Ethereum is preparing for its most consequential upgrade since The Merge. Glamsterdam, a hard fork targeted for mid-2026, introduces a package of eight Ethereum Improvement Proposals (EIPs) that collectively restructure how blocks are built, how gas is priced, and how censorship resistance is enf...
"Block level access lists coming in Glamsterdam allow blocks to be verified in parallel. ePBS coming in Glamsterdam has many features..." — Vitalik Buterin, Ethereum Co-Founder, February 27, 2026
Ethereum is preparing for its most consequential upgrade since The Merge. Glamsterdam, a hard fork targeted for mid-2026, introduces a package of eight Ethereum Improvement Proposals (EIPs) that collectively restructure how blocks are built, how gas is priced, and how censorship resistance is enforced at the protocol level. The upgrade arrives at a critical juncture: ETH has fallen 36% year-to-date, trading below $2,000, while two block builders control over 90% of Ethereum's block production — a centralization crisis that threatens the network's foundational neutrality.
The headline features — enshrined Proposer-Builder Separation (ePBS via EIP-7732), Block-Level Access Lists (EIP-7928), and a progressive gas limit increase from 60 million toward 200 million — are not incremental improvements. They represent a structural overhaul of Ethereum's economic plumbing. For the $112 billion staked on the network, the $27 billion locked in DeFi on Aave alone, and the growing fleet of Layer 2 rollups settling billions daily on Ethereum's base layer, the stakes of getting Glamsterdam right are existential.
This report examines the economic architecture of each major proposal, analyzes the centralization dynamics Glamsterdam aims to solve, and assesses what the upgrade means for validators, L2 operators, DeFi protocols, and institutional participants building on Ethereum.
Ethereum's block production has quietly become one of the most centralized processes in all of crypto. Two sophisticated operators — Beaverbuild and Titan Builder — now construct over 90% of all Ethereum blocks. This duopoly exists because of Maximal Extractable Value (MEV): the profit available from reordering, inserting, or censoring transactions within a block. Annual MEV extraction on Ethereum exceeded $3 billion in 2025, with sandwich attacks alone accounting for $289.76 million — 51.56% of a measured $561.92 million MEV volume in one study period.
The current system relies on MEV-Boost, an off-chain relay infrastructure developed by Flashbots. Under this model, block builders assemble blocks optimized for MEV extraction and bid for the right to have their block proposed by validators. Proposers — the 1.1 million active validators staking 35.8 million ETH — act as passive auctioneers, accepting the highest-paying block without seeing its contents. The problem: this relay infrastructure is not enshrined in the protocol. It depends on trusted intermediaries, creates single points of failure, and has no protocol-level enforcement mechanism if a relay misbehaves.
Flashbots attempted a partial fix in late 2024 with BuilderNet, a decentralized block-building network running on Trusted Execution Environments (TEEs) and co-operated by Flashbots, Beaverbuild, and Nethermind. But BuilderNet remains an off-chain, voluntary coordination layer. Glamsterdam's answer is to bring the entire mechanism into the protocol itself.
EIP-7732, the centerpiece of Glamsterdam, introduces enshrined Proposer-Builder Separation (ePBS). This moves the block auction — currently handled by external relays — directly into Ethereum's consensus layer. The mechanics work through a commit-reveal flow:
This design eliminates trusted relays entirely. Builders cannot renege on commitments because the protocol enforces them. Proposers cannot frontrun or censor because they never see transaction contents until finalization. The protocol itself becomes the auctioneer.
The economic implications are significant. Currently, validators earn MEV payments through off-chain side channels with no protocol-level transparency. Under ePBS, builder payments become visible on-chain, creating a transparent and auditable market. This matters enormously for institutional stakers — entities like Coinbase (managing 1.84 million ETH, 5.1% of all validators) and Figment (1.48 million ETH, 4.1%) — who need regulatory clarity on the provenance of their yield.
Critically, ePBS also decouples block building centralization from staking centralization. Today, there is an economic incentive for stakers to delegate to pools affiliated with dominant builders, since those builders extract more MEV and can share higher tips. By commoditizing the auction mechanism, ePBS reduces this gravitational pull.
EIP-7928 introduces Block-Level Access Lists (BALs) — a mechanism requiring each transaction in a block to pre-declare which accounts and smart contracts it will interact with. This seemingly technical change has profound throughput implications.
Currently, Ethereum executes transactions sequentially. Each transaction must be processed in order because the EVM cannot know in advance whether two transactions touch the same state. BALs solve this by making state dependencies explicit. If two transactions touch entirely different accounts and contracts, nodes can execute them simultaneously.
The practical effect is that Ethereum can safely increase its gas limit — and therefore its throughput — without proportionally increasing the hardware burden on node operators. BALs are the engineering precondition for the gas limit increases that follow.
For Layer 2 rollups, BALs carry an additional benefit. By making state access patterns predictable at the block level, rollups can more efficiently plan their data availability submissions and batch execution. Base, Coinbase's L2, has publicly identified BALs as one of the most promising Glamsterdam proposals, noting they "support increased blob capacity and more efficient execution performance."
Ethereum's gas limit — the maximum computational work allowed per block — has sat at approximately 60 million since early 2025. Glamsterdam begins a phased escalation that could triple this capacity.
Tomasz Stańczak, co-director of the Ethereum Foundation, outlined the trajectory: the gas limit rises to 100 million in H1 2026, with a target of 200 million following ePBS deployment. Stańczak indicated that 300 million gas per block is theoretically achievable before year-end, contingent on parallel execution performance.
