L2Beat’s live tracker. That growth looked like validation of the rollup-centric roadmap. In practice, it produced a fragmented landscape where most chains cannot generate enough fee revenue to cover their own operating costs.

Aggregate L2 total value locked peaked at approximately $50 billion in March and has since pulled back to around $38 billion as of May 30. The contraction is not uniform. Arbitrum (ARB) and Base, Coinbase‘s (COIN) L2, together hold roughly 45% of all L2 TVL, leaving the remaining 35-plus chains to divide the rest, according to DefiLlama’s L2 dashboard.

> The top two L2 networks by TVL command a combined share that exceeds the combined share of the next 15 chains. This is not healthy competition. It is the early stage of a monopolar market.

Vitalik Buterin has repeatedly argued that the long-term architecture of Ethereum should resemble a hub-and-spoke model, with specialized chains inheriting Ethereum’s security. What nobody modeled cleanly was how quickly spoke differentiation would collapse into commoditization when the marginal cost of launching a rollup dropped to near zero.

Also Read: The Ethereum L2 Shakeout, How 90% Of Rollups Lost Their Reason To Exist

How EIP-4844 Broke The Cost Moat

Before March 2024’s Dencun upgrade, L2 operators paid substantial fees to post transaction data to Ethereum’s calldata layer. Those data costs created a natural floor: chains that could not generate enough transaction volume to amortize posting costs would fail economically. It was brutal, but it was a filter.

EIP-4844, which introduced blob-carrying transactions, reduced L2 data posting costs by between 90% and 99% depending on the chain. Arbitrum’s average transaction fee fell from roughly $0.50 to under $0.01 almost overnight. Optimism (OP) fees dropped similarly. For users, this was an obvious win.

> L2 data costs fell by up to 99% after EIP-4844 activated in March 2024, eliminating the economic filter that had previously culled unviable rollup operators.

For the competitive landscape, the consequence was more complex. When data costs collapsed, the last remaining cost-based differentiator between chains disappeared. An undifferentiated general-purpose L2 could now exist on near-zero operating cost, meaning the culling mechanism was removed before most weak chains were forced to exit. The result is a zombie cohort: chains that are technically live, maintain Discord servers, issue governance tokens, and attract zero organic usage.

Also Read: Dow Hits Record as Ceasefire Hopes Lift Markets, Broadcom Drags Nasdaq Lower

Where The Users Actually Are

Daily active address counts tell the story more starkly than TVL does. TVL is sticky because it reflects locked capital, often from early insiders and incentivized liquidity programs. Daily active addresses reflect genuine user behavior.

Base surpassed Arbitrum in daily active addresses for the first time in Q4 2025 and has maintained that lead into this year, driven by Coinbase’s consumer product integrations and its social app ecosystem, according to Dune Analytics dashboards tracking L2 activity. zkSync Era and Polygon zkEVM saw daily active address counts fall 60% and 72% respectively between September 2025 and May of this year, per the same datasets.

> Base crossed 2 million daily active addresses in April, a figure that places it in the same league as several prominent Layer 1 chains, while dozens of named L2s recorded fewer than 500 daily active addresses on the same day.

The bifurcation maps neatly onto distribution advantage. Base inherits Coinbase’s 100-million-plus verified user base. Arbitrum built a DeFi ecosystem deep enough that protocol-level liquidity creates its own gravity. Every other general-purpose L2 is effectively asking users to bridge assets to a chain that offers the same DeFi experience as Arbitrum or the same consumer apps as Base, but without the brand trust or liquidity depth.

Also Read: SpaceX Lines Up Retail Investors for Record $75 Billion IPO

The Revenue Math Does Not Work For Most Chains

Fee revenue is the cleanest measure of whether a chain has found product-market fit. A chain that generates fees is a chain with demand. A chain that must continuously emit tokens to attract TVL and active addresses is running a subsidy program, not a network.

As of May 2026, L2Beat’s revenue module shows that fewer than 10 L2 networks generate enough gross fee revenue to cover their sequencer operating costs without relying on token incentives. Base, Arbitrum, and OP Mainnet together account for more than 90% of all L2 fee revenue. The fourth and fifth largest revenue generators are application-specific chains, not general-purpose rollups. The rest of the field generates revenues measured in thousands of dollars per month.

