Ethereum launched glamsterdam-devnet-0 on April 24, 2026 — the first unified test environment combining enshrined Proposer-Builder Separation (ePBS) and Block-Level Access Lists (BALs) into a single network. The devnet marks the beginning of a multi-month path toward the protocol's most consequen...
"Glamsterdam progress is slow but steady: enshrined Proposer-Builder Separation implementation is proving to be trickier than anticipated." — Ethereum Foundation, Checkpoint #9 (April 2026)
Ethereum launched glamsterdam-devnet-0 on April 24, 2026 — the first unified test environment combining enshrined Proposer-Builder Separation (ePBS) and Block-Level Access Lists (BALs) into a single network. The devnet marks the beginning of a multi-month path toward the protocol's most consequential upgrade since the Merge, targeting a gas limit increase from 60 million to 200 million, parallel transaction execution approaching 10,000 TPS, and projected gas cost reductions of up to 78%.
The stakes are structural. Currently, two entities — Titan Builder (52.16% of blocks) and Flashbots' BuilderNet (24.63%) — control over 76% of Ethereum's block production through off-chain relay infrastructure. ePBS would move this market on-chain, eliminating reliance on trusted intermediaries and reducing MEV extraction by an estimated 70%. For a network with 1.1 million active validators securing 36 million staked ETH ($83.9 billion at current prices), the governance implications of centralised block production are not theoretical.
The Ethereum Foundation's April 10 Checkpoint #9 blog post confirmed that Q2 2026 mainnet activation is "unlikely." Q3 remains the realistic target, contingent on devnet stability, client releases, security audits, and public testnet deployments on Holesky and Sepolia.
Glamsterdam does not arrive in a vacuum. The Fusaka (Fulu-Osaka) upgrade activated on December 3, 2025, and delivered an eightfold increase in blob capacity through PeerDAS (Peer Data Availability Sampling). The rollout was phased: maximum blob capacity rose from 9 per block to 15 on December 9, then to 21 on January 7, 2026.
The results were measurable. Typical Layer 2 transaction costs fell from approximately $0.50 in late 2025 to $0.20–$0.30 in the weeks following activation, according to Fidelity Digital Assets. PeerDAS eliminated the requirement that every node download and store all blob data, reducing bandwidth requirements while expanding data throughput. Post-activation telemetry showed healthy validator participation and smooth block propagation with no material disruptions.
However, Fusaka also exposed the blob fee revenue problem. Throughout most of 2024 and 2025, the blob base fee sat at or near 1 wei — effectively zero. Rollups enjoyed near-free data availability while Ethereum's base layer captured almost none of the value. EIP-7918, introduced alongside Fusaka, established a reserve floor price for blob gas to address this. The blob base fee subsequently surged by roughly 15 million times relative to its pre-Fusaka floor, establishing the first meaningful price signal in Ethereum's data availability market.
Core developers are already planning further blob parameter optimisations (BPOs) targeting 48 blobs per block by mid-2026, with a long-term target of 128 blobs per slot under full Danksharding.
Glamsterdam comprises two simultaneous upgrades across Ethereum's execution and consensus layers. The two headline features are:
Enshrined Proposer-Builder Separation (EIP-7732). Today, over 90% of Ethereum blocks are built using Flashbots' MEV-Boost relay system — an off-chain, trust-dependent mechanism. Under ePBS, the proposer and builder roles are formally separated at the protocol level. The proposer commits to a block header; a separate builder constructs the execution payload. The auction that currently takes place through external relays moves into the protocol itself.
The practical effect: validators no longer need to trust third-party relay operators. The protocol enforces the separation, removing a centralisation vector that has grown more acute as the builder market consolidated.
Block-Level Access Lists (EIP-7928). BALs introduce structured declarations specifying exactly which storage slots and accounts each transaction will read or write. A hash of this access list — the "BAL root" — is embedded directly in the block header. This enables the EVM to identify non-conflicting transactions and execute them in parallel, rather than sequentially.
Until the April 24 devnet launch, these two features were tested on separate networks (epbs-devnet and bals-devnet). The generalised devnet-0 is the first environment where both components coexist and interact — a necessary precondition for identifying integration bugs and performance characteristics at scale.
The economic case for ePBS rests on data. As of April 2026, Ethereum block production is dominated by a small number of entities:
| Builder | Block Share (24h) | |---------|------------------| | Titan Builder | 52.16% | | BuilderNet (Flashbots) | 24.63% | | Quasar | 15.06% | | Eureka | 2.41% | | Beaverbuild | 1.81% | | Others | 3.93% |
The Herfindahl-Hirschman Index (HHI) for block building stands at 3,186 — 2.5 times higher than the concentration level among miners prior to the Merge, according to an empirical study published in Big Data and Cognitive Computing (MDPI). The top three builders control 91.85% of blocks and 86.8% of all MEV.
This concentration emerged through exclusive order flow agreements. In April 2023, Titan secured an exclusive deal with Banana Gun, which became a top Telegram trading bot. Exclusive access to high-value transaction flow gave Titan a structural advantage in building more profitable blocks, crowding out smaller builders.
