Ethereum's Great Migration: Layer 2s Now Process Double the Mainnet's Transactions — and 50 Rollups Are Already Dead

Ethereum's mainnet processes roughly one million transactions per day. Its Layer 2 networks handle two million. That single statistic captures one of the most consequential shifts in blockchain history — and it is happening faster than almost anyone predicted.

But the story is not simply about more transactions happening elsewhere. It is about which rollups are capturing that activity, which are quietly dying, and what the entire migration means for the economic model that made Ethereum deflationary in the first place.

The Numbers Behind the Shift

By early 2026, Layer 2 networks collectively process approximately two million daily transactions, roughly double the Ethereum mainnet's volume. But the distribution of that activity is staggeringly uneven.

Three networks dominate the landscape:

• Base commands 46.58% of Layer 2 DeFi TVL and recently recorded 11.57 million transactions in a single 24-hour period — a figure that would have seemed implausible for any blockchain just two years ago.
• Arbitrum holds 30.86% of L2 TVL, averaging 1.5 million daily transactions and maintaining the deepest liquidity pools outside of Ethereum mainnet itself.
• Optimism rounds out the top three with approximately 800,000 daily transactions, anchored by its OP Stack framework that powers Base and an expanding roster of chain deployments.

Together, these three networks process nearly 90% of all Layer 2 transactions. The remaining 10% is spread across more than 50 competing rollups — many of which are, by any honest assessment, already dead.

The Zombie Chain Graveyard

The term "zombie chain" has entered the crypto lexicon to describe Layer 2 networks that technically continue to operate but have lost meaningful activity. The list is growing disturbingly fast.

Kinto has shut down entirely. Loopring closed its wallet. Blast's total value locked collapsed by 97%. Pirate Nation and Polygon zkEVM have ceased operations. According to 21Shares research, more than 40 Layer 2 projects are at risk of becoming zombie chains by the end of 2026 — networks that still produce blocks but move no real money.

In February 2026, Vitalik Buterin himself weighed in, publicly criticizing "copypasta" L2 chains that offer no meaningful differentiation. His comments underscored a growing consensus: the era of launching a rollup with a forked codebase, running an airdrop campaign, and hoping for organic adoption is definitively over.

Three factors are driving the consolidation:

1. The Dencun fee collapse. EIP-4844's blob transactions, live since March 2024, reduced L2 data posting costs by 90%. Arbitrum's gas fees dropped from $0.37 to $0.012. This eliminated the fee premium that differentiated rollups and triggered unsustainable price wars among competitors with no other moat.

2. The airdrop farming cycle. Many Layer 2 networks experienced artificial growth during token generation events, only to see usage collapse once farming incentives ended. The pattern is now so predictable that analysts can forecast the post-TGE activity cliff with startling accuracy.

3. Distribution beats technology. Base's dominance is not primarily a technical achievement. It is a distribution achievement. Backed by Coinbase's 100+ million user base, Base demonstrated that in the rollup era, access to users matters more than marginal improvements in execution speed or proving systems.

The ETH Tokenomics Paradox

The migration to Layer 2s created what analysts call Ethereum's "scaling paradox." The very success of Ethereum's rollup-centric roadmap is undermining the economic model that made ETH deflationary.

Here is the mechanism: EIP-1559, implemented in August 2021, burns a portion of every transaction fee on the Ethereum mainnet. When mainnet activity was high, this burn rate exceeded new ETH issuance, making the total supply shrink. Ethereum was deflationary — a powerful narrative for price appreciation.

But as activity migrates to Layer 2s, mainnet fees plummet. After the Pectra upgrade in May 2025, the average daily ETH burn fell to approximately 3.26 ETH per day — a 71% decrease from prior levels. Daily gas fee revenue declined from a peak of $23 million to $6.3 million. IntoTheBlock reported a 50% weekly drop in total ETH fees alongside a decline in daily active addresses from 350,000 to 280,000.

