Ethereum's Glamsterdam upgrade activated on the Sepolia test network on September 28, 2026 at 14:44:48 UTC, marking the first public testnet deployment of a protocol overhaul that bundles ten EIPs across the consensus and execution layers. The upgrade targets a 3.3x increase in the block gas limi...
"This is probably the largest fork we've had since the Merge. We're working on devnets with all the EIPs in them right now." — Parithosh Jayanthi, Ethereum Foundation Test Network Coordinator
Ethereum's Glamsterdam upgrade activated on the Sepolia test network on September 28, 2026 at 14:44:48 UTC, marking the first public testnet deployment of a protocol overhaul that bundles ten EIPs across the consensus and execution layers. The upgrade targets a 3.3x increase in the block gas limit — from the current ~60 million to 200 million — and introduces two structural changes to how blocks are built and processed: Enshrined Proposer-Builder Separation (EIP-7732) and Block-Level Access Lists (EIP-7928).
The Sepolia fork proceeded without dissent from six client teams. A second testnet activation on Hoodi is tentatively scheduled for October 26, 2026, with mainnet deployment projected in Q4 2026 but not yet confirmed. ETH traded at $2,690 on the day of activation, with Ethereum's market capitalization at $328.4 billion.
The scope of Glamsterdam is material. It restructures the validator-builder relationship that currently routes over 90% of Ethereum blocks through third-party MEV-Boost relays, and it lays the infrastructure for parallel transaction execution — a prerequisite for sustained gas-limit increases without proportional growth in hardware requirements.
Glamsterdam is a coordinated hard fork combining consensus-layer changes (codenamed "Gloas") and execution-layer changes (codenamed "Amsterdam"), specified under EIP-7773. It follows the Fusaka upgrade (December 2025) and precedes Hegotá, tentatively scheduled for H1 2027.
The upgrade addresses three objectives: faster transaction processing through parallelization, faster node synchronization via snap v2, and long-term sustainability against state database bloat. According to Ethereum Foundation researcher Toni Wahrstätter, Glamsterdam represents "one of the biggest forks seen yet," with "things looking great" as the upgrade approaches its final testing phase.
The fork transition on the Lodestar client was measured at approximately 500 milliseconds, down from roughly 20 seconds in earlier devnet iterations — an indicator of client readiness improvements during the multi-month devnet testing cycle that began in early 2026 and concluded with the Soldøgn interop devnet on May 2, 2026.
Currently, over 90% of Ethereum blocks are routed through MEV-Boost relays — off-chain infrastructure operated by third parties including Flashbots, BloXroute, and Titan Relay. The top two builders (Titan and Quasar) construct approximately 73% of all blocks, with the top four builders accounting for over 95%.
EIP-7732 moves the proposer-builder handoff and payment process into Ethereum's protocol rules, removing the network's dependency on external relay infrastructure. The change introduces a Payload Timeliness Committee (PTC) — a new validator committee duty that verifies whether builders deliver block payloads on time. Every solo staker and institutional operator will need to update their client software to support this new duty.
The proposal extends the data propagation window from approximately 2 seconds to 9 seconds, creating headroom for larger blocks and higher blob capacity. According to Figment Staff Protocol Analyst Benjamin Thalman, the upgrade addresses two primary institutional staker concerns: block building architecture improvements and faster exit liquidity for large staked ETH positions.
Block-Level Access Lists map every account and storage slot touched per block, along with post-execution state values. Nodes receive this map before processing, enabling three downstream capabilities:
BALs do not deliver parallel execution on their own. They provide the data structure that makes parallel execution technically feasible and safe at higher throughput levels.
Beyond the two headline proposals, Glamsterdam bundles eight additional EIPs:
| EIP | Function | |-----|----------| | EIP-8037 | Sets fixed cost per byte of new state; creates a separate gas reservoir for state growth | | EIP-7976 | Increases calldata floor cost | | EIP-7981 | Increases EIP-2930 access list cost | | EIP-7954 | Raises maximum deployed contract code size from 24 KiB to 64 KiB; initcode limits from 48 KiB to 128 KiB | | EIP-7778 | Removes gas refunds from block-level accounting | | EIP-8024 | Adds SWAPN, DUPN, and EXCHANGE stack-manipulation opcodes | | EIP-7843 | Adds SLOTNUM opcode to expose the current slot number | | EIP-7708 | Makes ETH transfers and burns emit logs, similar to ERC-20 token events |
The gas repricing EIPs (7976, 7981, 8037) collectively address state growth sustainability. EIP-8037 is particularly notable: by carving out a dedicated gas reservoir for state creation, it separates the cost of adding new state from general computation costs — a mechanism designed to prevent long-term database bloat that currently affects full node operators.
