Ethereum's Glamsterdam hard fork — a combined execution-layer (Amsterdam) and consensus-layer (Gloas) upgrade — entered its final multi-client devnet phase on June 16, 2026, with ten EIPs locked for inclusion. The upgrade targets a 200-million gas-limit floor, roughly tripling current L1 capacity...
"This is probably the largest fork we've had since the Merge." — Parithosh Jayanthi, Ethereum Foundation DevOps Engineer (CoinDesk, June 2026)
Ethereum's Glamsterdam hard fork — a combined execution-layer (Amsterdam) and consensus-layer (Gloas) upgrade — entered its final multi-client devnet phase on June 16, 2026, with ten EIPs locked for inclusion. The upgrade targets a 200-million gas-limit floor, roughly tripling current L1 capacity from the 60-million range, while introducing protocol-level proposer-builder separation and parallel transaction execution. Mainnet activation is projected for Q3 2026, though no firm date is locked.
The fork arrives at a critical juncture. Ethereum L1 processes roughly 1 million transactions per day against a combined L2 throughput running at a 5:1 to 10:1 ratio. L2 fee payments to Ethereum collapsed over 90% year-over-year, from an estimated $113 million in 2024 to $10 million in 2025, according to Yellow Research. Monthly protocol revenue fell 60–80% in the quarters following EIP-4844 activation. Glamsterdam represents a direct attempt to make L1 execution competitive again — not by abandoning the rollup roadmap, but by widening the base layer's economic surface area.
Glamsterdam bundles ten Ethereum Improvement Proposals scheduled for inclusion, as listed under EIP-7773: EIP-7708, EIP-7732, EIP-7778, EIP-7843, EIP-7928, EIP-7954, EIP-7976, EIP-7981, EIP-8024, and EIP-8037. An additional 25+ EIPs remain under consideration, according to Phemex's technical review. The upgrade follows the Fusaka hard fork, which activated on December 3, 2025, at epoch 411392 (block 23,935,694).
The architecture separates into two functional layers. On the consensus side, the Gloas component restructures how blocks are proposed and validated. On the execution side, Amsterdam modifies how transactions are processed, priced, and parallelized. The combination changes the block propagation window from approximately two seconds to roughly nine seconds — the structural prerequisite for raising the gas limit to the 200-million target without degrading validator performance.
EIP-7732: Enshrined Proposer-Builder Separation (ePBS). This proposal moves the block-building and block-proposing functions into Ethereum's core protocol. Currently, 90%+ of Ethereum blocks flow through third-party MEV relays — principally Flashbots — that sit between proposers and builders. ePBS eliminates this external dependency by enshrining the separation on-chain. The stated objectives: reduce manipulation vectors for maximal extractable value (MEV), increase transparency in block construction, and remove a single-point-of-failure risk in relay infrastructure. Over $550 million in MEV is extracted annually from Ethereum alone, according to Calmops research. Flashbots Protect had saved users over 4,600 ETH in MEV and 2,200 ETH in gas fees by March 2026.
EIP-7928: Block-Level Access Lists (BALs). Blocks pre-declare the accounts and smart-contract storage slots they will interact with. When transactions touch non-overlapping state, validators can process them in parallel rather than sequentially. This is the foundational mechanism enabling higher throughput under the raised gas limit. Without BALs, tripling the gas limit would impose proportionally longer validation times on node operators, threatening decentralization.
EIP-7904 recalibrates gas costs across opcode categories. Current gas prices were set years ago and no longer reflect execution costs on modern hardware. The repricing makes compute-heavy operations cheaper while increasing costs for state-access operations, which carry higher long-term storage burden.
Quantified impact, according to CryptoRank and Phemex analysis:
As Jayanthi described to crypto.news: "High-level compute gets cheaper and state gets more expensive." The repricing is designed to push contract design toward lighter state footprints — a long-term sustainability measure as Ethereum's state database grows.
| Phase | Date | Status | |-------|------|--------| | Soldøgn interop devnet | ~May 2, 2026 | Concluded | | glamsterdam-devnet-5 | June 2026 | Running | | glamsterdam-devnet-6 | June 2026 | Running | | Public testnets | TBD | Pending devnet completion | | Mainnet activation | Q3 2026 (est.) | No firm date |
Everstake's June 2026 technical review cites end of August 2026 as the current best estimate. Jayanthi described the current phase as "the last phase before we work on hardening and then shipping the testnets." Historical precedent suggests a two-to-four-month gap between public testnet deployment and mainnet activation. Analysts at SpotedCrypto assign a 20–30% probability of further slip into Q4 2026.