The math is straightforward: fewer mainnet transactions mean less fee burn, which means less deflation, which weakens a core investment thesis for ETH.

Yet the picture is more nuanced than the bears suggest. Total daily ETH burn held at roughly $1.2 billion through February 2026, still exceeding the network's 0.8% annual inflation rate. Ethereum remains technically deflationary — but barely, and the margin is shrinking.

The resolution to this paradox may lie in blob fees. EIP-4844 introduced a separate fee market for blob data that L2s post to Ethereum for security. As L2 transaction volumes grow, blob demand increases, and blob fees become a new revenue stream for the network. Analysts at VanEck estimate blob fees could generate $200 billion in cumulative value for Ethereum by 2030, effectively replacing the old mainnet fee burn model with a new data availability revenue model.

Why Distribution Won the War

The consolidation story is really a distribution story. Base did not win by building a faster rollup. It won by being Coinbase's rollup.

Built on the OP Stack, Base launched with immediate access to Coinbase's massive user base, fiat on-ramps, and compliance infrastructure. Consumer applications — the kind that generate sustained, non-speculative transaction volume — flocked to the network because users were already there.

Arbitrum took a different path, building the deepest DeFi ecosystem on any Layer 2 through aggressive grants programs, a robust developer community, and early mover advantage in attracting blue-chip protocols. Its $40.52 million in net inflows in January 2026 demonstrated continued institutional confidence.

Optimism's strategy is perhaps the most architecturally ambitious: rather than competing solely on its own chain, the OP Stack powers an expanding "Superchain" of interconnected rollups. Base is built on it. So are chains from Sony, Coinbase, and other major players. Optimism is betting that capturing the infrastructure layer will prove more valuable than winning any single chain war.

The lesson for the broader ecosystem is clear: in a world where rollup technology is increasingly commoditized, competitive moats come from users, liquidity, and institutional partnerships — not from marginal technical improvements.

What Comes Next: PeerDAS and the Blob Expansion

The roadmap ahead addresses both the scalability ceiling and the economics question.

PeerDAS, scheduled for late 2026, will increase the blob target from 3 to 16-32 blobs per block. This expansion ensures that growing L2 demand does not outpace Ethereum's data availability capacity — a concern that became real as blob space utilization approached saturation during peak activity periods.

The Glamsterdam fork, planned for the same timeframe, introduces parallel processing capabilities that could push theoretical throughput to 10,000 transactions per second on the mainnet. While most users will never transact directly on L1, increased mainnet capacity benefits L2s by reducing settlement costs and improving finality times.

Perhaps more importantly, the rollup ecosystem is beginning to solve its fragmentation problem. Cross-rollup bridging protocols, shared sequencer networks, and standardized messaging layers are reducing the friction of moving assets between L2s. The user experience of interacting with multiple rollups is converging toward something that feels like a single network — which was always the goal of Ethereum's modular scaling vision.

The New Ethereum

The Ethereum that emerges from this migration is fundamentally different from the one that existed even two years ago. Direct user activity on the mainnet is becoming minimal. The base layer is evolving into a settlement and data availability engine — the "court system" of a decentralized economy, as Vitalik has described it, invoked only when disputes need resolution or permanent records need anchoring.

The network's value proposition is shifting from "the world's most active smart contract platform" to "the most secure settlement layer for a constellation of execution environments." Whether that narrative supports ETH's valuation at current levels depends on whether blob fees and settlement demand can replace the mainnet fee revenue that made ETH deflationary.

For developers and projects building on Ethereum's Layer 2 ecosystem, the message is unambiguous: the window to launch a competitive general-purpose rollup has closed. The winners are decided. What remains is building applications on the platforms that won — and building them for users who are already there.

BlockEden.xyz provides high-performance RPC endpoints and API services across Ethereum and its leading Layer 2 networks, including Arbitrum and Optimism. Whether you are building on the mainnet settlement layer or deploying applications on the rollups that captured the market, our infrastructure is designed to keep pace with the migration.