Ethereum's Layer 2 ecosystem is undergoing a violent consolidation. Of the 100+ rollups that launched over the past two years, three networks — Base, Arbitrum, and Optimism — now process nearly 90% of all L2 transactions. Blast's total value locked has collapsed 97%. Kinto and Loopring have shut ...
"We don't need more copypasta EVM chains." — Vitalik Buterin, Ethereum Co-Founder, February 3, 2026
Ethereum's Layer 2 ecosystem is undergoing a violent consolidation. Of the 100+ rollups that launched over the past two years, three networks — Base, Arbitrum, and Optimism — now process nearly 90% of all L2 transactions. Blast's total value locked has collapsed 97%. Kinto and Loopring have shut down entirely. Usage across smaller L2s has dropped 61% since mid-2025. The rollup gold rush is over, and what remains is a Darwinian shakeout that will define Ethereum's economic architecture for the next decade.
The catalyst was not market collapse but infrastructure success. Ethereum's Dencun upgrade slashed L2 data costs by approximately 90%, triggering fee wars that pushed most rollups deep into loss-making territory. Only Base — backed by Coinbase's 9.3 million monthly active traders — turned a profit in 2025, earning roughly $55 million while capturing 62% of total L2 revenue. The rest are burning through treasuries with no viable path to sustainability. Even Ethereum's co-founder has acknowledged the reckoning: Vitalik Buterin stated in February 2026 that the rollup-centric roadmap "no longer makes sense," calling on L2s to stop replicating generic EVM chains and instead specialize or perish.
This report examines the economic forces driving L2 consolidation, maps the winners and casualties, and analyzes what the emerging three-pillar structure means for Ethereum's value capture model.
The Ethereum scaling thesis was simple: move execution off-chain, post proofs on-chain, inherit Ethereum's security. Between 2023 and 2025, over 100 rollup mainnets deployed on Ethereum, fueled by Rollup-as-a-Service (RaaS) providers like Caldera, Conduit, and AltLayer that reduced launch costs to under $10,000. Every exchange, every DeFi protocol, every gaming studio seemed to want its own chain.
The result was a liquidity fragmentation crisis. Capital that once concentrated on Ethereum mainnet scattered across dozens of siloed environments, each with its own bridge, its own token, and its own thinly spread user base. By late 2025, L2BEAT tracked over 100 Ethereum-deployed mainnets. The vast majority had negligible activity.
The fundamental error was treating "launch a rollup" as a business model. A rollup is infrastructure — it generates revenue only when users pay fees. And when 50+ chains compete for the same pool of Ethereum-native users, the economics collapse into a race to zero.
Ethereum's March 2024 Dencun upgrade introduced EIP-4844 (proto-danksharding), which created a new "blob" data layer that reduced L2 data posting costs by approximately 90%. The upgrade was a technical triumph — and an economic wrecking ball.
Before Dencun, L2s could sustain margins by charging users modest fees while paying substantial costs to post data on Ethereum. The spread between user fees and settlement costs was the L2 business model. Dencun collapsed the cost side of that equation, forcing L2s into direct fee competition with each other. When your settlement costs drop 90%, every competitor's costs drop 90% too, and users migrate to whoever charges the least.
The Pectra upgrade in May 2025 compounded this by increasing blob capacity further, driving per-transaction costs toward fractions of a cent. For well-capitalized networks with large user bases, this was a gift. For small rollups with thin treasuries, it was a death sentence.
Base's average daily revenue reached approximately $185,000 per day in 2025. Arbitrum generated roughly $55,000 per day. Most other L2s failed to generate meaningful revenue at all — many producing less than $1,000 in daily fees while burning through millions in operational costs, team salaries, and ecosystem incentives.
The consolidation has produced a clear hierarchy.
Base dominates by every metric. Coinbase's L2 captured 62% of total L2 revenue in 2025, earning $82.6 million — a 30-fold increase from 2024. Its TVL peaked above $5.6 billion in October 2025, accounting for 46.6% of all L2 DeFi TVL. Base handles over 60% of all L2 transactions. The critical advantage is distribution: Coinbase's 9.3 million monthly active traders provide an unmatched onboarding funnel. Base is the only L2 that turned a profit.
Arbitrum holds the second position with approximately $2.4–2.8 billion in TVL and over 1.5 billion cumulative transactions. It represents roughly 31% of L2 DeFi TVL. Arbitrum's Timeboost feature, launched in April 2025, generated approximately $2 million in additional fee revenue — a modest sum that underscores how difficult monetization has become. Still, Arbitrum's deep DeFi ecosystem (Aave, GMX, Uniswap deployments) gives it structural stickiness.
Optimism takes a different strategic approach through its Superchain architecture, which captures fee portions from over 40 affiliated L2s. While its direct TVL is smaller at roughly $430 million, the Superchain model creates an aggregation layer that generates network effects across chains like Worldchain, Zora, and Mode. Optimism is not yet breakeven, but its architecture bets on becoming the "franchise model" of L2s.
Together, these three process nearly 90% of all L2 transactions and control the vast majority of L2 economic value.
The other side of consolidation is elimination. The 21Shares State of Crypto report, published in early 2026, described the situation bluntly: most current Ethereum L2s are unlikely to survive beyond 2026.
Full shutdowns:
Near-death collapses:
Zombie status:
The pattern is consistent with traditional platform economics: network effects concentrate activity around winners, and the marginal cost of switching between L2s (via bridges) is low enough that users follow liquidity. Once TVL begins draining from a smaller L2, the process accelerates — fewer users means fewer fees, means less development, means fewer users.
Perhaps the most significant signal came from Ethereum's co-founder. On February 3, 2026, Vitalik Buterin published a post stating that the original rollup-centric roadmap — which positioned L2s as Ethereum's primary scaling mechanism — "no longer makes sense."
