Ethereum Glamsterdam 2026: Fees Down 78% — What Changes for Users

Ethereum has spent the last two years proving it can scale its Layer 2 ecosystem. The Dencun upgrade in March 2024 slashed L2 transaction fees by roughly 90% through EIP-4844 blob transactions. Pectra, deployed in May 2025, refined validator operations, introduced EIP-3074 authcall for account abstraction, and increased the blob count. Both upgrades were consequential. Neither touched the execution layer's core bottleneck: the sequential processing model that limits how many transactions Ethereum can actually execute on Layer 1 per second.

Glamsterdam is different. It targets execution directly. The two headline EIPs — 7732 and 7928 — address, respectively, how blocks are built and how transactions within those blocks are processed. Together with a proposed gas limit increase that validators are discussing as a matter of social consensus rather than protocol rule, they represent the most significant overhaul of Ethereum's base execution layer since the Merge in 2022.

This article breaks down each component, explains the cryptographic and economic mechanics for a non-developer audience, and assesses the real-world implications — including the legitimate criticisms that researchers and node operators are raising about the pace and scale of these changes.

What Is the Glamsterdam Upgrade?

The name Glamsterdam is unofficial — a community codename coined by Ethereum developers and researchers during planning sessions, blending Glasgow and Amsterdam, two cities historically associated with Ethereum developer summits and AllCoreDevs meetings. Ethereum upgrades have a tradition of geographic names: Berlin, London, Shanghai, Cancun, Prague. Glamsterdam follows that convention while acknowledging that its scope spans multiple thematic clusters that were discussed across both cities.

In technical terms, Glamsterdam is a coordinated hard fork — a backward-incompatible protocol change that all Ethereum nodes must adopt simultaneously at a specific block number (or epoch number, in post-Merge terminology). It is not a soft fork and it is not optional. Once the network reaches the designated transition point, any node running older software is cut off from the canonical chain.

The upgrade sits third in the post-Dencun sequence. To understand why Glamsterdam is necessary, it helps to trace the trajectory: Dencun (March 2024) introduced EIP-4844 proto-danksharding, which created a separate data lane (blobs) for Layer 2 rollups, reducing their data posting costs by up to 90% and triggering a wave of L2 activity. Pectra (May 2025) followed with housekeeping: validator credential improvements, EIP-3074 authcall, increased blob targets for high-throughput L2s, and several EVM Object Format (EOF) changes that cleaned up legacy opcodes. Neither upgrade increased the raw capacity of Ethereum's execution layer. Glamsterdam does.

The three pillars of Glamsterdam are: EIP-7732 (Enshrined Proposer-Builder Separation), EIP-7928 (Block-Level Access Lists), and a gas limit increase from approximately 60 million to 200 million gas units per block — the last of which is a validator governance decision rather than a protocol specification change, making it unusual among Ethereum scaling milestones.

EIP-7732 — Enshrined Proposer-Builder Separation Explained

To understand EIP-7732, you need to understand the problem it solves: the current architecture of Maximal Extractable Value (MEV) on Ethereum.

Every block on Ethereum is proposed by a validator — a participant who has staked 32 ETH to participate in consensus. The validator has the right to include whatever transactions they choose in their block, subject to protocol rules. In theory, this is a neutral process. In practice, the ordering of transactions within a block creates enormous economic opportunity: a sophisticated actor who can choose the sequence of transactions can front-run trades, sandwich decentralized exchange orders, and extract value from ordinary users through price manipulation. This is MEV.

The current solution is MEV-Boost, an off-chain relay system developed by Flashbots. In MEV-Boost, validators outsource block construction to specialized builders — firms with sophisticated software that assembles the most profitable transaction ordering possible. Builders compete by offering validators a share of the MEV they extract. Validators, in turn, are economically incentivized to accept builder blocks rather than constructing their own. As of 2025, over 85% of Ethereum blocks were built by external builders through MEV-Boost relays.