The original target was June 2026. That date has passed without testnet deployment, confirming slippage of at least two months from initial aspirational timelines.
Glamsterdam's urgency is partly economic. Post-EIP-4844 (Dencun upgrade, March 2024), Ethereum's fee revenue structure changed materially:
The L1-L2 transaction ratio tells the structural story: combined L2 daily transactions run between 5:1 and 10:1 against Ethereum mainnet throughout Q1 2026, surpassing 10 million transactions on peak days versus Ethereum's roughly 1-million daily ceiling. Users pay L2 sequencers, not Ethereum validators, for those transactions.
Glamsterdam's effects on rollup economics are indirect but material. Lower L1 congestion reduces costs for the operations rollups depend on: proof settlement, bridge transactions, forced exits, and contract upgrades. The extended ePBS propagation window may further compress data-posting costs for rollups.
Current L2 market concentration, per SpotedCrypto:
L2 transfer costs are already minimal — Base and World Chain median USDC transfer: $0.02; OP Mainnet simple ETH transfer: approximately $0.0007. Glamsterdam's primary L2 benefit is operational cost reduction for rollup operators, not end-user fee savings, which are already negligible.
The throughput gap between Ethereum L1 and Solana remains large. Solana sustains 2,000–4,000 TPS during normal operation with spikes above 100,000 TPS in stress tests. Ethereum averages 15–20 TPS on mainnet.
Post-Glamsterdam, proponents cite up to 10,000 TPS-equivalent throughput under realistic workloads with the 200-million gas limit. That figure narrows the gap significantly but remains a design target, not a tested production number. Ethereum's aggregate throughput including L2s already exceeds 5,000 TPS.
The competitive framing has shifted. Ethereum positions as a modular settlement layer for high-value finance; Solana positions as a monolithic execution layer for consumer applications. Glamsterdam does not alter this strategic divergence — it strengthens the settlement layer's processing capacity.
As of June 15, 2026, per beaconcha.in and CoinLaw data:
The developer milestone, confirmed by Consensys co-founder Joseph Lubin who posted "We got there" on X, represents cumulative contributors to Ethereum L1, major L2s, and ecosystem repositories after filtering for bots.
Timeline risk. Glamsterdam has already slipped from its June 2026 aspirational target. Multi-client devnets are running, but public testnet deployment has not begun. A Q4 2026 or early 2027 mainnet activation remains plausible.
Fee-burn paradox. Cheaper L1 execution and continued L2 scaling may further reduce ETH burn rates, exacerbating the inflation dynamic. The 0.83% annualized inflation rate reflects an Ethereum that burns far less ETH than it issues — a reversal of the "ultrasound money" thesis.
Validator hardware requirements. A 200-million gas limit with parallel execution increases computational demands on node operators. The BAL mechanism is designed to mitigate this, but real-world performance under sustained load remains unproven.
MEV redistribution. ePBS eliminates relay dependencies but does not eliminate MEV extraction. Value may shift from relay operators to protocol-level auction mechanisms without reducing the total MEV burden on users.
Complexity risk. Ten EIPs in a single fork, affecting both execution and consensus layers, constitute the largest coordinated protocol change since the September 2022 Merge. Interaction effects between proposals are difficult to model fully in devnet conditions.
Glamsterdam is Ethereum's most ambitious single-fork upgrade since the Merge, targeting structural changes to block production, transaction execution, and gas economics simultaneously. The technical scope is large: ePBS removes external relay dependencies, BALs enable parallel processing, and gas repricing realigns costs with modern hardware capabilities.
The economic context is equally significant. Ethereum L1 has experienced a sustained decline in fee revenue as L2 activity grew and EIP-4844 compressed blob costs to near-zero. Glamsterdam does not reverse this dynamic directly — L2 transactions will continue to pay L2 sequencers — but it expands L1 capacity for direct execution, potentially attracting activity that currently bypasses the base layer entirely.
Whether the 200-million gas target and 10,000 TPS-equivalent throughput materialize in production remains unverified. The upgrade is in devnet, not testnet. The original June timeline has slipped. The gap between design targets and deployed reality will determine whether Glamsterdam represents a substantive L1 recovery or an incremental step in Ethereum's ongoing modular evolution.