At 200 million gas, Ethereum's L1 could process up to 10,000 transactions per second — a dramatic increase from the current effective throughput of roughly 30-60 TPS. This does not replace L2 scaling. Instead, it creates headroom for L1 to handle settlement, complex DeFi operations, and cross-rollup coordination while L2s continue to serve high-volume consumer applications.
The gas repricing is handled through EIP-7904, which benchmarks gas costs against actual computational overhead. Operations currently overpriced relative to modern hardware (such as certain cryptographic precompiles) get cheaper, while underpriced operations get more expensive. The result is a more accurate relationship between gas consumed and real-world node resources used.
Glamsterdam introduces ePBS, but Ethereum's censorship resistance plan extends beyond it. Vitalik Buterin's March 2, 2026 roadmap post laid out a multi-stage progression:
Stage 1 — Glamsterdam (H1 2026): ePBS creates a permissionless builder market. Proposers can outsource block construction without trusting intermediaries. However, dominant builders can still censor transactions within their blocks.
Stage 2 — FOCIL in Hegota (H2 2026): Fork-Choice Enforced Inclusion Lists (EIP-7805) add a committee of 16 randomly selected validators who create mandatory inclusion lists. If a block builder ignores valid transactions from these lists, the block is rejected by the fork-choice rule. This ensures any valid transaction reaches finality within a bounded number of slots — even if every builder attempts to censor it.
Stage 3 — "Big FOCIL" and Encrypted Mempools (future): Buterin envisions expanding inclusion lists to cover nearly all transactions, reducing the builder's role to state computation and minor MEV optimization. Encrypted mempools would prevent builders from even seeing transaction contents until after ordering, eliminating sandwich attacks and frontrunning at the protocol level.
FOCIL has been controversial. Critics argue it could expose validators to legal liability by forcing inclusion of sanctioned transactions. Proponents counter that it is precisely this property that makes Ethereum credibly neutral — the protocol does not discriminate, and no single actor can be compelled to censor.
ePBS transparently routes builder payments through the protocol, improving yield attribution. The 3.3% average staking APY on 35.8 million staked ETH could see composition shifts as MEV becomes more evenly distributed. Large staking operations gain regulatory clarity; smaller operators gain fairer access to MEV proceeds.
Higher gas limits and BALs directly benefit rollup economics. More blobspace at lower cost reduces L2 operating expenses — already down to approximately $135,100 daily post-Dencun. Base, Arbitrum, and Optimism can scale user throughput without increasing settlement costs. EIP-7918's blob fee floor ensures Ethereum retains meaningful pricing power in the data availability market, preventing a race to zero.
Parallel execution enables more complex atomic operations per block. Protocols like Aave — which crossed $1 trillion in cumulative lending volume in February 2026 on $27.2 billion TVL — can support denser liquidation cascades, multi-collateral operations, and real-time oracle updates without competing for scarce sequential block space. The MEV reduction from ePBS and the eventual encrypted mempool protects users from sandwich attacks, reducing the hidden tax on swaps and liquidations.
The transparency of on-chain builder auctions, combined with quantum-resistant signature migration paths (EIP-8141) and state tree efficiency gains (EIP-7864), positions Ethereum as infrastructure that institutional compliance teams can underwrite. This matters as firms like VanEck, WisdomTree, and Securitize expand their on-chain exposure through Aave Horizon and similar institutional DeFi products.
Block building centralization is Ethereum's most urgent structural risk. Two builders control 90%+ of block production. Glamsterdam's ePBS replaces trusted relay infrastructure with protocol-enforced auctions — the single most important change since The Merge.
The gas limit is tripling. From 60 million to 100 million in H1 2026, targeting 200 million post-ePBS. Block-Level Access Lists enable this by unlocking parallel transaction execution without compromising decentralization.
Censorship resistance becomes protocol-enforced. FOCIL (coming in Hegota, H2 2026) introduces mandatory inclusion lists. Combined with ePBS, this creates a two-phase system where no actor — builder or proposer — can systematically exclude valid transactions.
L2 economics improve structurally. Higher gas limits, cheaper blobs, and predictable state access reduce rollup settlement costs while EIP-7918 preserves Ethereum's data availability pricing power.
ETH's price disconnect is a feature, not a bug. At $1,940 and falling, ETH trades at a steep discount to the economic activity it settles. Glamsterdam does not solve the narrative problem, but it materially improves the network's economic plumbing — the value accrual mechanism itself.
Glamsterdam is not a speculative roadmap feature. It is an engineering response to measurable centralization. Two block builders controlling 90% of production is not a hypothetical risk — it is today's reality. The $3 billion annual MEV extraction flowing through opaque off-chain relays is not a future concern — it is the current architecture.
What makes Glamsterdam consequential is its scope. ePBS restructures the economic relationship between proposers and builders. BALs unlock parallel execution for the first time. Gas limit increases deliver 3x capacity. And the FOCIL roadmap following in Hegota creates the first protocol-enforced censorship resistance mechanism in a major blockchain.
For the $112 billion staked on Ethereum, the billions flowing through DeFi, and the institutional capital cautiously positioning on-chain, these are not abstract improvements. They are structural changes to how value flows through the network — and who captures it. Glamsterdam does not make ETH cheap or expensive. It makes Ethereum harder to capture, harder to censor, and harder to ignore.