This architecture works, but it creates serious structural problems. First, it is entirely off-chain: the relay system operates outside the Ethereum protocol, with no on-chain enforcement, no slashing conditions for misbehavior, and no transparency guarantees. Second, it is deeply centralized: a handful of builders construct the vast majority of blocks, meaning a small number of firms have outsized control over transaction censorship. Third, it creates systemic fragility: if the dominant relays go offline or are compromised, the MEV market breaks in ways the protocol cannot recover from gracefully.

EIP-7732 — Enshrined Proposer-Builder Separation (ePBS) — solves this by bringing the proposer-builder relationship into the core protocol. Under ePBS, the split between the validator (proposer) and the block-builder is formalized in the consensus layer. Builders submit sealed bids for the right to construct the execution payload of each block. The winning builder's payload is committed to in the consensus layer and revealed only after the proposer has committed to it — a cryptographic mechanism that prevents last-minute switches while ensuring builders cannot withhold blocks.

The practical implications for users are significant. Because the builder selection process is now on-chain and governed by protocol rules, censorship resistance improves dramatically: validators cannot quietly exclude transactions on behalf of builders without it being visible and attributable on-chain. The reliance on off-chain relays — a chokepoint for regulatory action and a systemic risk — is eliminated. And because builders must post on-chain commitments, their behavior is subject to Ethereum's slashing and accountability mechanisms for the first time.

What ePBS does not do is reduce fees directly. It is a structural reform of how block production works, not an increase in block capacity. Its contribution to Glamsterdam is to make the system more trustworthy and decentralized, creating the institutional foundation on which higher-throughput changes (BALs, higher gas limits) can safely rest. A more decentralized, censorship-resistant block production system is better positioned to handle the increased throughput that the rest of Glamsterdam introduces.

EIP-7928 — Block-Level Access Lists and Parallel Execution

If EIP-7732 is about who builds blocks, EIP-7928 is about how the transactions inside those blocks are processed. And here is where the throughput numbers become dramatic.

Today, the Ethereum Virtual Machine (EVM) executes transactions sequentially — one after another, in the order they appear in the block. This is a deliberate design choice, not an accident. Sequential execution is simple to reason about: transaction A runs completely before transaction B begins, so there is no risk of two transactions simultaneously modifying the same piece of state in inconsistent ways. But sequential execution is also a severe bottleneck: even if a modern computer could process ten transactions simultaneously, the EVM processes only one.

Parallel execution is the obvious solution. The reason it has not been implemented sooner is the state conflict problem: to safely execute transactions in parallel, you need to know in advance whether any two transactions will touch the same piece of Ethereum state (the same contract storage slot, the same account balance). If transaction A and transaction B both try to write to the same storage slot, you cannot run them simultaneously without creating undefined behavior. Identifying these conflicts in real time, during execution, is computationally expensive and effectively negates the benefit of parallelism.

EIP-7928 solves this through Block-Level Access Lists (BALs). Under the proposal, transactions must pre-declare — at the block level, before execution begins — which state locations they intend to read and write. This pre-declaration is not enforced by the protocol in the same way that gas limits are enforced (a transaction that accesses state it did not declare simply fails), but it provides the EVM with the information it needs to build a conflict graph before execution starts.

With a complete pre-declared conflict graph, the EVM scheduler can identify which transactions have no overlapping state access — and execute them simultaneously. Transactions that do conflict are serialized, as before. In practice, on a block with representative mainnet traffic (a mix of simple ETH transfers, ERC-20 approvals, DEX swaps, and NFT operations), the majority of transactions operate on entirely disjoint state and can be parallelized. Researchers modeling EIP-7928's impact on representative block samples estimate that between 60% and 80% of transactions could be parallelized in typical blocks.

