Ethereum's Layer 2 networks processed more than 90% of ecosystem transactions in Q1 2026 while returning less than 8% of their revenue to the mainnet in data-availability fees. The result: a 90% year-over-year collapse in L1 value capture from L2 activity, according to L2BEAT and ethreportseth.x...
"We should think explicitly about economics of ETH. We need to make sure that ETH continues to accrue value even in an L2-heavy world." — Vitalik Buterin, Ethereum Co-Founder
Ethereum's Layer 2 networks processed more than 90% of ecosystem transactions in Q1 2026 while returning less than 8% of their revenue to the mainnet in data-availability fees. The result: a >90% year-over-year collapse in L1 value capture from L2 activity, according to L2BEAT and ethreportseth.xyz data compiled through March 2026. Ethereum's 30-day transaction fee revenue fell to approximately $10.3 million, placing it third behind Tron (~$25 million) and Solana (~$20 million), while Coinbase's Base chain alone generated $3.66 million in 30-day protocol revenue — three times Ethereum's own figure.
The December 2025 Fusaka upgrade introduced EIP-7918, a blob fee floor mechanism that prevents data-availability pricing from collapsing to negligible levels. Fidelity Digital Assets estimated the measure would have generated an additional $78.6 million (24,641 ETH) in cumulative blob-fee revenue had it been active since the Dencun upgrade in March 2024. Whether this repricing mechanism can close the value-extraction gap — or merely slow its widening — remains the central question for ETH holders in 2026.
The numbers tell a straightforward story. In 2024, Ethereum L2 networks earned $277 million in aggregate revenue and remitted $113 million to L1 in data-availability fees — a 41% capture rate. In 2025, revenue dropped to $129 million earned, with approximately $10 million paid to L1 — an 8% capture rate. The decline did not result from lower L2 activity. Daily active addresses on Base alone ranged between 550,000 and over 1 million in early 2026, representing roughly 70% of all L2 active addresses. Smart contract calls on the Ethereum ecosystem exceeded 40 million per day in February 2026, surpassing 2021 bull-market peaks.
The Dencun upgrade (March 2024) introduced blobs, a separate data-availability fee market that reduced L2 posting costs by 50-90%. The economic consequence was that L2 sequencer operators internalized a larger share of the spread between user fees and settlement costs. Base's data-posting costs to Ethereum mainnet fell to under $11,000 per month by late 2025, down from $3.8 million in February 2024.
Ethereum now hosts approximately $162 billion in stablecoin supply — 52% of the global market — yet this activity has not translated into proportional ETH value capture. As CryptoQuant noted in March 2026, "capital flows, rather than network activity, now explain ETH price dynamics more effectively."
March 2026 data from L2BEAT and ethreportseth.xyz shows a concentrated market. Base and Arbitrum together account for 77% of L2 DeFi TVL. The top three chains — Base, Arbitrum, and OP Mainnet — process approximately 90% of all L2 transactions.
| L2 Network | 2025 Revenue | Daily Revenue (Avg) | L2 DeFi TVL Share | DAU Range | |---|---|---|---|---| | Base | ~$93M | ~$185K | 46.6% (~$4.2B) | 550K-1M+ | | Arbitrum | ~$42M + $5M TimeBoost | ~$55K | 30.9% (~$2.8B) | 250-470K | | OP Mainnet | ~$26M | — | — | ~82K | | Others | <$5M each | — | — | Declining |
Base is the only consistently profitable L2. Its net profit in 2025 reached approximately $55 million. Most smaller L2s operate at losses. The "zombie chain" phenomenon is accelerating: Blast's TVL collapsed from a $2.7 billion peak to $55 million (a 97% decline), smaller L2 usage fell 61% since June 2025, and Kinto shut down entirely in September 2025 following a $1.6 million exploit.
Ethereum L1 30-day protocol revenue, at $1.22 million, placed it fifth — behind Base ($3.66M), Tron, Polygon, and Solana. The L2 ecosystem now generates more protocol revenue than the settlement layer it depends on.
The Fusaka upgrade, activated on December 3, 2025, introduced EIP-7918 as a direct response to the value-leakage problem. The mechanism establishes a dynamic floor price for blob fees tied to L1 execution gas costs. Blob fees cannot fall below 1/15.258 of the L1 execution base fee. This prevents the fee market from collapsing to 1 wei during periods of low demand — a condition that persisted on 93% of days between the Dencun and Fusaka upgrades, according to Fidelity Digital Assets.
The activation caused an immediate repricing. The blob base fee surged by approximately 15 million times from its pre-Fusaka level, per KuCoin data. Fidelity's analysis, published in November 2025, calculated that had EIP-7918 been active since the Dencun upgrade, it would have generated $78.6 million (24,641 ETH) in additional cumulative blob-fee revenue. Since the Prague-Electra upgrade alone (May 2025), the figure was $9.8 million (3,037 ETH). The additional cost per blob was estimated at approximately $6.02.
