Ethereum Just Rehearsed Its Biggest Block-Building Change in Years

On October 6 at 13:53 UTC, Ethereum's Glamsterdam upgrade activated on the Sepolia test network, marking the start of public testing for the most significant change to how Ethereum builds blocks in years. This was not a mainnet launch and no ETH holder needed to do anything. It was the dress rehearsal. 

Glamsterdam rewrites the mechanics underneath every transaction on the network, including the way orders get sequenced and the way value leaks out of trades through MEV, and that plumbing matters to anyone trading on or against Ethereum even though the live rollout is still months away.

The upgrade went live at a specific point in the chain, epoch 353,024, and it bundles two structural changes that have been in research for years.

The headline change is Enshrined Proposer-Builder Separation, known as ePBS and defined in EIP-7732. Today, the validators who propose Ethereum blocks almost all rely on off-chain relays and software like MEV-Boost to actually assemble those blocks. That separation exists, but it lives outside the protocol, held together by trusted middlemen. ePBS writes the split between proposing a block and building a block directly into Ethereum itself, removing the dependence on those external relays.

The second change is Block-Level Access Lists, defined in EIP-7928. The key points of what Sepolia is now testing:

• ePBS moves builder commitments, payload reveals, and proposer payments into Ethereum's own consensus process rather than trusted off-chain relays

• Block-Level Access Lists let the network process transactions in parallel and give validators more time to verify execution payloads

• A set of state-cost changes reprices how the network charges for the storage that larger blocks create

• Testing moves to the Hoodi testnet next, tentatively October 27, before any mainnet activation

Vitalik Buterin framed the longer arc as Ethereum evolving into a hybrid of blockchain and advanced cryptography. Glamsterdam is the step that makes the block-building layer ready for it.

From the analysis we have built across this series all year, the part of Glamsterdam that matters most to a trader is the one getting the least attention in the general coverage.

Every trade you place on an Ethereum-based venue sits inside a block that someone builds. The builder decides the order of transactions in that block, and that ordering power is where maximal extractable value, or MEV, comes from. 

Builders can place their own transactions ahead of yours, sandwich your trade between theirs, or reorder a block to capture value that would otherwise have been yours. Today that power sits with a small set of off-chain relays that validators trust by convention. ePBS pulls that arrangement into the protocol itself, which over time changes who controls transaction ordering and how transparent that process is.

For a trader, ordering is not an abstraction. It is the difference between the price you expected and the price you got. The same forces that produce slippage on a fast market are shaped by how blocks are built and who gets to build them. A structural change to that layer is a structural change to execution quality, and from what we have observed tracking market structure this year, execution quality is where most traders quietly lose an edge they never measured.

The temptation with a major upgrade headline is to treat it as a reason to position in ETH. That would be a mistake here, and the timeline is the reason.

October 6 was a testnet activation. The mainnet rollout is targeted for the fourth quarter of 2026 with no confirmed date, and large upgrades routinely slip as testing surfaces edge cases. Glamsterdam has already been pushed back once. The Sepolia fork also went live with roughly a seven-day audit buffer, half the usual two-week window, and developers had flagged concerns about builder-level attacks disrupting the testnet. None of that threatens real funds, but it does mean the path from here to mainnet is a monitoring exercise, not a settled date.

Ethereum is trading in a market that has spent the last month reacting to macro and regulatory developments far more than to protocol upgrades. A successful testnet fork is a constructive signal for the Q4 target. It is not a near-term price catalyst, and treating a rehearsal as the main event is how traders get caught paying for news that will not land for months.

Know your liquidation price and your drawdown buffer before you enter, not after. Glamsterdam tells you where Ethereum's architecture is heading over the coming quarters. It tells you nothing about the next candle, and the gap between those two timeframes is exactly where discipline earns its keep.