The coming yr can be crucial for Ethereum scaling. In 2026, the Glamsterdam fork will bring perfect parallel processing to the chain and increase the gas limit from 60 million today to 200 million.
A big variety of validators will as a substitute switch from re-executing transactions to verifying zero-knowledge (ZK) proofs. This puts Ethereum Layer 1 heading in the right direction to scale as much as 10,000 transactions per second (TPS) and possibly beyond, although this goal won’t be reached in 2026.
Meanwhile, the number of information blobs will increase (perhaps as much as 72 or more per block), allowing Layer 2s (L2s) to process a whole lot of 1000’s of transactions per second. L2s are also becoming easier to make use of; ZKsync’s recent Atlas upgrade allows funds to stay on mainnet but trade within the fast on-chain execution environment on ZKsync’s Elastic Network.
The planned Ethereum Interoperability Layer will enable seamless cross-chain operations between L2s, privacy can be a spotlight, and improved censorship resistance is targeted for the Heze-Bogota fork by the tip of the yr.
Ethereum in 2026: The Glamsterdam Fork
Ethereum developers are currently finalizing which Ethereum Improvement Proposals (EIPs) to incorporate within the Glamsterdam hard fork, expected in mid-2026. The predominant changes confirmed are Block Access Lists and Enshrined Proposer Builder Separation. Neither sound particularly interesting, but they’ve the potential to speed up the blockchain before switching to ZK technology.
Eventually the core developers will give you cool names for things like “Firedancer,” but until then we’ll keep on with the boring technical names they select.
Glamsterdam: Block access lists (EIP-7928)
Although “block access lists” sound like a censorship scheme, the upgrade actually enables “perfect” parallel block processing.
Previously, Ethereum ran in single-lane mode, with a really long queue of transactions executed one after the opposite so as. Block access lists allow throughput to be scaled to a multi-lane highway, processing multiple transactions concurrently.
The term refers to a card included in every block, developed by the block manufacturer who first ran every thing on some fancy high-end equipment. The map tells Ethereum customers which transactions affect which other transactions, accounts and storage spaces, and what the post-transaction state differences are. This allows them to bundle the transactions and execute them on multiple CPU cores concurrently without conflicts.
It also allows clients to load all vital data from disk into memory first, quite than reading the disk over and another time, which Trintinalia calls “the most important bottleneck now we have.”
Perfect parallel processing will allow Ethereum to perform higher transactions per second and have larger block sizes without increasing the gas limit.
The upgrades carried out in 2026 will scale Ethereum L1 to 10,000 TPS. Source: Growthepie
Glamsterdam: Anchored separation between applicants and developers
The technique of separating block builders and proposers has already begun with MEV Boost, an off-protocol solution that uses centralized relays as intermediaries and settles about 90% of blocks. Enshrined Proposer Builder Separation (ePBS) integrates this process directly into Ethereum's consensus layer for trustless operations.
The idea behind separating the 2 is that block creators compete to pick out and arrange transactions in the very best possible solution to create a block, while proposers determine which block to propose. The aim is to alleviate the centralization pressure of the utmost extractable value (MEV) and improve security, decentralization and censorship resistance.
However, from a scalability perspective, the predominant advantage of ePBS is that it provides more time to generate and distribute ZK proofs across the network. Validators are currently penalized for being slow, which negatively impacts the wait for ZK proofs to be validated. EPBS provides more time for obtaining and validating ZK evidence.
This may give validators more time to acquire proof (and validators more time to generate proof), explained Ethereum researcher Ladislaus von Daniels, adding that ePBS decouples block validation from block execution and, in that sense, provides one other variant of deferred execution.
“This makes opt-in zkAtesting way more incentivized compatible for validators.”
Ethereum Foundation researcher Justin Drake estimates that about 10% of validators will switch to ZK after this point, which is able to allow for further increases in gas limits.
Ethereum Foundation researcher Justin Drake demonstrates ZK proof validation. Source: EthProofs
Ethereum L1 gas increase and L2 blob goal upgrades
The gas limit (which refers back to the throughput on the L1) has already been increased to 60 million. In 2026 it is anticipated to rise significantly – although there are different estimates as to how high it’s going to be.
“I believe in 2026 I'm expecting 100 million pretty soon. Anything beyond that might be just too speculative to think about,” said Gary Schulte, lead blockchain protocol engineer at Besu client. He added that moving to delayed execution could allow for higher gas limits.
Tomasz Stańczak, co-director of the Ethereum Foundation, said on the recent Bankless Summit that the limit would rise to 100 million in the primary half of 2026 and predicted that it will double to 200 million after ePBS. Further improvements could lead to as much as 300 million gas per block being possible by the tip of the yr.
Ethereum creator Vitalik Buterin was more cautious. At the tip of November he said: “For next yr, expect continued growth, but more targeted/less uniform growth. For example, one possible future is: a five-fold increase in gas limits together with a five-fold increase in gas costs for operations whose processing is comparatively inefficient.” Buterin mentioned things like storage, precompiles, and calls to contracts with large contract sizes.
Ethereum will grow in 2026. Source: TenaciousBit
Ethereum 2026 Fork #2: Heze-Bogota
Expect a number of the EIPs acquired by Glamsterdam to be included on this fork, but in line with Forkast, the Fork-Choice Inclusion Lists (FOCIL) are the one EIPs currently on the “Consideration for Inclusion” list. This was intended for Glamsterdam, but was pushed forward after a heated debate as it will have required an excessive amount of work and made life too difficult.
It focuses not on scaling, but on the cypherpunk ideal of censorship resistance by giving multiple validators the flexibility to mandate the inclusion of certain transactions in each block.
“This is a censorship resistance mechanism that ensures that if you’ve not less than a part of the network being honest… your transaction can be included in some unspecified time in the future,” Trintinalia said.
Stay tuned for Part 2 as we delve deeper into scaling the L1 with ZK proofs in 2026.
