Ethereum's Layer 2 networks collectively retained approximately $119 million in sequencer profit in 2025 while paying less than $10 million in blob fees back to the base layer — a value capture ratio below 8%. The imbalance has contributed to a 95% decline in Ethereum mainnet fee revenue from 202...
"If you create a 10,000 TPS EVM where its connection to L1 is mediated by a multisig bridge, then you are not scaling Ethereum." — Vitalik Buterin, Ethereum Co-Founder, February 2026
Ethereum's Layer 2 networks collectively retained approximately $119 million in sequencer profit in 2025 while paying less than $10 million in blob fees back to the base layer — a value capture ratio below 8%. The imbalance has contributed to a 95% decline in Ethereum mainnet fee revenue from 2021 peaks, pushed ETH into net inflation at 0.23% annually, and coincided with a 55% price decline from August 2025 highs. Standard Chartered estimates that Coinbase's Base chain alone removed $50 billion from ETH's market capitalization through fee diversion.
The structural question is no longer theoretical. Seventy-three active rollups secure over $48 billion in TVL, but three networks — Base, Arbitrum, and Optimism — process roughly 90% of all L2 transactions. Usage across smaller rollups declined 61% since June 2025, with 50+ chains operating as technical zombies. Vitalik Buterin acknowledged in February 2026 that the original rollup-centric roadmap "no longer makes sense," calling for a fundamental rethinking of how L2s relate to the base layer.
This report examines the economic mechanics of the L2-L1 value flow, quantifies the extraction gap, and assesses whether recent protocol-level interventions — including blob fee floors and direct L1 scaling — can realign incentives.
The economic relationship between Ethereum L1 and its L2 ecosystem inverted after the Dencun upgrade (March 2024) introduced blob transactions via EIP-4844. Prior to Dencun, L2 networks posted transaction data as calldata to Ethereum mainnet, paying standard gas fees. Post-Dencun, dedicated blob space reduced L2 data posting costs by 80–90%.
The financial outcome is stark:
| Metric | Pre-Dencun (2023) | Post-Dencun (2025) | Change | |--------|-------------------|---------------------|--------| | L2 total revenue | ~$274M | ~$129M | -53% | | L2 fees paid to Ethereum | ~$80M+ (calldata) | ~$10M (blobs) | -87% | | L2 retained profit | ~$194M | ~$119M | -39% | | L2 retention rate | ~71% | ~92% | +21pp |
The aggregate numbers mask a concentration effect. Base generated $185,291 in average daily revenue over the past 180 days — $67.6 million annualized — dwarfing Arbitrum's $55,025 per day ($20.1 million annualized). Together with Optimism, the three dominant networks capture the vast majority of L2 economics.
Daily gas revenue on Ethereum mainnet fell from approximately $23 million at peak activity to $6.3 million. Mainnet fees dropped roughly 95% to $0.01 per transaction by early 2026.
Coinbase's Base chain presents the clearest illustration of the value extraction dynamic. According to CryptoSlate, Base earned over $94 million in profit in 2024 while contributing just $4.9 million to Ethereum in blob fees — a 5.2% passthrough rate.
Base's operating economics in context:
Coinbase reported total Q1 2026 revenue of $1.4 billion, down 21% quarter-over-quarter, with a net loss of $394 million driven by unrealized crypto losses. Base's sequencer fees appear within Coinbase's "other transaction revenue" line item. The company disclosed that over 90% of onchain agentic stablecoin transactions in Q1 occurred in USDC on Base, positioning the chain as a vertically integrated stack serving Coinbase's institutional and retail businesses.
Standard Chartered's Global Head of Digital Assets Research, Geoffrey Kendrick, estimated in March 2025 that Base alone removed $50 billion from ETH's market capitalization by diverting fee revenue. The bank subsequently cut its year-end ETH price target from $10,000 to $4,000, citing "structural decline" from L2 fee cannibalization.
The "ultrasound money" thesis — that ETH would become deflationary through EIP-1559 fee burning — has been functionally suspended.
As of April 2026, approximately 120.7 million ETH are in circulation, growing at an annual inflation rate of 0.23%. Validator issuance runs at approximately 1,700 ETH per day. Daily burn rates have fallen well below issuance, driven by the collapse in mainnet fee revenue.
The blob mechanism has produced modest burn contributions. Ethereum earned approximately 212 ETH in blob fees over the 30 days preceding May 2026 — roughly 7 ETH per day, negligible against 1,700 ETH in daily issuance.
The BPO (Blob Parameter Override) schedule has expanded capacity aggressively:
These capacity expansions tripled blob space in a single month, further reducing per-blob pricing and compressing the already-thin revenue stream flowing back to L1. Core developers plan to reach 48 blobs per block by mid-2026, with a long-term target of 128 blobs per slot under full Danksharding.