> The top three L2s by revenue capture more than 90% of all rollup fee income, leaving roughly 70 other live chains to split the remainder.

Token incentive programs have masked this dynamic for years. Protocols like Blast and Scroll ran aggressive airdrop campaigns that inflated active address metrics through Sybil activity. Academic analysis of airdrop-driven onchain activity, including work published on SSRN examining farming behavior across EVM chains, consistently shows that address counts inflated by token incentives revert toward baseline within 60 to 90 days of distribution. That reversion is now visible across most of the 2024 and 2025 airdrop cohorts.

Also Read: Treasury Yields Ease as Traders Brace for May Jobs Report

The Three Archetypes That Are Surviving

Across the L2 landscape, three structural patterns characterize the chains that are not only surviving but growing. Understanding these archetypes is more useful than studying any individual chain in isolation.

The first archetype is the brand-anchored chain. Base is the clearest example. Its survival is not primarily a technical story. Coinbase’s compliance infrastructure, its fiat on-ramp dominance, and its regulatory standing in the United States give Base a distribution moat that no independent rollup team can replicate. Kraken‘s forthcoming NFS chain and Robinhood‘s reported L2 exploration follow the same logic.

The second archetype is the application-specific rollup. Chains like dYdX (which migrated to a Cosmos (ATOM) appchain but proved the concept), ApeChain for the Yuga Labs NFT ecosystem, and Xai for gaming demonstrate that a rollup built around one application or one community retains users because the users have no rational reason to leave. The application is the moat, not the chain technology.

The third archetype is the ZK-native chain with a verifiable technical advantage. Starknet and Scroll occupy this space, though Scroll has struggled with user retention post-airdrop. The verifiable proof advantage matters most for institutional users who require cryptographic guarantees for compliance purposes, a demand that is growing as tokenized real-world assets scale.

> The three surviving L2 archetypes are brand-anchored distribution chains, application-specific rollups with captive demand, and ZK-native networks serving institutional compliance requirements.

Also Read: Cod Prices Squeeze British Fish and Chip Shops

Why General-Purpose Chains Have No Reason To Exist

CoinDesk’s June 4 analysis put it bluntly: many general-purpose chains no longer have a reason to exist. The logic is structural, not sentimental.

A general-purpose EVM rollup in 2026 offers the same smart contract environment as Ethereum mainnet, the same DeFi protocols that have already deployed on Arbitrum and Base, and faster finality at lower cost. The finality and cost arguments were compelling in 2021 and 2022. They are not compelling today because every other chain in the category offers the same thing.

Differentiation in blockchain networks historically comes from one of three sources: performance, ecosystem liquidity, or community identity. General-purpose L2s launched after 2023 rarely have any of the three in sufficient quantity to compete. Performance is commoditized by shared proving infrastructure. Ecosystem liquidity flows to the chains already at the top of the TVL rankings. Community identity requires years of brand-building that most rollup teams have neither the budget nor the patience to pursue.

The academic framing for this dynamic is Metcalfe’s Law applied to liquidity networks. A 2023 paper available on arXiv examining network effects in DeFi protocols found that liquidity concentration follows a power law, with the top network in any category capturing a disproportionate share of activity as the total participant count grows. The implication for L2s is that fragmentation across many general-purpose chains does not produce a competitive market. It produces a temporary fragmentation that resolves toward consolidation once switching costs fall low enough for users to express preferences.

Also Read: Bitcoin Treasury Firms Lose $62 Billion in Market Value

The Superchain And AggLayer As Consolidation Plays

The most sophisticated response to the fragmentation problem has come from the two largest L2 ecosystems, and both responses involve absorbing smaller chains rather than competing with them.

Optimism’s Superchain framework invites other rollups to deploy on the OP Stack and share sequencing infrastructure, security, and eventually a unified messaging layer. As of May 2026, more than 20 chains have deployed on the OP Stack, including Base, Mode, Zora, and Redstone. The Superchain strategy turns what would otherwise be competitors into distribution nodes for the broader Optimism ecosystem.

Polygon’s (MATIC) AggLayer takes a different approach. Rather than requiring chains to adopt a specific execution environment, AggLayer uses ZK proofs to create unified liquidity across chains that connect to the aggregation layer. Polygon co-founder Sandeep Nailwal told Bloomberg in April that AggLayer already connects more than 15 chains and is targeting 100 by end of year.