The throughput implication is substantial. Combined with the gas limit increase discussed in the next section, community and research models estimate that EIP-7928 contributes the majority of the ~78% fee reduction and pushes Ethereum's theoretical transaction processing capacity toward 10,000 transactions per second on Layer 1 — a 300x improvement over the current practical limit of 15–30 TPS.

That 10,000 TPS figure is theoretical, and the distinction matters. It assumes ideal transaction composition, maximum parallelizability, and validators running hardware capable of handling the increased computation. Real-world throughput will be lower. But even at 20–30% of theoretical maximum, the improvement in Layer 1 capacity would be transformational for use cases that currently rely entirely on Layer 2 solutions — including DeFi protocols, gaming applications, and real-world asset tokenization platforms where L1 settlement costs dominate economics.

Gas Limit Increase: From 60M to 200M

The third leg of Glamsterdam is the most unusual: a proposed increase in the block gas limit from approximately 60 million gas units to 200 million — a more than 3x expansion — achieved not through an EIP but through validator social consensus.

The gas limit on Ethereum is not a fixed protocol constant. It is a parameter that each validator can adjust incrementally within a small range (1/1024 of the current limit per block, either up or down). The current level of approximately 36–60 million (a range reflecting the gradual increase validators have been implementing throughout 2025) is the result of validators collectively voting to raise it from the post-Merge baseline. When a sufficient majority of validators signal the same direction, the limit drifts accordingly.

The discussion around raising the gas limit to 200M has been ongoing in Ethereum research circles since at least late 2024. The core argument is simple: block capacity is currently artificial. The hardware requirements to run a full Ethereum node have not grown proportionally with the gas limit adjustments made so far, and modern client software (especially Go-Ethereum and Nethermind with their latest optimizations) can handle significantly more computation per block than validators currently allow. A 200M gas limit would increase the amount of computation — and therefore the number of transactions — that can fit in each 12-second block slot.

Combined with EIP-7928's parallel execution, the gas limit increase acts as a force multiplier. More gas per block means more transactions per block; parallel execution means those transactions complete faster and at lower per-unit cost. The ~78% fee reduction estimate that appears in community discussions and research blog posts derives primarily from this combination: the expanded block space (more supply) meeting parallel execution efficiency gains (lower per-transaction overhead).

However, this is also where the most substantive criticism of Glamsterdam's scope is concentrated, and that criticism deserves careful attention.

State bloat: More transactions per block means more state changes per block. Ethereum's state — the complete record of all account balances, contract code, and storage — grows with every transaction. At a gas limit of 200M with full blocks, the rate of state growth could increase by a factor commensurate with the gas limit increase. This creates long-term storage challenges for node operators, particularly home node runners and smaller staking operations. Ethereum researchers including Vitalik Buterin have written extensively about state growth as an existential challenge for decentralization; Glamsterdam's gas limit increase accelerates the timeline on which statelessness or state expiry solutions become urgent.

Bandwidth requirements: Larger blocks require more bandwidth to propagate across the network. Ethereum's 12-second slot time is a constraint: if a block takes more than a few seconds to reach most validators, it risks being orphaned. A tripling of the gas limit — and corresponding block size — puts pressure on validators running on commodity internet connections, disproportionately affecting home stakers in regions with slower upload speeds. This is a decentralization concern: if only well-resourced data center operators can keep up with block propagation at 200M gas, the validator set risks further concentration.

Client diversity risk: EIP-7928's parallel execution is computationally complex. Ethereum's goal of client diversity — having multiple independent software implementations of the protocol, so that a bug in one client does not bring down the whole network — means all major clients (Geth, Nethermind, Besu, Erigon, Reth) must implement parallel execution correctly and consistently. Subtle differences in how clients handle access list declarations, conflict resolution, or edge cases could lead to consensus splits. The testing burden for Glamsterdam's execution changes is therefore substantially higher than for previous upgrades.

