Ethereum Network Updates 2026: Glamsterdam Upgrade Explained

What Is the Glamsterdam Upgrade?

Ethereum upgrades every 12–18 months through coordinated hard forks — protocol changes that all nodes must adopt simultaneously. Glamsterdam follows Dencun (March 2024, which cut Layer 2 fees by ~90% via EIP-4844 blob transactions) and Pectra (May 2025, which improved validator operations and account abstraction). But those two upgrades left one core bottleneck untouched: Ethereum's Layer 1 execution model still processes transactions one at a time, sequentially.

Glamsterdam changes that. It targets the execution layer directly, introducing two major EIPs and coordinating with validators on a gas limit increase that together represent the most aggressive throughput expansion Ethereum has attempted since the Merge in September 2022.

EIP-7732: Better Block Building, Less MEV Extraction

Every Ethereum block is built by a validator — someone who has staked 32 ETH. Today, most validators outsource the actual block construction to specialized builders via off-chain relay services. These builders optimize transaction ordering to extract Maximal Extractable Value (MEV): front-running trades, sandwiching DEX orders, and capturing arbitrage at the expense of ordinary users.

The current system is entirely off-chain: no on-chain accountability, no transparency guarantees, and deep concentration — a handful of builders construct the majority of Ethereum blocks. EIP-7732, called Enshrined Proposer-Builder Separation (ePBS), brings this relationship into the core protocol. Builders submit sealed, on-chain bids for the right to construct each block's execution payload. The process is now governed by Ethereum's consensus rules, with slashing conditions for misbehavior.

For everyday users, the practical effects are:

• Improved censorship resistance: Validators can no longer quietly exclude your transactions without it being visible and attributable on-chain.
• Reduced systemic risk: Off-chain relays — a regulatory chokepoint and single point of failure — are eliminated from the critical path.
• More MEV accountability: Builder behavior is subject to Ethereum's on-chain accountability mechanisms for the first time.

EIP-7928: Parallel Transaction Execution

This is the upgrade that most directly drives the fee reduction estimate. Currently, the Ethereum Virtual Machine (EVM) processes transactions in strict sequence: one transaction completes before the next begins. This is safe — sequential processing prevents state conflicts — but it leaves enormous computational capacity on the table.

EIP-7928 introduces block-level access lists. Before a block is executed, every transaction declares in advance which parts of Ethereum's state (accounts, storage slots, contracts) it will read or write. With this complete conflict map available upfront, the EVM scheduler can identify which transactions have no overlapping state access — and run them simultaneously.

Researchers analyzing representative Ethereum blocks estimate that 60–80% of typical transactions can be parallelized — simple ETH transfers, ERC-20 approvals, DEX swaps on separate trading pairs, and NFT operations that touch entirely disjoint state. Transactions that do conflict are still serialized, as before.

Under ideal conditions, EIP-7928 opens a theoretical throughput ceiling of roughly 10,000 transactions per second. Real-world performance will be lower — hardware requirements increase, and actual parallelizability depends on what users are actually doing on the network. But even at 20–30% of theoretical maximum, the improvement in L1 capacity would be transformational for use cases currently forced onto Layer 2 because of L1 costs.

Gas Limit Increase: From 60M to 200M

The third component of Glamsterdam is unusual: the gas limit increase is not an EIP. The Ethereum block gas limit is a parameter that validators adjust incrementally by social consensus — a supermajority of validators signaling the same direction causes the limit to drift. The current level of ~60 million gas units is already the result of gradual increases validators have been implementing since the Merge.

The proposal circulating in Ethereum research forums would raise this to 200 million gas units per block — a more than 3x expansion. The core argument: modern Ethereum client software (Go-Ethereum, Nethermind) can handle significantly more computation per block than the current limit allows, and hardware requirements have not grown proportionally with previous adjustments.

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 computational cost. This combination — more supply, lower processing overhead — is the primary driver behind the ~78% average fee reduction estimate that has circulated in research blog posts and EIP author analyses.

What Changes for Ethereum Users

For someone holding ETH or using DeFi applications, the practical effects of Glamsterdam break down as follows:

Activity	Before Glamsterdam	After Glamsterdam
Simple ETH transfer	$0.50–$5 at peak	Estimated $0.10–$1 at peak
DEX swap (Uniswap)	$2–$30 at peak	Estimated $0.50–$7 at peak
Complex DeFi interaction	$10–$100+ at peak	Estimated $2–$20 at peak
NFT mint	Variable, often $5–$50	Estimated 70–80% lower
Block MEV impact	Unaccountable off-chain	On-chain, auditable

Fee estimates are illustrative based on research models. Actual fees depend on post-upgrade network demand.

If you hold ETH as an investment, Glamsterdam does not change tokenomics or issuance. It does, however, have potential market implications: lower fees tend to increase network utility, which historically correlates with higher demand for ETH as gas. Whether this translates to price impact depends on factors well beyond the protocol level.

For a deep technical analysis of each EIP and the research behind the 78% estimate, read our full investigation: Ethereum Glamsterdam 2026: Fees Down 78% — What Changes for Users.

---