> The Superchain and AggLayer together represent a consolidation play disguised as a scaling play: both frameworks absorb independent rollups into a shared economic layer controlled by the originating team.

Neither framework is neutral. Chains that join the Superchain cede some sequencer revenue and governance autonomy to Optimism’s governance structure. Chains connecting to AggLayer inherit Polygon’s ZK stack dependencies. The trade-off is survival in exchange for independence, and most small L2 teams will take that deal because the alternative is irrelevance.

Also Read: Trump Deploys Wartime Powers for $700M Coal Push

Tokenomics And The Governance Token Trap

Most independent L2 networks launched governance tokens as a combination of fundraising mechanism, community incentive, and theoretical decentralization signal. In practice, L2 governance tokens have become the clearest indicator of which chains are in structural decline.

When a chain’s native fee revenue is insufficient to sustain sequencer operations, the team faces a binary choice: raise external capital or use token inflation to pay for ongoing costs. Token inflation is the path of least resistance, and it is the path most L2 teams have taken. The result is a predictable death spiral. Inflation increases token supply and depresses price. Lower token price reduces the purchasing power of future emission budgets. Teams then increase emission rates to maintain dollar-denominated incentive levels, accelerating the spiral.

The OP token and ARB token have both avoided the worst version of this dynamic because fee revenue from mainnet activity provides a genuine cash flow base. OP Mainnet generated approximately $4.2 million in gross fee revenue in April, according to L2Beat revenue data. That revenue funds sequencer costs and contributes to the Optimism Collective treasury without requiring inflationary emissions.

> For the 70-plus L2s generating less than $100,000 in monthly fee revenue, governance token inflation is not a growth strategy. It is a slow liquidation of founder and early investor stakes disguised as community reward.

An analysis of 30 L2 governance token trajectories published in the Electric Capital Developer Report 2025 found that token price and developer retention are the two metrics most predictive of long-term chain survival. Chains that lost more than 50% of their active developer count over a 12-month period universally saw token prices fall by 70% or more. The developer drain from smaller L2s to Base, Arbitrum, and Ethereum mainnet has been consistent since Q3 2025.

Also Read: South Korea’s Jensen Huang Obsession

Institutional Demand Is Reshaping The Priority Stack

One structural shift that is not yet fully priced into L2 market structure is the arrival of institutional demand for blockchain infrastructure. Tokenized real-world assets, institutional DeFi, and regulatory-compliant settlement layers are creating a demand profile that was absent from the original rollup design thesis.

The scale is measurable. Total tokenized real-world assets on public blockchains surpassed $20 billion in April, according to rwa.xyz, with the majority sitting on Ethereum mainnet or on a small number of compliant chains. BlackRock (BLK) and Franklin Templeton (BEN), two of the largest asset managers in the world, have both deployed tokenized fund products on Ethereum and select L2s.

For institutional users, the choice of L2 is not primarily a fee or performance decision. It is a legal, compliance, and counterparty risk decision. Institutions need chains where sequencer operators are identifiable entities, where data availability is provably sufficient for audit purposes, and where the proving system can be independently verified. These requirements favor ZK-native chains and brand-anchored chains with regulated operator entities over anonymous sequencer teams.

> Institutional tokenized assets demand provable compliance infrastructure, a requirement that effectively disqualifies most anonymous or under-resourced L2 operators from serving this market.

Coinbase is particularly well-positioned here. Its regulatory standing, its relationships with institutional custody clients, and Base’s technical architecture create a compliance surface that competitors cannot easily replicate. The firm’s decision to tie its first Fannie Mae-backed mortgage product to Bitcoin (BTC) collateral, reported by multiple outlets in June, signals an ambition to become the primary regulated crypto infrastructure provider across multiple asset classes.

Also Read: India’s RBI Holds Rates Steady Amid Iran War Pressure on the Rupee

What Consolidation Will Actually Look Like

Consolidation in the L2 ecosystem will not resemble the sharp collapses of the 2022 Terra-Luna implosion or the FTX bankruptcy. It will be a slow, unglamorous process of activity migration, team attrition, and eventual chain abandonment.