These are not hypothetical concerns raised by critics of Ethereum's direction; they are active items in the Ethereum Magicians forum and AllCoreDevs call notes. The Ethereum community's willingness to discuss them openly is part of why the upgrade timeline remains Q3 2026 rather than Q1 — devnets and testnets are expected to run for an extended period before mainnet deployment is confirmed.

How Much Could Fees Drop? The 78% Estimate

The ~78% average gas fee reduction figure is the most-cited statistic in Glamsterdam coverage, and it requires careful unpacking. Where does it come from, what does it actually mean, and what assumptions does it rest on?

The estimate originates from a combination of Ethereum research blog posts, EIP author analyses, and independent modeling by DeFi analytics teams. It is not a protocol guarantee, a roadmap commitment, or a number that the Ethereum Foundation has officially endorsed. It is a research estimate, and like all estimates, it embeds significant assumptions.

The primary mechanism driving the fee reduction is supply-demand dynamics. Ethereum gas fees are set by EIP-1559's base fee mechanism, which adjusts the minimum fee per gas unit based on how full recent blocks have been. When blocks are consistently full (demand exceeds supply), the base fee rises. When blocks have spare capacity (supply exceeds demand), the base fee falls. A tripling of the gas limit — all else being equal — triples the supply of block space. Basic economics: more supply at the same demand level means lower prices.

The parallel execution layer adds a separate efficiency gain. By reducing the computational overhead per transaction (through simultaneous rather than sequential processing), EIP-7928 allows more transactions to be processed within the same gas budget, which effectively lowers the gas cost of complex operations and reduces the competition for block space at peak times.

The 78% figure assumes that demand for Ethereum block space does not proportionally expand to fill the new capacity in the short to medium term — a reasonable assumption given that most transaction demand growth on Ethereum since Dencun has been captured by L2s rather than L1. If L1 fees drop sharply, some activity that currently executes on L2s may migrate back to L1, partially refilling the new capacity. This demand response would dampen the fee reduction below 78% in equilibrium.

To put the number in concrete terms: at the time of writing, a standard ERC-20 token transfer on Ethereum mainnet during moderate network congestion costs approximately $0.80–$2.50 depending on the time of day and current demand. A 78% reduction would bring that range to approximately $0.18–$0.55. A complex DeFi interaction (multi-hop DEX swap, for example) that currently costs $8–$25 on L1 could fall to $1.76–$5.50. These are illustrative figures based on the estimate, not guarantees, and actual fees will vary with market conditions.

For context, it is useful to recall how Dencun's EIP-4844 affected L2 fees: before the upgrade, posting a batch of transactions to Ethereum as an L2 rollup cost approximately $0.50–$2.00 per user transaction in data fees alone. After EIP-4844, that cost fell to fractions of a cent. L2 adoption subsequently accelerated significantly, with daily transaction counts on networks like Base, Arbitrum, and Optimism increasing several-fold within months of the upgrade. Glamsterdam's impact on L1 fees could trigger a comparable adoption wave for use cases that need L1 settlement guarantees.

Timeline: When Does Glamsterdam Launch?

The Q3 2026 target — meaning July through September 2026 — is the current planning horizon for Glamsterdam, but it is worth being precise about what "target" means in the context of Ethereum's upgrade process.

Ethereum upgrades do not have a fixed launch date announced months in advance. They proceed through a staged process: EIP finalization (specification locked, no further changes), devnet deployment (small private networks run by core developers to catch implementation bugs), testnet deployment (public testnets including Sepolia and Holesky, which anyone can run), and finally mainnet activation at a specific block number agreed by client teams and the community.

As of May 2026, EIP-7732 has been marked "candidate for inclusion" in Glamsterdam by the AllCoreDevs process, meaning core developers have agreed it meets the bar for deployment. EIP-7928 is in late-stage review, with the specification largely finalized but implementation testing still underway across multiple clients. The gas limit increase, being a validator governance matter rather than a protocol specification, can proceed independently of the EIP schedule — validators can begin signaling for 200M before, during, or after the Glamsterdam fork itself.

