Ethereum processed 200.4 million base-layer transactions in Q1 2026, the first quarter above the 200 million mark and a 43% increase from Q4 2025's 145 million. Active addresses reached 12.6 million. First-time users rose 82% quarter-over-quarter to 284,000. Capital inflows exceeded $2.1 billion,...
"More activity does not cleanly translate into more burn or more holder value." — CoinDesk analysis of Ethereum Q1 2026 on-chain data, April 17, 2026
Ethereum processed 200.4 million base-layer transactions in Q1 2026, the first quarter above the 200 million mark and a 43% increase from Q4 2025's 145 million. Active addresses reached 12.6 million. First-time users rose 82% quarter-over-quarter to 284,000. Capital inflows exceeded $2.1 billion, according to Artemis. By every usage metric, Ethereum just posted the strongest quarter in its ten-year history.
ETH, the token that is supposed to capture that value, traded at $2,328 as of April 17, 2026 — more than 50% below its August 2025 peak near $5,000 and roughly flat against its April 2021 price. Since the Merge in September 2022, Bitcoin has appreciated approximately 251%. Ethereum has gained roughly 40%. The ETH/BTC ratio hit 0.028 in February 2026, its lowest level since 2020, before recovering modestly to 0.0313.
This report examines the structural reasons why Ethereum's record network activity is not translating into token value, the economic mechanics of the Layer 2 fee model, and what the data implies for ETH as a productive asset.
Ethereum's quarterly transaction count followed a U-shaped trajectory. From a trough of approximately 90 million quarterly transactions in 2023, activity climbed to the 100–120 million range throughout 2024, reached 145 million in Q4 2025, and then jumped 43% to 200.4 million in Q1 2026.
The address-level data reinforces the picture. Artemis data shows 12.6 million active addresses, reflecting a reported 1,704% increase quarter-over-quarter. Some 284,000 first-time participants joined the network between January and March 2026, an 82% increase from the prior quarter.
Net capital inflows to Ethereum exceeded $2.1 billion in Q1, making it the leading blockchain by this measure for the period.
Yet ETH's market capitalization at approximately $280 billion reflects none of this momentum. The token is priced as though usage is declining, not surging.
The disconnect originates in Ethereum's fee structure. Daily gas revenue declined from approximately $23 million at its peak to $6.3 million. Average transaction fees fell to approximately $0.01 for basic ETH transfers in January 2026, down from around $0.41 in February 2025. By March 2026, average fees settled at $0.16–$0.22.
The cause is architectural, not accidental. The Dencun upgrade (March 2024) introduced EIP-4844, which created blob space — a dedicated data area for Layer 2 rollups that reduced their data posting costs by approximately 60%. Monthly L1 fee revenue, which had exceeded $100 million at its peak, fell below $15 million by Q4 2025.
Under EIP-1559, Ethereum's base fee is burned. Lower fees mean less burn. Less burn means the deflationary supply mechanism weakens. The protocol effectively traded revenue for scalability.
Ethereum's transaction fee revenue for Q1 2025 was approximately $217 million. January 2025 alone generated $150.8 million. By February, it had fallen to $47.5 million. The trajectory continued downward into 2026.
The Layer 2 ecosystem illustrates the value capture problem. In 2024, L2 networks collectively earned approximately $277 million in sequencer revenue and paid approximately $113 million to Ethereum L1 in data fees — a 41% capture rate. In 2025, L2s earned $129 million and paid approximately $10 million to L1 — a capture rate of roughly 8%.
The numbers for individual L2s underscore the shift. Base, Coinbase's L2, generated approximately $92 million in revenue in 2024 but paid only $4.9 million to L1 in blob fees, a 5% capture rate. Base's average daily revenue over the last 180 days stands at $185,291. Arbitrum generates approximately $55,025 per day.
Blob utilization, despite the increase in L2 activity, remains at approximately 29% of the 14-blob target introduced with Fusaka. The blob market is in structural surplus. L2s are posting data cheaply, and the fee floor introduced by EIP-7918 has not materially changed the economics.
The result: Ethereum L1 subsidizes the growth of L2 networks that capture the vast majority of sequencer revenue. Arbitrum holds $16.63 billion in TVL. Base processes the largest share of retail transactions. Both benefit from Ethereum's security guarantees while returning a fraction of their revenue to the settlement layer.
ETH supply at the Merge (September 15, 2022) stood at approximately 120,520,000. As of April 2026, supply has increased to approximately 120.7–121.5 million, a net addition of roughly 950,000 ETH. The network is technically inflationary at an annualized rate of approximately 0.23%.
