Ethereum's Glamsterdam hard fork, targeted for H1 2026, is the network's third protocol upgrade in twelve months. The fork enshrines Proposer-Builder Separation (ePBS) via EIP-7732 and introduces Block-Level Access Lists (BALs) via EIP-7928. Together, these changes raise the gas limit from 60 mil...
"The resulting gas limit increase will be phased, reaching 100 million per block initially and 200 million once ePBS is fully operational." — Tomasz Stańczak, Co-Executive Director, Ethereum Foundation
Ethereum's Glamsterdam hard fork, targeted for H1 2026, is the network's third protocol upgrade in twelve months. The fork enshrines Proposer-Builder Separation (ePBS) via EIP-7732 and introduces Block-Level Access Lists (BALs) via EIP-7928. Together, these changes raise the gas limit from 60 million to an initial 100 million per block — potentially 200 million post-ePBS — targeting 10,000 transactions per second on the base layer. A companion suite of gas repricing EIPs projects a 78% reduction in fees for simple transfers and smart contract calls.
The upgrade arrives at a point of tension. ETH trades at approximately $2,194 as of April 9, 2026, down roughly 39% from its October 2025 high above $3,600. Ethereum's annual net inflation sits at 0.23%, with daily validator issuance of ~1,700 ETH outpacing a burn rate that collapsed to 50–70 ETH/day in early 2025. L2 fee payments to Ethereum L1 fell over 90% year-over-year, from $113 million in 2024 to $10 million in 2025. Glamsterdam is less about competitive positioning against rival L1s and more about whether Ethereum's base layer can recapture enough economic activity to sustain its security model.
Glamsterdam bundles two headliner EIPs and a package of supporting gas-repricing proposals. It is the consensus and execution layer upgrade following Pectra (May 2025) and Fusaka (December 2025). Vitalik Buterin outlined eight EIPs defining the full scope in a February 2026 blog post, framing 2026 as the year Ethereum should "take back lost ground in terms of self-sovereignty and trustlessness."
The two primary changes:
| EIP | Layer | Function | |-----|-------|----------| | EIP-7732 (ePBS) | Consensus | Enshrines block builder market into protocol | | EIP-7928 (BALs) | Execution | Enables parallel transaction processing |
Over 25 additional non-headliner EIPs remain under consideration, though the inclusion of FOCIL (Fork-Choice Enforced Inclusion Lists, EIP-7805) has been deferred to the subsequent Hegota fork to avoid timeline slip.
Currently, approximately 90% of Ethereum blocks are built through MEV-Boost, an out-of-protocol relay system developed by Flashbots. Two builders control the majority of block production — a concentration level that has tripled since 2022, according to Flashbots data.
EIP-7732 moves the proposer-builder market directly into the consensus layer. Under the new mechanism:
The practical effect is structural. Validators no longer depend on external relay infrastructure to participate in block production. The nine stakeholders who ranked ePBS as their first choice during the All Core Developers Consensus Call #162 (ACDC #162) cited relay dependency and censorship risk as primary concerns.
However, ePBS introduces a known trade-off. As Buterin acknowledged: "ePBS only prevents builder centralization from spilling over into the staking layer, but the issue of builder centralization itself remains." Research led by Shuzheng Wang indicates that builders who secure more private order flow win more auctions and retain more profit, creating feedback loops that favor scale. Simulations suggest ePBS could increase profit concentration among a small number of builders.
The mitigation roadmap extends beyond Glamsterdam. Buterin outlined a multi-stage approach: ePBS first, then FOCIL in Hegota, followed by encrypted mempools in subsequent forks. Each layer is designed to progressively constrain builder power without compromising throughput.
Ethereum currently processes transactions sequentially within each block. EIP-7928 introduces Block-Level Access Lists that make storage slot information explicit at the block level, identifying which accounts and smart contracts each transaction will interact with.
This enables parallel processing. Transactions that do not touch the same state can execute simultaneously across multiple CPU cores — effectively converting Ethereum's single-lane execution model into a multi-lane system.
The performance implications are significant when combined with the gas limit increase:
| Metric | Pre-Glamsterdam | Post-Glamsterdam (Projected) | |--------|----------------|------------------------------| | Gas limit per block | 60M | 100M (initial), 200M (post-ePBS) | | Effective TPS | ~15–30 | Up to 10,000 (target) | | Transaction execution | Sequential | Parallel (non-conflicting txs) |
The 10,000 TPS target is aspirational and assumes full ePBS deployment with the 200M gas limit. Initial gains at the 100M level will be more modest.
Glamsterdam bundles EIP-7904 and companion proposals that recalibrate gas costs based on actual computational resource benchmarks rather than legacy pricing assumptions.
The projected result: a 78.6% reduction in gas fees across both simple ETH transfers and complex smart contract interactions. This repricing is independent of the gas limit increase — fees fall because operations are priced more efficiently, not just because block space expands.
For context, the Ethereum base fee on L1 has been well below the 16 gwei threshold needed for net deflation through most of late 2025 and early 2026. A 78% fee reduction, if realized, further compresses per-transaction burn — amplifying the supply dynamics challenge discussed below.
