Ethereum Finalizes “Fusaka” Upgrade, Advancing Its L1 + Rollup Roadmap

• Ethereum’s Fusaka upgrade enables PeerDAS, boosting Layer 2 speed while cutting transaction costs.
• Sharding now begins but L1 still can’t handle L2-scale loads, mempool remains unchanged.

Ethereum’s mainnet has officially integrated the “Fusaka” upgrade, finalizing a major technical change in the ongoing effort to improve transaction processing and data handling. Fusaka adjusts the foundation for Ethereum’s Layer 1 and Layer 2 interaction. The main purpose is to lower Layer 2 transaction fees, expand data storage capacity, and smooth out the use of rollups, without adding pressure to individual network nodes.

Ethereum Introduces True Data Sharding

One of the central additions in the upgrade is Peer Data Availability Sampling (PeerDAS) through EIP-7594, which allows validators to verify only parts of large data blobs instead of downloading the entire dataset. It is expected to have a strong impact on how rollups like Arbitrum and Optimism operate.

The increased data availability leads to lower rollup costs. Ethereum developers estimate that fees on Layer 2s could drop by 40 to 60 percent. This cost reduction is made possible because rollups now have greater room to publish their data, supported by larger blob capacity and improved throughput introduced by Fusaka.

Ethereum co-founder Vitalik Buterin highlighted the inclusion of PeerDAS in Fusaka, describing it as a breakthrough for Ethereum. “PeerDAS in Fusaka is significant because it literally is sharding,” he said. He added that Ethereum is now finalizing blocks without requiring a full data view from any one validator, calling it “client-side probabilistic verification” rather than consensus by vote.

Ethereum Raises Block Gas Limit to 60M

The upgrade also raised Ethereum’s block gas limit from around 45 million to a target of 60 million. This change allows more transactions to be processed per block. Developers have set a long-term goal of 150 million, aiming to improve network responsiveness. To avoid single transactions clogging the system, a gas cap per transaction has been enforced through EIP-7825.

Another change involves the Blob Parameter Only (BPO) fork switch. Introduced through EIP-7892, it allows smaller and more targeted blob capacity updates in the future. This design avoids the need for complete hard forks every time the network wants to grow its blob capacity. As we reported, the first BPO fork is already scheduled for December 9, 2025, followed by another on January 7, 2026.

Vitalik, however, pointed out that while PeerDAS marks real progress, Ethereum still lacks full sharding in a few areas. First, Ethereum L1 can’t yet handle transaction scaling beyond constant-factor improvements. Second, block building still depends on centralized builders having full access to data. Third, the absence of a sharded mempool limits efficiency during high usage.

Despite these missing parts, Buterin called Fusaka a “fundamental step forward” and projected that the next two years will refine PeerDAS, increase scale gradually, and prepare the network to handle scaling at the Layer 1 level when zero-knowledge EVMs mature. 

Following the upgrade, ETH’s price climbed by 4.51%, trading at $3,193.37, reflecting  strong investor interest in technical improvements.