His reasoning was twofold. First, progress among L2s toward decentralization has been "slower and more difficult than expected." Most L2s still rely on centralized sequencers, meaning they offer cheaper execution but not the trust-minimized security that was originally promised. Second, Ethereum itself is now scaling at Layer 1, with fees remaining low and gas limits expected to increase significantly throughout 2026.
Buterin was explicit about what he doesn't want to see: "Make yet another EVM chain and add an optimistic bridge to Ethereum with a one-week delay" — the approach taken by the majority of existing rollups. He called these "copypasta EVM chains" and argued they add no meaningful value to the ecosystem.
Instead, Buterin outlined two directions he views as legitimate: tightly integrated app-specific systems where Ethereum handles settlement, accounts, or verification; and institutional chains that publish cryptographic proofs to Ethereum. Everything else, he implied, is infrastructure overhead without economic justification.
This represents a fundamental shift in Ethereum's official narrative — from "L2s are the shards of Ethereum" to "L2s are a spectrum, and most of them aren't useful."
21Shares projects the surviving L2 landscape will coalesce around three structural pillars:
Pillar 1: Ethereum-Aligned Chains. Networks like Linea that route economic value back to Ethereum's base layer, maintaining tight integration with mainnet security and composability. These function as extensions of Ethereum rather than independent platforms.
Pillar 2: High-Performance Contenders. Specialized chains like MegaETH that target near-real-time execution for latency-sensitive applications. These compete on throughput and speed rather than cost, serving trading, gaming, and high-frequency use cases that Ethereum L1 cannot serve even with scaling upgrades.
Pillar 3: Exchange-Backed Networks. Base (Coinbase), Mantle (Bybit), Ink (Kraken), and BNB Chain (Binance) leverage massive existing user bases for distribution. These are the clearest winners because their business model doesn't depend on L2 fees alone — they drive volume to parent exchange businesses.
The economic implication is stark: the L2 layer is becoming an oligopoly. And in oligopolistic markets, value accrues to the few incumbents with distribution advantages, not to the best technology.
Viewed through the lens of economic value distribution, the L2 consolidation reveals a structural reality about blockchain economics: infrastructure layers cannot sustain more than a handful of profitable operators.
The subsidy problem persists. Most L2s that launched between 2023 and 2025 were funded by venture capital and token incentive programs. Their user bases were largely mercenary — attracted by airdrop farming, not genuine utility. When incentives dried up, users left. This mirrors the broader pattern identified across the blockchain ecosystem where 85–90% of all value flows remain subsidy-driven.
Fee revenue is insufficient for fragmented infrastructure. Total L2 fee revenue across all Ethereum rollups was approximately $120.7 million in 2025. Split across 50+ active chains, that averages to roughly $2.4 million per chain — far below the operational costs of maintaining a team, sequencer infrastructure, and ecosystem development. Only Base's dominant market share made the economics work.
Value is migrating to distribution, not protocol. The winning L2s all share one characteristic: they control user distribution through a parent entity (Coinbase, exchange integrations, Superchain affiliates). Pure-play rollups without captive user bases are discovering that "build it and they will come" does not apply to interchangeable EVM execution environments.
Ethereum L1 faces fee dilution. As L2s capture more execution, Ethereum's base layer collects less in fees. Post-Dencun, Ethereum burns only approximately 40,000 ETH annually while issuing roughly 960,000 ETH to stakers — creating 0.7–0.8% annual inflation. The L2 consolidation doesn't solve this: even dominant L2s pay minimal settlement costs to Ethereum, raising questions about L1's long-term value capture.
Three L2s now control 90% of all rollup transactions. Base (60%+), Arbitrum (~20%), and Optimism (~10%) have established dominance. The remaining 50+ rollups split the scraps.
Only Base was profitable in 2025, earning $55 million on $82.6 million in revenue. The fee compression triggered by Dencun has made standalone L2 economics unviable for most operators.
Vitalik Buterin has effectively retired the rollup-centric roadmap, calling generic L2s "copypasta" and stating Ethereum itself will provide sufficient blockspace in 2026.
The L2 layer is becoming an exchange-dominated oligopoly. Distribution through Coinbase, Bybit (Mantle), Kraken (Ink), and Binance (BNB Chain/opBNB) matters more than technical differentiation.
Smaller rollups face existential risk. Blast collapsed 97%. Kinto and Loopring shut down. Usage across remaining small L2s dropped 61%. Treasury runways are shrinking with no fee revenue to extend them.
Ethereum L1's fee capture problem deepens. The consolidation concentrates value at the L2 layer, while Ethereum mainnet drifts further into inflationary territory post-Dencun.
The Ethereum L2 ecosystem is not growing — it is contracting. The proliferation of 100+ rollups was never sustainable, and the market is now enforcing the economic discipline that hype cycles delayed. What emerges is an oligopolistic structure where three to five well-capitalized, distribution-advantaged networks capture nearly all L2 economic activity, while dozens of zombie chains slowly wind down.
For investors, the signal is clear: L2 tokens for chains outside the Big Three carry existential risk. For builders, the lesson is that launching a rollup is no longer a defensible strategy — you need either a captive user base, a differentiated use case, or deep integration with Ethereum's base layer. For Ethereum itself, the consolidation raises uncomfortable questions about whether its base layer can capture sufficient value when its most active users are increasingly one step removed, paying fees to Coinbase's chain rather than to Ethereum validators.
The rollup thesis is not dead. But the rollup gold rush is. What replaces it is a mature, concentrated market that looks far more like traditional cloud infrastructure — dominated by a few players, commoditized at the margins, and governed by distribution power rather than technical novelty.