The target gas limit under Glamsterdam is 200 million per block, up from the current approximately 60 million. This represents a 233% increase in base-layer capacity.
Projected impact according to Datawallet analysis:
The gas limit increase is not automatic at fork activation. It requires validators to signal support by adjusting their client configurations, following the same social consensus mechanism used in previous gas limit adjustments. BALs (EIP-7928) provide the technical safety net — without the ability to identify parallelizable transactions, a 3.3x gas limit increase would impose proportionally higher hardware requirements on validators.
Calldata repricing EIPs included in Glamsterdam directly affect Layer 2 rollups that post transaction data to Ethereum L1. Lower calldata costs reduce L2 settlement expenses, which flow through to lower user-facing transaction fees on networks including Arbitrum, Optimism, and Base.
The upgrade does not pivot Ethereum away from its rollup-centric roadmap. It hardens L1 as a higher-capacity settlement and data availability layer. Combined with the blob capacity expansion from Dencun (March 2024) — which reduced L2 costs by approximately 90% — Glamsterdam continues the pattern of making L1 a more capable substrate for L2 operations.
EIP-7708, which makes ETH transfers emit logs like ERC-20 events, has secondary utility for rollup indexers and cross-chain bridges that currently require custom logic to track native ETH movements.
| Milestone | Date | Status | |-----------|------|--------| | Soldøgn interop devnet | May 2, 2026 | Completed | | Final devnet phase | June 2026 | Completed | | ACDC Call #186 confirmation | September 3, 2026 | Completed | | Sepolia testnet fork | September 28, 2026 at 14:44:48 UTC | Activating | | Hoodi testnet fork | October 26, 2026 (tentative) | Pending | | Mainnet activation | Q4 2026 (no confirmed date) | Pending |
The Sepolia fork activated at epoch 351232, slot 11,239,424. Six client teams confirmed readiness without dissent. During the September 3 All Core Developers Consensus call, Lido representative Dima Gusakov noted a preference for "both dates a bit later — like a week or so to have more room" for testing, but the proposed schedule was adopted.
Barnabas Busa, an Ethereum Foundation developer, underscored the stakes of the testing phase: "If we have 4-5 months of non-finality, it's game over."
The Hoodi testnet fork on October 26 remains contingent on stable devnet performance. Ethereum's official roadmap lists mainnet deployment as expected in Q4 2026, but the specific date has not been confirmed. The realistic activation window spans November to December 2026.
Two notable proposals were deferred from Glamsterdam:
EIP-7782 (6-second slots): Shelved due to immature implementation and conflicts with ZK-proving timelines. Reducing slot time from 12 seconds to 6 seconds would double throughput but requires ZK provers to generate proofs within tighter windows — a capability not yet production-ready.
FOCIL/EIP-7805 (Fork-Choice Inclusion Lists): Moved to Hegotá (H1 2027) to avoid interaction complexity with ePBS. Inclusion lists are designed to limit censorship by ensuring certain transactions cannot be excluded from blocks, but their interaction with the new proposer-builder separation mechanism required additional design work.
Additionally, EIP-7610 was cancelled and superseded by EIP-8253, which targets Hegotá.
Glamsterdam is the most structurally significant Ethereum upgrade since the Merge in September 2022. Where the Merge changed the consensus mechanism, Glamsterdam changes how blocks are built, validated, and processed — and it does so while targeting a 3.3x expansion in base-layer capacity.
The economic implications are direct. A 200M gas limit, if adopted by validators post-fork, reduces per-transaction costs and increases L1 revenue capacity without requiring users to migrate to L2s for cost savings. For Layer 2 rollups, lower calldata costs further compress settlement expenses. For institutional stakers, ePBS replaces an opaque relay market with in-protocol rules governing builder behavior and payment.
The risk lies in execution. The Hoodi testnet and mainnet activation remain unconfirmed, and the interaction between ePBS and higher gas limits at scale has not been tested under adversarial conditions on a public network. The September 28 Sepolia fork is the first data point. The next 60-90 days will determine whether Glamsterdam ships in 2026 or slips into 2027.