MEV (Maximal Extractable Value) represents the profit validators and builders extract by reordering, inserting, or censoring transactions within a block. According to Flashbots data, MEV-Boost handles ~90% of Ethereum block production, with market concentration among builders increasing steadily.
ePBS is projected to reduce MEV extraction by up to 70%, according to Bitfinex research. But the mechanism shifts rather than eliminates centralization pressure.
Research findings paint a nuanced picture:
The economic tension is clear: ePBS protects validators from centralization contagion but may entrench a builder oligopoly. Whether FOCIL and encrypted mempools — scheduled for later forks — arrive fast enough to prevent this entrenchment is an open question.
Glamsterdam's L1 scaling ambitions collide with a structural problem: Ethereum's Layer 2 ecosystem has grown to $38–44 billion in total value secured (per L2BEAT, March 2026), but the fee value flowing back to L1 has collapsed.
L2 Fee Capture Data (Ethereum L1 Revenue from L2s):
| Year | L2 Revenue | Paid to L1 | L1 Capture Rate | |------|-----------|------------|-----------------| | 2024 | ~$277M | ~$113M | 41% | | 2025 | ~$129M | ~$10M | 8% |
The 90%+ drop in L1 payments coincided with Fusaka's blob expansion (EIP-7918), which established a minimum blob fee floor but hasn't yet reversed the trend. Blob utilization sits at approximately 29% of the 14-blob target — a massive surplus that suppresses fee revenue.
Three L2s dominate: Base ($4.2B DeFi TVL, 46.6% L2 DeFi market share), Arbitrum ($2.8B, 30.9%), and OP Mainnet ($550M, 6%). Only Base generates a profit, with ~$55M net from $93M in sequencer revenue in 2025. The remaining 50+ L2s collectively process approximately 10% of all L2 transactions.
Glamsterdam's L1 scaling may recapture some activity currently pushed to L2s by high base-layer fees. But the 78% fee reduction simultaneously compresses per-transaction burn, potentially accelerating Ethereum's shift from deflationary to inflationary dynamics. ETH supply has grown by approximately 950,000 tokens since The Merge (September 2022), with an annualized inflation rate of ~0.23%.
As of April 2026:
The June timeline carries meaningful risk. Over 25 non-headliner EIPs remain under consideration, and the decision to exclude FOCIL from Glamsterdam was made specifically to avoid delays. Ethereum's two previous forks (Pectra and Fusaka) each shipped within weeks of target, but neither included consensus-layer changes as structurally significant as ePBS.
The Hegota fork (tentatively H2 2026) is planned to introduce Verkle Trees, reducing node storage requirements by approximately 90%, alongside FOCIL. If Glamsterdam slips into Q3, Hegota may push into 2027.
Glamsterdam's 10,000 TPS target, if achieved, would narrow the throughput gap with competing L1s. For comparison:
| Network | Real-World TPS | Theoretical Max | Avg. Transaction Cost | |---------|---------------|-----------------|----------------------| | Ethereum (current) | 15–30 | ~30 | $1–50+ | | Ethereum (post-Glamsterdam) | Up to 10,000 | 10,000 | ~78% lower | | Solana | 600–3,300 | 65,000 | $0.00025 | | Ethereum L2s (combined) | 40,000+ | 100,000+ | $0.10–1.00 |
Solana's Firedancer validator client adoption is expected to push its throughput toward 10,000+ TPS by mid-2026. A Glamsterdam-upgraded Ethereum at 10,000 TPS would reach theoretical parity with Solana's near-term target — though at substantially higher per-transaction costs.
The comparison, however, obscures different design philosophies. Ethereum maintains over 1 million validators with $70 billion+ in staked collateral (37 million ETH, 30.6% of circulating supply). Solana operates approximately 1,500 validators. The security model trade-offs are reflected in staking yields: Ethereum offers 3.5–4.5% APY versus Solana's variable rates.
Glamsterdam represents Ethereum's most significant structural change since The Merge. EIP-7732 and EIP-7928 address real infrastructure deficiencies: relay dependency, sequential processing, and gas mispricing. The technical case for the upgrade is sound.
The economic case is more complex. Ethereum faces a revenue problem: L2s have absorbed transaction volume while returning diminishing fees to the base layer. A 78% fee reduction improves user experience but further pressures the burn mechanism that underpins ETH's monetary narrative. The network's shift from deflationary to mildly inflationary is unlikely to reverse absent a substantial increase in L1 transaction volume.
The MEV question remains half-answered. ePBS isolates validators from builder centralization but leaves the builder market itself concentrated. The full mitigation stack — FOCIL, encrypted mempools — is at least one to two forks away. Whether builder oligopoly calcifies in the interim period is a structural risk that protocol researchers have acknowledged but not yet resolved.
Devnet-5 testing is underway. Public testnet activation on Holesky and Sepolia is expected in the coming weeks. If the June target holds, Ethereum will have shipped three hard forks in thirteen months — a pace not seen since the pre-Merge era. The technical execution track record supports cautious confidence in delivery. The economic sustainability question Glamsterdam raises, however, extends well beyond the fork itself.