The most likely consolidation path for most small L2s is migration into a Superchain or AggLayer framework, rebranding as an application-specific chain serving one protocol or community, or simply losing developer activity until the chain becomes a ghost network maintained by automated sequencers posting empty blocks.

A smaller number of chains will attempt to raise additional venture capital to extend runway and build toward differentiation. The appetite for that capital is constrained. Cryptocurrency venture funding in Q1 of this year totaled approximately $4.9 billion across all deals, according to Galaxy Digital’s quarterly VC report, down from the $10-billion-plus quarters of the 2021 to 2022 cycle. Infrastructure deals, which dominated 2022 and 2023 funding, are losing wallet share to application layer and AI-crypto intersection bets.

> The L2 consolidation will not be a crash. It will be a slow draining of developer attention, user activity, and venture backing from the long tail toward three to five dominant networks over 18 to 24 months.

The timeline matters for token holders and founders. Teams that move early to find a strategic home within a larger ecosystem, whether through the Superchain, AggLayer, or an acquisition, will preserve optionality. Teams that wait for market conditions to improve are likely to find that the window for a credible pivot closes before the market turns.

Also Read: Cardano Craters 16% as Broadcom AI Miss Triggers Broad Crypto Selloff

The Chains Most Likely To Emerge From The Shakeout

Identifying the survivors requires applying the three archetype filters to the current field, not ranking chains by current TVL or token price. Current TVL rankings are lagging indicators. Structural position is what matters for the 18-month horizon.

Base has the strongest structural position in the entire L2 field, combining brand-anchored distribution, regulatory standing, and an application ecosystem that is growing through Coinbase’s consumer product strategy rather than through mercenary liquidity. Arbitrum retains the deepest DeFi protocol stack and the longest institutional relationships in the independent rollup category. Its governance community is the largest and most active of any L2, a leading indicator for developer retention.

Among ZK-native chains, Starknet’s Cairo-based execution environment creates genuine technical lock-in for applications built natively in Cairo, a smaller but defensible moat. zkSync Era’s developer tooling ecosystem, built around its own ZK stack, serves a similar purpose.

Among application-specific chains, the survivors will be those tied to protocols or communities with genuine user demand that is independent of token incentives. ApeChain, gaming-focused chains with live titles, and financial application chains built around specific regulated products all fit this description.

> The L2 survivors share one characteristic above all others: they have a reason to exist that is separate from the generic promise of cheap, fast Ethereum transactions.

The losers share a different characteristic. They are chains that launched between 2023 and 2025 with a general-purpose EVM environment, an airdrop campaign, a governance token, and no application or community that would be meaningfully harmed by the chain’s disappearance. By that description, more than half of the current L2 field is a candidate for eventual abandonment.

Read Next: China Targets Micro Drama Violence and Misogyny in Two-Month Crackdown

Conclusion

The Ethereum Layer 2 shakeout is a predictable consequence of two intersecting forces: the commoditization of rollup infrastructure through shared proving and sequencing toolkits, and the fee compression triggered by EIP-4844 that removed the last economic filter separating viable chains from zombie projects. The result is a market where most chains are technically functional and economically insolvent at the same time.

The survivors are not necessarily the most technically sophisticated chains. They are the chains with the most defensible distribution advantages, whether those advantages come from a regulated institutional parent, a captive application community, or a ZK proof system that satisfies compliance requirements that competitors cannot meet. Technology is table stakes in a commoditized infrastructure market. Distribution and compliance surface are the actual moats.

For anyone allocating attention, capital, or developer hours across the L2 ecosystem, the analytical frame should shift. The question is no longer “how many transactions per second can this chain process?” The question is “what user or application exists that would be meaningfully harmed by this chain’s disappearance?” If the honest answer is “none,” the chain’s long-term prospects are poor regardless of its current TVL or token price. The shakeout has arrived, and it will be slow enough to be denied until it is complete.

Read Next: Asia Tech Stocks Slide After Broadcom Earnings Shock Rattles AI Trade

—

Daniela Kirova

Senior Writer

Daniela Kirova is a finance and cryptocurrency journalist at Nonce Media. Her writing covers economics, digital assets, technology, and innovation, with a focus on making complex financial topics accessible to broad audiences. A multilingual translator fluent in English, German, and Bulgarian, she brings a background in psychology to her analysis of market behavior and investor sentiment.