Ethereum completed its Fusaka upgrade on December 3, marking one of the network’s most essential steps toward long-term scalability.
The upgrade builds on a series of changes following Merge 2022 and follows previous Dencun and Pectra releases that lowered layer 2 rates and increased blob capacity.
Fusaka goes further by reorganizing how Ethereum checks whether data is available, expanding the channels through which layer 2 networks such as Arbitrum, Optimism, and Base publish compressed batches of transactions.
This is done through a new system called PeerDAS, which allows Ethereum to verify large amounts of transaction data without downloading it from every node.
Buterin said Husaka was ‘incomplete’.
However, Ethereum co-founder Vitalik Buterin warned that Husaka should not be viewed as a finished version of sharding, the network’s long-term expansion plan.
Buterin noted that PeerDAS represents the first practical implementation of data sharding. However, he pointed out that several important components remain unfinished.
According to him, Ethereum now makes more data available at a lower cost, but the overall system conceived over the past decade still requires work across multiple layers of the protocol.
Taking this into account, Buterin highlighted three gaps in Fusaka’s sharding:
First, Ethereum’s base layer still processes transactions sequentially. This means that execution throughput did not increase with the new data capacity.
Second, block builders, specialized actors who assemble transactions into blocks, continue to download the entire data payload even though validators are no longer needed, creating centralization risk as data volume increases.
Lastly, Ethereum still uses a single global mempool, which forces all nodes to process the same pending transactions and limits the scalability of the network.
His message essentially frames Fusaka as the foundation for the next development cycle. He said:
“Over the next two years, there will be time to improve the PeerDAS mechanism, carefully scale it up while continuing to ensure stability, use it to scale L2, and once ZK-EVM matures, transition in-house to also scale Ethereum L1 gas.”
Glamsterdam becomes the next focus.
Fusaka’s most direct successor is the Glamsterdam upgrade, targeted for 2026.
As Fusaka expands Ethereum’s data bandwidth, Glamsterdam wants to ensure that the network can handle the resulting operational load.
The headline feature is the proposer-builder separation, known as ePBS. This change shifts block composition into the protocol itself, reducing Ethereum’s dependence on the small number of external block builders that currently dominate the market.
As data volume grows under Fusaka, its builders will gain more influence. ePBS is intended to prevent this outcome by formalizing how builders bid on blocks and how validators participate in the process.
Running alongside ePBS is a complementary feature called block-level access lists. This list requires the builder to specify which part of the Ethereum state the block will affect before execution begins.
The client team says this allows the software to schedule tasks more efficiently and lays the foundation for future parallelization. This is an essential step as the network prepares for more computational load.
Together, ePBS and Access List form the core of Glamsterdam’s market and performance reforms. This is considered a structural prerequisite for operating large data systems without sacrificing decentralization.
Other planned Ethereum upgrades
Beyond Glamsterdam is Verge, another roadmap milestone centered around the Verkle tree.
This system reorganizes how Ethereum stores and verifies network state.
Instead of requiring a full node to store the entire state locally, Verkle trees allow blocks to be verified with compressed proofs, significantly reducing storage requirements. In particular, this was partially addressed in Fusaka.
For node operators and validators, this is consistent with one of Ethereum’s core priorities: ensuring that running nodes remains accessible without the need for enterprise-grade hardware.
This work is important because Fusaka’s success increases the amount of data Ethereum can collect. However, if you do not change state management, the cost of maintaining the chain may eventually go up.
The Verge aims to ensure the opposite and make Ethereum easier to run even though it handles more data.
Ethereum has since focused on purge updates, a long-term effort to remove accumulated historical data and eliminate technical debt, making the protocol lighter and easier to operate.
In addition to these changes, there’s Splurge, a suite of upgrades designed to improve the user and developer experience.
This will be achieved through improved account abstraction, new approaches to MEV mitigation, and ongoing encryption improvements.
Global settlement layer
Taken together, these updates form successive steps in the same ambition.
“Ethereum is positioning itself as a global payments layer capable of supporting millions of transactions per second through its Layer 2 ecosystem while maintaining the security guarantees of the base chain.”
Ecosystem figures over long periods of time increasingly reflect that framing. Ethereum co-founder Joseph Lubin said:
“The global economy will be built on Ethereum.”
Lubin pointed out that the network has operated without interruption for nearly a decade and was responsible for settling more than $25 trillion in value last year.
He also noted that Ethereum currently accounts for the largest share of stablecoins, tokenized assets, and real-world asset issuance, with ETH itself becoming a productive asset through staking, re-staking, and DeFi infrastructure.
His remarks capture the broader theme behind the current roadmap: a payments platform that runs continuously, absorbs global financial activity, and is open to any participant seeking verification or transactions.
According to CoinGecko, its future depends on three outcomes: The network must remain scalable and roll up to handle high volumes of activity at a predictable cost. Security must be maintained by relying on thousands of independent validators, whose ability to participate is not limited by hardware requirements. And it must remain decentralized so that anyone can run a node or validator without specialized equipment.