Daily new issuance to stakers runs at approximately 1,700 ETH per day. During Q1 2025, the burn rate fell to 50–70 ETH per day. For Ethereum to achieve net-zero inflation — the breakeven point for the "ultrasound money" thesis — average gas prices need to sustain approximately 16 gwei. Current usage patterns fall well short of that threshold.
Total ETH burned since EIP-1559's activation in August 2021 is approximately 4.6 million. The mechanism works — but only at fee levels that the protocol's own scaling roadmap has deliberately reduced.
Pre-Merge, Ethereum inflated at 4–5% annually. The current 0.23% rate is an improvement by any standard. But the narrative marketed to investors was deflation, not merely low inflation. The gap between expectation and reality has a price impact.
Ethereum's validator set has grown to over 1.1 million active validators. Approximately 35.9–36 million ETH, or 29.6–31.1% of total supply, is staked. The staking ratio hit a record 31.1% in March 2026.
Staking yields have declined to 2.6–3.1% as the validator set expanded, driven by the inverse relationship between validator count and per-validator reward. Restaking protocols can temporarily boost combined yields above 8–15%, but base protocol yields are compressing.
The declining yield creates a competitive problem. U.S. Treasury yields, money market funds, and even some bank savings accounts offer comparable or superior returns with lower risk. The DeFi lending rate on major protocols has fallen below traditional savings account yields, according to concurrent webthreepedia research. ETH staking yields face the same compression.
Institutional engagement with Ethereum is accelerating even as ETH's price stagnates. BlackRock's BUIDL fund, a tokenized money market fund on Ethereum, has grown to over $1.8 billion in on-chain assets. Franklin Templeton, JPMorgan, and UBS have conducted live asset settlement pilots on Ethereum infrastructure.
Ethereum hosts approximately $180 billion in stablecoin supply, representing 60% of the global stablecoin market. The Ethereum Foundation deposited 22,517 ETH (approximately $46 million) into staking on March 30, 2026.
DeFi TVL on Ethereum stands at roughly $70 billion, with the network maintaining approximately 68% of total DeFi TVL across all chains. Aave leads at $26.46 billion, followed by Lido at $17.96 billion and Uniswap at approximately $6.8 billion.
The paradox: institutions use Ethereum as settlement infrastructure while the token that secures it declines. Stablecoin issuers, tokenized fund operators, and L2 sequencers all derive value from Ethereum's security and finality. But the economic design routes that value through fees — and fees are structurally low.
Ethereum's development roadmap has begun to address the value accrual gap, though incrementally.
Fusaka (December 2025) introduced PeerDAS, which distributes blob data across the network so each full node holds only approximately one-eighth of blob data. This cuts per-node bandwidth and storage requirements by 90% and enables eightfold blob capacity scaling. Crucially, Fusaka also introduced EIP-7918, a blob base fee floor tied to L1 execution gas cost. Fidelity Digital Assets estimated the floor would have generated approximately $78.6 million in additional blob revenue had it been active since Dencun.
Glamsterdam, expected in the first half of 2026, targets block building, validator-builder interactions, fee pricing, and censorship resistance at the protocol level. The upgrade aims to align fees with actual computational cost, creating what developers describe as a more rational fee market. The upgrade also introduces Enshrined Proposer-Builder Separation (ePBS) and parallel execution for up to 10,000 TPS theoretical capacity.
Neither upgrade directly solves the token value accrual problem. Both expand capacity, which at current demand levels means lower per-unit fees. The bet is that capacity creates demand — that cheaper transactions attract volumes sufficient to compensate for lower unit economics.
Ethereum's Q1 2026 data presents a case study in the difference between network utility and token value. The protocol has succeeded at its stated engineering goal: cheap, scalable transactions that serve as a settlement layer for a growing ecosystem of L2 networks, stablecoin issuers, and institutional users.
The cost of that success has been borne by ETH holders. The fee burn mechanism that was supposed to create deflationary supply dynamics requires fee levels that the scaling roadmap has deliberately lowered. L2 networks capture the majority of sequencer revenue. Staking yields have compressed toward traditional finance benchmarks. Supply is mildly inflationary.
Ethereum's value proposition may be shifting from "ultrasound money" — a narrative dependent on fee-driven deflation — to something closer to digital infrastructure: a settlement layer whose value derives from security guarantees and network effects rather than monetary scarcity. Whether that shift supports a $280 billion market capitalization, or something substantially different, depends on whether the protocol can recapture a meaningful share of the economic activity it enables.
The data does not resolve that question. It does establish that record usage, by itself, is insufficient.