I read through Binance's explainer on Ethereum's 2026 roadmap, published on June 8, 2026, and the two problems it keeps returning to are the ones that matter: who assembles a block, and how much of a block the network can process at once. Everything on the schedule follows from those, and the upgrade names are just the labels on them.
The headline item is Glamsterdam, the next upgrade discussed. It brings two changes: enshrined proposer-builder separation, usually shortened to ePBS, and block-level access lists. Together they aim to make block production less dependent on outside software and let Ethereum execute more transactions at the same time. Further out sits Hegotá, with a feature list still being settled.
What an Ethereum upgrade changes
Ethereum is a blockchain built to run decentralized applications, and ETH is the token that pays for using it. An upgrade rewrites the rules nodes follow, the way data is stored and the way blocks are put together. You rarely see the change directly. You see it in fees, in confirmation times and in whether an app feels responsive on a busy day.
Each upgrade builds on the last, which is why the 2026 proposals make more sense next to what came before.
| Upgrade | Year | Main change |
|---|---|---|
| Dencun | 2024 | EIP-4844 adds blobs, a new data type that carries rollup data more cheaply |
| Pectra | 2025 | EIP-7691 and EIP-7702 raise blob capacity and add account abstraction for wallets |
| Fusaka | 2025 | PeerDAS (EIP-7594) improves data availability while keeping validator requirements low |
| Glamsterdam | 2026 | ePBS and block-level access lists |
| Hegotá | late 2026 | Feature list still being finalised; focus on scale, stability and usability |
Read the table as a sequence. Dencun and the two 2025 upgrades worked on the cost of posting data from Layer 2 networks. Glamsterdam moves the focus to how Ethereum itself builds and executes blocks. Hegotá is the promised follow-through.
Why demand keeps pressure on block space
Block space is the limited capacity a block gives to transactions. When more people want it than a block can hold, they bid against each other, and that is where congestion, higher fees and slower inclusion come from. Every application on Ethereum draws from that same pool.
Stablecoins, tokenized real-world assets, decentralized finance and enterprise deployment all keep the pressure on, and none of them stops when the network gets busy. A stablecoin transfer still needs to settle and a tokenized asset still needs its record updated. That is the strongest part of the case for the roadmap, in my view: the demand comes from products people already use.
The fixes therefore have to be structural. Making a block cheaper to verify, or letting more of a block be processed at once, changes the economics for everyone downstream. The schedule is continuous maintenance rather than a one-off event, which is the right frame for a network already carrying real volume.
Glamsterdam, part one: ePBS
A block gets made in two jobs. Validators propose and attest to blocks; builders assemble the contents and hand them over. Proposer-builder separation keeps those roles apart, so a validator does not have to be the best at ordering transactions to do its job.
At the moment that separation runs on independent third-party software. It works, but something as basic as assembling a block rests on off-chain components. Enshrined proposer-builder separation moves the function into the core protocol, where network-level rules govern it instead of outside code.
Binance credits the change with dependability, auditability and decentralization. I find that order sensible. Once block assembly is written into the protocol, anyone can read the rules, and the behaviour gets the same review as the rest of the chain. It also lowers the risk that a handful of off-chain services become an unofficial chokepoint.
Block-level access lists and parallel execution
Before a block runs, Ethereum does not know which account balances or which parts of smart contract storage the transactions will touch. Contract storage is a contract's internal memory, the values it keeps between calls. Because that is unknown in advance, nodes process transactions one after another, and a busy block becomes a queue.
Block-level access lists fix the information problem. Each block declares up front which states it will touch. With that list in hand, a node can process independent transactions at the same time instead of in strict sequence. The creator presents this as a route to higher throughput and more predictable fees, since the cost of a transaction is less likely to depend on what else landed in the same block.
The argument assumes the transactions in a block can be paired off without touching the same state. Two swaps against the same pool cannot run at once, however good the access list is. The practical gain depends on how much real traffic is independent. What I take from it is the direction of travel: Ethereum is making its own execution layer faster before leaning harder on outside networks.
How the Ethereum Foundation splits the work
The foundation organises its technical work into three tracks: scaling capacity, improving the user experience, and hardening the Layer 1 chain. Scaling means handling more throughput overall. User experience work simplifies how people move across applications and wallets.
Hardening the base chain keeps consensus and execution secure as more activity arrives. Consensus is how nodes agree on what happened; execution is how they run the transactions they agreed on. If either grows fragile, every app on top inherits the problem, which is why the least glamorous track probably matters most over a long horizon.
The foundation's roadmap post names the people leading each track: Ansgar, Marius and Toni for scaling; Barnabe and Matt for user experience; Fredrik, Pari and Thomas for hardening Layer 1. Named owners and separate tracks are easier to hold to account than one vague promise of faster and cheaper transactions.
Why rollups carry the traffic, and what blobs do
Ethereum's scaling plan is Layer 2 first. Rollups such as Arbitrum, Optimism and Base process the bulk of activity at lower fees and faster speeds, while the base chain stays responsible for settlement and security. Settlement is the final record of what happened; security is the guarantee that nobody quietly rewrites it.
Blobs make that division affordable. A blob is a chunk of data attached to a block, introduced with EIP-4844 in the Dencun upgrade in 2024, and rollups use them to post their transaction records back to Ethereum. The detail that matters is how long the data sticks around: unlike permanent storage, blob data persists only long enough for proofs to be validated, which cuts the cost of posting it.
Cheaper data for rollups means cheaper transactions for the people using them, and the creator ties the fall in Layer 2 costs directly to this design. Pectra followed in 2025 with EIP-7691 and EIP-7702, raising blob capacity and adding account abstraction, which lets wallets handle more of the signing logic on a user's behalf. Fusaka in 2025 added PeerDAS through EIP-7594, a way for nodes to check that data is available without each of them holding all of it, while keeping validator hardware requirements low.
Keep one distinction in mind. Lower fees and faster confirmations on a rollup are not lower fees on Ethereum itself, and the two can move in different directions.
Hegotá and the quantum question
Hegotá is the next named upgrade after Glamsterdam, targeted for late 2026. Its feature list is still being finalised and the stated focus is scale, stability and usability. I would treat any firm Hegotá feature claim you see elsewhere with suspicion until the Ethereum Foundation publishes it.
Quantum computing could eventually break the conventional cryptography that protects wallets and signatures. Tomasz Stanczak of the Ethereum Foundation describes a quantum resistance roadmap already in motion, with researchers standardising parameters and engineers testing code implementations.
Rather than waiting for one all-in-one fix, Ethereum is developing in steady, modular steps so the chain can adapt as demand grows and new technical risks appear. That fits how cryptographic migrations tend to go: they are slow, they touch wallets and clients and user habits, and starting early is the whole point.
Binance's explainer gives no activation date for Glamsterdam and no final feature list for Hegotá. The Ethereum Foundation's own materials are the place to check before you plan around either, and the note on that page says plainly that features and timelines may evolve.
What to check when the next upgrade lands
When the next upgrade headline lands, look past the name and ask what it changes about blocks. Glamsterdam's two items are concrete enough to track: whether ePBS ships and removes the dependence on outside block-building software, and whether block-level access lists deliver more predictable fees under real load. Both can be checked against the chain.
Activity matters too. If rollups keep absorbing demand, the number that matters for most users is the cost of a transaction on Arbitrum, Optimism or Base rather than the cost on the base chain. Blob capacity is the lever to follow through Pectra, Fusaka and whatever comes after.
Protocol work of this size rarely runs on schedule. Hegotá's feature list is still being written, so a delayed upgrade done properly beats a rushed one that breaks settlement on the network everything else relies on.