The paradox: each capacity expansion makes L2 operations cheaper, attracting more activity, but simultaneously reduces the per-unit cost of settlement — meaning blob revenue grows only if volume scales faster than price declines. To date, volume has not kept pace.
The L2 landscape has undergone rapid consolidation. Of 73 active rollups listed on L2Beat, the top three — Base, Arbitrum, and Optimism — command approximately 77% of L2 DeFi TVL and process ~90% of all L2 transactions.
The remaining networks face existential economics:
Median per-transaction fees across major L2s in Q1–Q2 2026 illustrate the compression:
| Network | Median Tx Fee | |---------|--------------| | Base | $0.02 | | OP Mainnet | $0.03 | | Arbitrum One | $0.04 | | zkSync Era | $0.05 | | Scroll | $0.06 |
At these fee levels, smaller rollups without significant transaction volume cannot generate sufficient revenue to sustain development teams, security audits, and infrastructure costs. The result is a growing population of technically live but economically dead chains — what 21Shares termed "zombie chains."
Ethereum's core developers and the broader ecosystem have begun responding to the value extraction dynamic through several mechanisms:
1. EIP-7918: Blob Base Fee Floor (Fusaka Upgrade, December 2025)
This proposal established a minimum price floor for blob transactions. Even during periods of low L2 data demand, rollups now pay a minimum fee proportional to the execution base fee. The intent is to prevent blob burn from collapsing to near-zero during quiet periods. Early data suggests the floor is binding but the revenue contribution remains modest.
2. Vitalik Buterin's L2 Rethink (February 2026)
Buterin publicly reversed course on the rollup-centric roadmap, posting on Ethereum Research that the model of L2s as "branded shards" is "no longer viable." He defined scaling Ethereum as creating "large quantities of block space backed by the full faith and credit of Ethereum" and proposed that L2 networks support ETH by either burning a portion of their fees or permanently staking them and donating the proceeds.
3. Optimism's Buyback Mechanism (January 2026)
The Optimism community voted to allocate 50% of all network revenue to buy back OP tokens monthly, marking the first time a major L2 directly linked token value to network usage. This addresses the L2 token value problem — most L2 tokens are down 95% from peaks — but does nothing to increase value flow to Ethereum L1.
4. Arbitrum's Timeboost (Launched April 2025)
Arbitrum introduced express transaction ordering through an auction mechanism that has generated approximately $2 million in fees. The feature monetizes MEV at the L2 level — revenue that, in a pre-L2 world, would have accrued to Ethereum validators.
5. Direct L1 Scaling
Ethereum is pursuing aggressive L1 gas limit increases through the Glamsterdam upgrade, targeting 200 million gas and 10,000+ TPS. This approach, covered in a separate report, would reduce the economic rationale for L2s by making the base layer directly competitive on throughput and cost.
Beyond the economic extraction question, the sequencer centralization issue compounds the problem. All major L2s — Base, Arbitrum, and Optimism — operate centralized sequencers controlled by their respective organizations.
This creates a dual concern:
Buterin flagged this directly, noting that "progress among layer-2s toward later stages of decentralization has been slower and more difficult than expected." A high-throughput chain connected through a multisig-controlled bridge does not meet Buterin's definition of "scaling Ethereum."
The gap between the decentralized-infrastructure narrative and the centralized-sequencer operational reality is a risk factor that regulators are increasingly examining.
The economic relationship between Ethereum and its L2 ecosystem has evolved into what one former Ethereum Foundation researcher described as a principal-agent problem: L2 operators benefit from Ethereum's security guarantees while internalizing the majority of fee revenue. The original vision — that L2 growth would feed back into ETH value through increased blob demand — has not materialized at scale.
The data points toward a structural misalignment rather than a temporary market condition. Blob capacity expansions continue to outpace demand growth, compressing per-unit settlement costs. Dominant L2s are building independent economic moats through vertical integration (Base/Coinbase) and token buybacks (Optimism). Meanwhile, smaller rollups are dying at an accelerating rate.
Ethereum's response is bifurcated: protocol-level mechanisms like EIP-7918 address the margin, while the Glamsterdam upgrade's direct L1 scaling represents a more fundamental strategic pivot. Whether these interventions arrive fast enough to arrest the value extraction trend remains an open question.
What the data does not support is the thesis that the current L2 economic model is self-correcting. Without structural changes to how sequencer revenue, MEV, and blob fees are allocated, the gap between L2 profit capture and L1 value accrual will likely widen as the ecosystem scales.