Glamsterdam

What Glamsterdam is

The Glamsterdam upgrade is the next scheduled hard fork on Ethereum’s roadmap, and it carries the network’s biggest bet yet on parallel execution. The name is a portmanteau, drawn from Gloas on the consensus layer and Amsterdam on the execution layer, the two halves of the network that every upgrade has to coordinate. Client teams have set a working target of the end of August 2026 while describing that date as aspirational and dependent on devnet readiness, the standard caveat for a change of this scope.

The two proposals

Two proposals define the release. The first is EIP-7732, Enshrined Proposer-Builder Separation, which writes into the protocol the division between the party that proposes a block and the party that builds its contents, and adds in-protocol commitments from builders. The second is EIP-7928, Block-Level Access Lists, which requires each block to declare in advance which parts of the network’s state its transactions will touch. Read together, the two enable parallel transaction execution: if the network knows ahead of time which transactions touch which data, it can process those that do not overlap at the same time rather than strictly one after another. That is the core of the Glamsterdam upgrade, and the reason it draws so much attention.

Parallel execution is the ambition behind both proposals. Most blockchains process transactions in a strict single file, one after another, because any two might touch the same data and conflict. Declaring in advance which data each transaction needs lets the network prove which ones are independent and run them together, which is how a general-purpose chain can raise throughput without simply asking every node for more raw power. It is a well-understood technique in conventional computing, and bringing it to a decentralised network safely is what makes the Glamsterdam upgrade difficult.

The capacity climb

The release also continues the long climb in block capacity. The gas limit path runs from around 60 million toward 200 million, with developers testing at a 150 million reference limit in order to derive accurate state pricing, the cost assigned to operations that read from or write to the network’s growing store of data. Getting that pricing right is the unglamorous work that keeps a higher ceiling safe: if heavy operations are underpriced, a larger block becomes an attack surface. The throughput ambition attached to this work has been cited at up to 10,000 transactions per second, an order of magnitude beyond what the base layer handles today. That figure is an aim rather than a measured result, and it depends on the parallel-execution machinery landing as designed.

What comes after

The Glamsterdam upgrade is not the end of the sequence. Hegotá is expected to follow in late 2026 or early 2027, bringing Verkle trees, a change to how the network stores its state that is designed to keep state growth manageable as usage rises. Where Glamsterdam concentrates on doing more work per block, Hegotá concentrates on stopping the accumulated data from becoming a burden, and the two are complementary parts of the same scaling effort.

What to watch

For anyone following Ethereum’s direction, the Glamsterdam upgrade is the clearest test yet of whether the base layer can raise throughput sharply without giving up the properties that make it trusted. Parallel execution, enforced builder commitments and careful state pricing are ambitious things to ship together, and the August 2026 target should be read as intent rather than a promise; the dates will move if the devnets are not ready. What to watch is not the calendar but the testing, because whether the parallel-execution design holds up under load is the question the glamsterdam upgrade ultimately has to answer. Our Ethereum explainer places this upgrade within the longer roadmap.