External PBS (via MEV-Boost) resulted in a 261% increase in proposer revenue and improved fairness across different-sized validators. But the tradeoff was centralising the builder market itself. Researchers estimate ePBS could reduce MEV extraction by up to 70%, redistributing value that currently accrues to a concentrated set of builders back to validators and, indirectly, to users.
Flashbots responded to the concentration problem by launching BuilderNet in November 2025, a multi-operator block building network using Trusted Execution Environments (TEEs). BuilderNet has captured 24.63% of blocks as of April 2026, partially decentralising what was previously a Titan-Beaverbuild duopoly. But BuilderNet remains an off-chain solution. ePBS would make the separation native to the protocol.
The throughput implications of Glamsterdam are substantial. The current Ethereum gas limit is 60 million per block. Glamsterdam's combination of ePBS and BALs enables a phased increase:
By increasing efficiency and the gas limit ceiling, Glamsterdam is projected to reduce Layer 1 gas costs by up to 78%, according to early devnet benchmarks reported by SpazioCrypto. The parallel execution capability enabled by BALs targets throughput of 10,000 TPS, compared to Ethereum's current double-digit TPS on Layer 1.
For context, Ethereum currently processes approximately 15–30 TPS on Layer 1. A move toward 10,000 TPS would represent a 300x–600x increase. However, this figure represents a theoretical ceiling under optimal parallelisation conditions. Real-world throughput will depend on the proportion of transactions with non-overlapping state access patterns.
The gas limit expansion also has implications for Ethereum's role relative to its Layer 2 ecosystem. Cheaper L1 execution could redirect some transaction volume from rollups back to mainnet, altering the economic relationship between L1 and L2s that has defined Ethereum's scaling strategy since EIP-4844.
Ethereum's economic model is in transition. The network has 1.1 million active validators, with 36 million ETH staked (approximately 30% of total supply). ETH trades at approximately $2,330 as of April 26, 2026, valuing staked ETH at $83.9 billion.
The blob fee floor introduced in Fusaka began generating measurable L1 revenue from L2 data posting. Industry estimates suggest blob fees could contribute 30–50% of total ETH burn by the end of 2026, depending on L2 activity scaling. This represents a structural shift from the prior regime where rollups contributed negligible fee revenue to the base layer.
Glamsterdam's gas cost reduction complicates this picture. If L1 transactions become 78% cheaper, more activity may occur on L1 rather than on rollups — generating direct execution fees but potentially reducing blob demand. The net revenue effect depends on the elasticity of demand: whether cheaper L1 gas generates enough additional volume to offset the per-transaction revenue decline.
The 25+ non-headliner EIPs under consideration for Glamsterdam include gas repricing proposals currently being debated in the Glamsterdam Repricings working group. These proposals would adjust the gas costs of specific opcodes to better reflect actual computational costs, potentially altering the economics for smart contract execution, storage operations, and cross-contract calls.
The Ethereum Foundation's Checkpoint #9 (April 10, 2026) provided the most authoritative timeline assessment:
The primary technical risk is ePBS complexity. Splitting block production into two parties acting in sequence inside consensus requires every part of the stack to handle "partial blocks" and two-party coordination — a paradigm shift from the current single-proposer model.
A secondary risk involves scope management. The Base engineering team has publicly warned that adding Fork-Choice Inclusion Lists (FOCIL, EIP-7805) alongside ePBS could delay the upgrade beyond 2026. The Ethereum Foundation resolved this by deferring FOCIL to the subsequent Hegotá upgrade, where it will serve as the sole headliner feature. This scope reduction improves Glamsterdam's timeline prospects but means censorship resistance improvements are delayed.
The Hegotá upgrade, with FOCIL as its headliner and Account Abstraction as a non-headliner feature, could follow Glamsterdam in a shorter timeframe than the gap between Fusaka and Glamsterdam, according to Checkpoint #9. Feature proposals for Hegotá opened on April 9, 2026.
Glamsterdam represents the largest structural change to Ethereum's block production since the transition to proof-of-stake. By enshrining proposer-builder separation in the protocol and enabling parallel execution through block-level access lists, the upgrade addresses two compounding problems: the centralisation of block building around a handful of entities, and the throughput ceiling that has pushed the majority of transaction volume to Layer 2 networks.
The devnet launch on April 24 establishes a concrete testing baseline. But the gap between devnet and mainnet — passing through client releases, security audits, and two public testnets — implies months of additional validation. The Ethereum Foundation's explicit acknowledgment that Q2 is unlikely and Q3 is "realistic" sets expectations accordingly.
The economic implications extend beyond gas costs. If Glamsterdam succeeds, it alters the value distribution between L1 validators, L2 operators, and the builder ecosystem that has captured an outsized share of MEV since the Merge. Whether that redistribution is net positive for ETH's value accrual depends on execution — both the technical kind and the market kind.