For Base specifically, Fidelity estimated an additional $30.6 million in annual blob fees under EIP-7918, compared to the $5.2 million Base paid in the prior year. This would raise Base's effective L1 capture rate from approximately 5.6% to roughly 38% — a material change, though still below the 41% rate seen in 2024 before Dencun compressed costs.
Post-Fusaka blob utilization stood at approximately 29% of the 14-blob target, with a median of 4 blobs per block. The January 2026 BPO2 parameter update raised the target to 14 blobs and the maximum to 21. Ethereum's PeerDAS implementation is expected to reduce bandwidth requirements by up to 80% for full nodes while supporting up to 8x higher blob throughput, further expanding capacity.
On February 18, 2026, Coinbase announced that Base would transition away from Optimism's OP Stack to a consolidated, independently managed codebase under a single GitHub repository (base/base). The move severed a three-year dependency that had channeled a substantial portion of Base's sequencer revenue to Optimism's treasury. By late 2025, Base was contributing between 70% and 96.5% of the sequencer fees flowing to the Optimism Collective — approximately 8,387 ETH out of roughly 14,000 ETH in total lifetime inflows.
The rationale was operational: Base processed roughly four times more transactions than Optimism, generated 144 times more decentralized-exchange volume, and produced 80 times more gas fees. The team stated it aims to double its upgrade pace to six major releases per year. OP token fell 4% in the 24 hours following the announcement.
The separation carries broader implications. Base was eligible for approximately 118 million OP tokens over six years under its original agreement. Its departure removes the largest revenue contributor to the Optimism Collective, raising questions about the sustainability of the Superchain model. Base's DeFi TVL stood at $3.85 billion at the time of announcement, and it accounted for 42% of new Ethereum ecosystem code deployed in 2024, with 4,287 active developers.
For Ethereum, the episode illustrates the governance challenge inherent in the rollup-centric roadmap: as L2s grow economically self-sufficient, their incentive to remit value to L1 weakens, absent binding protocol-level mechanisms like EIP-7918.
Ethereum's supply reached 120.69 million ETH as of April 2, 2026. Net issuance is positive at approximately 0.23% per year, according to 21Shares research. The "ultrasound money" narrative — premised on post-Merge fee burns exceeding issuance — has not held since Dencun structurally reduced mainnet fee generation.
The supply has grown by roughly 950,000 ETH since the Merge. With approximately 28.5% of ETH staked, staking yields are now derived primarily from consensus rewards and MEV rather than fee burns. 21Shares characterized ETH as "modestly inflationary during low-fee periods," noting that automatic deflation only returns during sustained activity spikes.
Ethereum's one-year change in realized capitalization turned negative in early 2026, indicating net capital outflows. The ETH price declined approximately 30% over a six-month period ending in March 2026, despite record network activity.
Projections for whether blob fees can restore deflationary dynamics vary. Analysts at Liquid Capital estimate blob fees may contribute 30-50% of total ETH burn by the end of 2026 if L2 adoption continues growing. Fidelity's data suggests EIP-7918 could add 200,000 to 400,000 ETH in annual burn, though this depends heavily on blob utilization rates reaching well above current 29% levels.
Zero general-purpose L2s have achieved Stage 2 decentralization as of March 2026. Five — Arbitrum, OP Mainnet, Base (as of January 2026), Scroll, and Starknet — have reached Stage 1, meaning they have functional fraud or validity proofs but retain upgrade mechanisms controlled by security councils.
Sequencer centralization remains the norm. All major L2s run single-operator sequencers, which order transactions and capture associated MEV. Shared-sequencer projects have not gained traction: Astria, once a leading shared-sequencer effort, shut down entirely in 2025. Arbitrum introduced a dynamic pricing model in January 2026, and zkSync's roadmap includes multi-node sequencing testnets in late 2025 with open participation targeted for 2026, though neither has delivered a production decentralized sequencer.
Buterin addressed this gap directly in February 2026, stating that the original rollup-centric roadmap "no longer makes sense" and that L2s relying on multisig-controlled bridges are "not scaling Ethereum." His revised framework positions L2s as a spectrum of Ethereum connectivity levels, with value measured by how closely they inherit mainnet security guarantees.
Ethereum's rollup-centric roadmap created a scaling solution that works technically but leaks value economically. L2s process the transactions, collect the fees, and remit a declining fraction to the settlement layer. EIP-7918 establishes the first protocol-level pricing floor, but the gap between what L2s earn and what L1 captures remains wide. Base's break from Optimism demonstrates that the most successful L2s are building toward economic independence, not deeper integration. Whether Ethereum's data-availability market can evolve into a durable revenue source — or remains a subsidized utility — will determine the economic coherence of the rollup model.