The historical cadence is informative: from EIP finalization to mainnet activation, major Ethereum upgrades have typically required 4–8 months. Dencun took approximately six months from EIP lock-in to mainnet. Pectra took approximately seven months. Given that EIP-7928 is more technically complex than either of those upgrades' execution-layer changes, a seven-to-nine-month timeline from now to mainnet activation is plausible, landing in late Q3 or early Q4 2026 as a realistic range.

The Ethereum Foundation and client teams have explicitly stated that they will not rush Glamsterdam's testnet phase. Given the complexity of parallel execution and the potential for consensus bugs, extended testnet periods are expected. If significant issues emerge on testnets — which is not unusual for major upgrades — the Q3 2026 target could slip into Q4 2026 without causing alarm. The community's priority is correctness over schedule.

For investors and users tracking the upgrade, the most reliable signal will be the AllCoreDevs call announcements and the official Ethereum blog, which publishes upgrade readiness updates. A mainnet activation block is typically announced two to four weeks before the target date once testnet results are satisfactory.

What This Means for Ethereum Users and Investors

For ordinary Ethereum users — people who hold ETH, interact with DeFi protocols, trade on decentralized exchanges, or use Ethereum-based applications — Glamsterdam's most direct effect is lower transaction costs. If the ~78% fee reduction materializes even partially, the economics of interacting directly with Ethereum Layer 1 change substantially. Smaller transactions that are currently uneconomical on L1 (micro-DeFi positions, small NFT purchases, low-value token transfers) become viable again without routing through an L2 bridge.

For developers building Ethereum applications, the BAL pre-declaration requirement in EIP-7928 introduces a new consideration: smart contracts will need to be written or updated to correctly specify their state access patterns in block-level access lists. This is an additional engineering requirement, and poorly specified access lists (transactions that access state they did not declare) will fail. The upgrade therefore has a meaningful adoption period during which the developer ecosystem must adapt tooling, libraries, and contract patterns to the new model.

For DeFi protocols, the combination of higher throughput and lower fees could significantly expand the addressable market. Automated market makers (AMMs) like Uniswap, liquidity protocols like Aave, and perpetual futures platforms are all constrained today by L1 execution costs. Glamsterdam does not eliminate L2s — their speed advantages remain — but it narrows the cost gap and makes L1 DeFi competitive again for transactions that require the maximum security guarantees that only L1 settlement provides.

For ETH as an asset, the analytical picture is nuanced. On one hand, more Ethereum activity (more transactions, more DeFi usage, more application deployment) typically means more ETH burned via EIP-1559's fee destruction mechanism — a deflationary pressure that has historically supported ETH's price. On the other hand, lower per-transaction fees mean less ETH burned per transaction, which could partially offset the volumetric increase. The net effect on ETH supply dynamics depends on whether the demand response (more transactions from lower fees) outpaces the per-unit reduction in fee burn. Research models on this question produce divergent results, and investors should approach any specific ETH price prediction tied to Glamsterdam with appropriate skepticism.

For those looking to position ahead of Glamsterdam or to increase their Ethereum exposure, accessing ETH through regulated, licensed platforms is important — particularly for European investors operating under MiCA's framework. Coinbase holds a MiCA CASP license through its Luxembourg entity and offers straightforward ETH spot trading with competitive fees. Binance's European operations are currently progressing through the MiCA authorization process; check the latest licensing status before opening a new account.

It is also worth placing Glamsterdam in the broader context of Ethereum's competition with other smart contract platforms. Solana, for example, has marketed its parallel execution architecture (Sealevel) as a fundamental advantage over Ethereum's sequential EVM. EIP-7928 is in part Ethereum's answer to that criticism — an acknowledgment that parallel execution is necessary for L1 scalability and an implementation of it within the EVM's existing architecture. Whether Ethereum's approach (an opt-in, pre-declared access list model) or Solana's approach (mandatory account declaration for all transactions) proves more practical at scale is a live technical debate. Glamsterdam will provide real-world data.

