Ethereum's Next Major Upgrade, Glamsterdam, Increases Block Capacity by 3.3x on Testnet

Author: William M. Peaster

Compiled by Deep潮 TechFlow

DeepChain Overview: Ethereum’s next major upgrade, Glamsterdam, went live on the Sepolia testnet on October 6. The gas limit increased from 60 million to 200 million within 11 hours, and Geth’s time to process equivalent blocks dropped from approximately 115 milliseconds to around 30 milliseconds. The mainnet launch remains targeted for the fourth quarter of 2026. In our view, a one-time 3.3x increase in L1 throughput is one of the most tangible fundamental catalysts for ETH in recent years. However, this upgrade has been delayed multiple times; the next key milestone is whether the Hoodi testnet will activate as scheduled around October 27. Meanwhile, the gas cost for new storage will rise nearly fivefold, making activities such as airdrops and minting that involve creating large numbers of new addresses significantly more expensive.

The gas limit on the Sepolia testnet reached Glamsterdam’s target of 200 million just 11 hours after the fork. Now, let’s look at how Ethereum’s next upgrade will make room for much larger blocks.

The consensus layer is called Gladius, the execution layer is called Amsterdam, and together they form Glamsterdam—Ethereum’s next major upgrade, which has just passed its first formal test.

The reason is that Glamsterdam went live on the Sepolia testnet yesterday (October 6), and the chain continued to finalize blocks normally throughout the switch without any critical failures.

Of course, this is currently only activity on the testnet, but it signals significant changes ahead for this veteran smart contract network. As Ethlabs researcher Barnabé Monnot said after the news broke:

Highly significant. Completing this upgrade means the L1 scale will increase by at least 3x, the contract size limit will rise by 2.5x, along with a host of practical new features. — Barnabé Monnot (@barnabemonnot)

Ethereum Foundation researcher Toni Wahrstätter expressed the same view, stating that Glamsterdam will bring about Ethereum's largest throughput increase since its launch.

Indeed, numerically, Glamsterdam is designed with a gas limit of approximately 200 million, representing a roughly 3.3-fold increase from the current 60 million. For comparison, the mainnet’s gas limit remained at 30 million for several years until validators gradually raised it to 60 million in three steps by 2025; even after the London upgrade in 2021 doubled the maximum block capacity, the target remained at 15 million.

However, this fork itself will not instantly triple Ethereum’s capacity like flipping a switch. It changes the way blocks are produced, making it safe to generate much larger blocks, after which validators will decide whether to raise the limit.

That said, let’s break down exactly what was changed in this Ethereum upgrade and what steps remain before going live on mainnet.

This fork bundled a total of 18 EIPs (Ethereum Improvement Proposals), but three of these proposals carried the bulk of the workload.

The first item is naturally the enshrined proposer-builder separation, or ePBS for short. Very simply put, this mechanism gives Ethereum validators more time.

Today, most Ethereum blocks are packaged by specialized builders and then delivered to validators via trusted intermediaries called relays. Under EIP-7732 (ePBS), proposers instead first commit to bids signed by builders, after which the builders reveal the full block.

This means that the work of verifying transactions within a block no longer needs to fit within the approximately 4-second attestation window. Under this EIP, most validators will henceforth have about 9 seconds.

The second key pillar of Glamsterdam is block-level access lists (BALs). In simple terms, these lists will enable new capabilities for parallel processing.

As background, current Ethereum clients typically process transactions in blocks one by one because they cannot beforehand determine which state elements each transaction will access. Therefore, EIP-7928 (BALs) will attach a map to each block listing all the state elements it touches along with their resulting values after execution.

With this guidance, clients can prefetch data and validate transactions in parallel. Using a factory analogy: ePBS is like giving the quality inspection team more time, while BALs ensures all necessary materials are prepared in advance, allowing multiple assembly lines to operate simultaneously.

In addition, larger contracts and some repricings are coming. EIP-7954 will increase the contract size limit from 24 KiB to 64 KiB, relaxing a long-standing restriction that forced large applications to split their code.

Meanwhile, this fork will increase the price of what should be scarce—the new state. According to EIP-8037, the cost of creating a new storage slot will rise from 20,000 gas to 97,920 gas; meanwhile, according to EIP-2780, the cost of sending a standard ETH transfer to an existing account remains at 21,000 gas.

The first Glamsterdam block on Sepolia was produced yesterday, with a gas limit of 60 million, matching the current limit on Ethereum mainnet. Subsequently, Sepolia’s limit reached its target of 200 million approximately 11 hours after the fork and has remained at this level since.

The proposer of each block can adjust the gas limit by only about 0.1%, so if every proposer increased it in every block, Sepolia would complete this journey in roughly four hours. The actual time taken, however, is more than twice as long, partly because the two dominant consensus clients, Prysm and Teku, continued to keep validators at the 60-million limit after the fork, unless validators explicitly chose to raise it to 200 million.

Earlier mentioned Wahrstätter then shared initial data showing that these adjustments have already begun to take effect. He found that, compared to before the fork, each client now processes units of gas faster—for example, a 40-million-gas block that Geth previously took about 115 milliseconds to execute now takes only about 30 milliseconds.

This is data from just one node per client, spanning less than a day, but it’s undoubtedly an encouraging start.

Next, after Sepolia, Glamsterdam will be deployed to Ethereum’s Hoodi testnet. The initially proposed activation date is October 27, though it has not yet been finalized; the mainnet target remains sometime in the fourth quarter of 2026.

Also remember that Glamsterdam was initially planned to launch in the first half of 2026, and Sepolia’s fork has been delayed from August 3 to October 6. Therefore, the schedule may still change further.

It is certain that the arrival of Glamsterdam will bring new changes and new challenges.

For example, some Ethereum tools will stop working if they do not promptly update for this fork. The Ethereum Foundation has warned that wallets, indexers, and gas estimation tools that assume a fixed maximum value for the gas limit must be updated to avoid issues.

Additionally, since state growth is being repriced, on-chain activities that create a large number of new accounts or new storage—such as airdrops, minting, and new wallets—will incur higher costs, while everyday transfers will not.

However, these details are ultimately minor issues. The most significant and widespread change here is that Ethereum L1 is about to become much more spacious, which is great news for all kinds of experiments within the Ethereum ecosystem—whether large or small, traditional finance or unconventional use cases.

For example, as I mentioned back in August, Fake World Assets briefly became the largest gas consumer on the mainnet during its most chaotic day. With Glamsterdam, similar traffic surges will have much more room to expand without driving up fees or crowding out everyone else.

In other words, Ethereum is making room for more dreams, more projects, and more activity. Now, it’s up to the Ethereum community to fill this space—addressing existing needs while exploring new directions.