Investors comparing ETH to Bitcoin should note the fundamental difference in upgrade philosophy: Bitcoin's protocol is deliberately minimal and conservative, with any change requiring near-unanimous community consensus and years of discussion. Ethereum's upgrade cadence is more aggressive, reflecting a different set of design priorities: programmability, scalability, and application support are considered core properties, not features to be added cautiously. Glamsterdam illustrates this difference — it is a complex, multi-EIP upgrade targeting execution layer performance, something that would be inconceivable in Bitcoin's protocol development culture.

Neither approach is objectively superior; they reflect different visions of what a blockchain should be. Bitcoin's conservatism provides stability and predictability; Ethereum's iterative development provides adaptability and performance improvements. Investors holding both assets are not making a mutually exclusive bet — they are expressing exposure to two different value propositions within the same asset class.

FAQ — Ethereum Glamsterdam Upgrade 2026

What is the Ethereum Glamsterdam upgrade?

Glamsterdam is the community codename for Ethereum's next major protocol upgrade, expected in Q3 2026. The name blends "Glasgow" and "Amsterdam" — two cities associated with prior Ethereum developer meetings. It follows the Pectra upgrade (May 2025) and focuses on execution-layer scalability through EIP-7732 (ePBS), EIP-7928 (Block-Level Access Lists), and a proposed validator-driven gas limit increase from ~60M to 200M gas units per block.

How much will Ethereum fees drop in 2026 after Glamsterdam?

Community and research estimates suggest average gas fees could fall by approximately 78% following Glamsterdam. This figure combines the effect of parallel transaction execution enabled by EIP-7928 (Block-Level Access Lists) with the increased block capacity from a gas limit raised to 200M. However, actual fee reduction depends on how demand grows to fill the new capacity — if usage scales proportionally, fees could partially recover over time. A standard ERC-20 transfer that costs ~$1.50 today could fall to ~$0.33 under the base case estimate.

What is EIP-7732 ePBS and why does it matter?

EIP-7732 introduces Enshrined Proposer-Builder Separation (ePBS), moving the block-building process from the off-chain MEV-Boost relay system into the core Ethereum protocol. Today, over 85% of Ethereum blocks are built by a handful of external MEV firms through relays that operate entirely off-chain. ePBS brings this relationship on-chain, subjecting builders to Ethereum's slashing conditions and accountability mechanisms for the first time. The result: improved censorship resistance, elimination of off-chain relay dependencies, and a more decentralized block production system. ePBS does not directly reduce fees but improves the structural integrity of the network.

What is EIP-7928 and how does it enable 10,000 TPS on Ethereum?

EIP-7928 introduces Block-Level Access Lists (BALs), which allow transactions to pre-declare which parts of Ethereum's state they will read or write before execution. Because non-conflicting transactions can be identified upfront, the EVM can execute them in parallel rather than sequentially. Combined with a higher gas limit, researchers estimate this could push theoretical throughput to approximately 10,000 transactions per second — compared to Ethereum's current practical limit of ~15–30 TPS on Layer 1. Real-world throughput will be lower, but even at 20–30% of theoretical maximum, the improvement in Layer 1 capacity would be transformational.

When does Glamsterdam launch? Could it be delayed?

Glamsterdam targets Q3 2026. As with all Ethereum upgrades, the final date depends on devnet testing, testnet deployment (Sepolia and Holesky), and client team readiness across all major Ethereum clients (Geth, Nethermind, Besu, Erigon, Reth). Given the complexity of EIP-7928's parallel execution changes, an extended testnet period is expected. If significant issues emerge, the timeline could slip to Q4 2026. The community has explicitly prioritized correctness over schedule, so delays should not be interpreted as failure — they are the process working as designed.