Ethereum completes Fusaka upgrade, unlocking up to 8 times the data throughput.

Author: Chloe, ChainCatcher

Ethereum officially completed a major upgrade called "Fusaka" this morning. This is the second hard fork since the Pectra upgrade in May of this year, symbolizing the Ethereum development team's determination to accelerate iterations and officially entering a cycle of two upgrades per year. Meanwhile, the price of ETH has climbed to a high of $3240 in recent days, an increase of about 20% from the recent low, reflecting positive market sentiment.

The Fusaka upgrade (named a combination of Fulu and Osaka) not only covers comprehensive optimizations of the consensus and execution layers, but also introduces as many as 12 Ethereum improvement proposals, focusing on improving network throughput, optimizing transaction speed, and revising the economic model to strengthen ETH's deflationary mechanism.

The Fusaka upgrade not only covers comprehensive optimizations of the consensus and execution layers, but also introduces as many as 12 Ethereum improvement proposals, focusing on improving network throughput, optimizing transaction speed, and revising the economic model to strengthen ETH's deflationary mechanism.

The core highlight of this upgrade is undoubtedly the "PeerDAS (Peer Data Availability Sampling)" technology. This innovation is considered a key step in Ethereum's scaling path, and it is expected to reduce transaction latency from minutes to milliseconds, bringing a more immediate user experience to decentralized applications and payment systems. Core developers witnessed the upgrade live, demonstrating Ethereum's commitment to improving mainnet speed and performance. The Fusaka upgrade process proceeded smoothly, reaching final confirmation within approximately 15 minutes after being triggered at 21:49 UTC last night. According to Coindesk, several core developers witnessed this moment together during the EthStaker livestream. Consensys engineer Gabriel Trintinalia stated, "Fusaka clearly demonstrates Ethereum's commitment to improving mainnet speed and performance. In the early development phase of the Fusaka upgrade, any features that could potentially delay a fork, such as those requiring more research or being overly complex, were downgraded and removed from the development scope." The Ethereum Foundation describes the upgrade as laying the foundation for "instant-feel user experiences," and the technology will unlock up to 8 times the data throughput. For L2 scaling solutions like Optimism and Arbitrum, this equates to submitting more data at a lower cost, thereby reducing transaction fees for end users while preserving significant growth potential for the network. PeerDAS Introduces "Sampling" Concept, Unlocking Up to 8x More Data Throughput. As previously reported, PeerDAS was originally planned for inclusion in Ethereum's major February upgrade, Pectra, but was delayed due to testing requirements. PeerDAS, short for Peer Data Availability Sampling, is a data processing mechanism designed to address bottlenecks on the Ethereum mainnet when processing L2 commit data. Simply put, since the Dencun upgrade introduced "Blobs" (a temporary data storage space designed specifically for L2) in 2024, validators have had to completely download and verify the entire contents of each Blob, leading to heavy network bandwidth burdens, limited processing performance, and indirectly increased transaction fees. According to Coin Metrics, since its introduction, Blob has seen significant adoption thanks to rollups like Base and Arbitrum. However, this has also led to Blob utilization often approaching saturation (currently close to the target of 6 Blobs per block), potentially triggering an exponential increase in rollup fees. As the demand for data availability continues to grow, Blob space has become a key bottleneck in Ethereum's scaling path. PeerDAS's innovation lies in introducing the concept of "sampling," where validators don't need to download the entire Blob but only check randomly selected "slices" (small fragments). Through this peer-to-peer network sharing and verification method, the system ensures that data availability and security are not compromised, while significantly reducing computational and storage requirements. According to official estimates, in addition to unlocking up to 8 times the data throughput, allowing the L2 network to submit more data at a lower cost, this upgrade is expected to lower the barrier to entry for small or new validator operators by reducing the resources required to run a small number of validators. However, Ethereum developers also point out that large institutions operating a large number of nodes, such as staking pools, will not see the same savings. While the savings for large institutional validators are smaller, overall, this will make Ethereum more inclusive and attract more participants to the ecosystem. "These improvements will take several months to fully take effect as we will only gradually increase the number of blobs to ensure the network can safely handle the increased throughput," said Marius Van Der Wijden, a core developer at the Ethereum Foundation. The Reserve Price Mechanism links Blob fees to L2 execution costs. Besides PeerDAS, the Fusaka upgrade also made precise adjustments to the economic aspects, particularly addressing the sluggish Blob fee market. Since the Dencun upgrade, due to oversupply, Blob fees have frequently fallen to an ineffective level of 1 wei, causing the ETH Burn Mechanism to almost fail in this area. According to Blockworks statistics, in the months following the Pectra upgrade, Blob fees were only around $900, and even the brief peak in November was only $23,000, contributing little to the deflationary pressure on ETH. To address this, EIP-7918 introduced the "Reserve Price Mechanism," linking Blob fees to L2 execution costs to prevent price collapse and ensure consistency with actual processing costs. This not only stabilizes market volatility but also allows Blob fees to contribute more effectively to ETH burning as L2 transaction volume grows. Combined with EIP-7892's "Blob-only hard fork" (BPO) feature, Ethereum will increase the target number of Blobs per block to 14 (maximum 21) by January 7, 2026, further expanding capacity. Other EIPs cover protocol cleanup and performance improvements, such as: EIP-7935: Increases the default block gas limit to 60 million, allowing for more computation and providing flexibility for future adjustments. EIP-7951: Adds native support for secp256r1 (P-256) signatures, allowing wallets to integrate device biometrics and improve user login convenience. EIP-7825 and EIP-7934: Set transaction and block size limits to prevent resource-intensive attacks such as DoS. EIP-7883: Increases gas costs for certain mathematical operations to ensure fair allocation of network resources. EIP-7642: Removes outdated message fields, simplifies the protocol, and cleans up code. Ethereum co-founder Vitalik Buterin stated in an article on the X platform that Ethereum has been continuously introducing "hard, fixed rules" in recent years to improve protocol security and long-term adaptability. He recalled that in 2021, EIP-2929 and EIP-3529 adjusted storage costs and reduced gas returns; in 2024, Dencun upgraded and weakened contract destruction instructions; and in 2025, a gas cap of 16,777,216 gas was set for a single transaction. Vitalik emphasized that these changes, by setting clear processing caps, help prevent DoS attacks, simplify clients, and expand efficiency margins. He anticipates adding more restrictions in the future, including limiting the total bytes of accessible code (short-term increase in costs for large contracts, medium-term adoption of binary trees and block-based billing), setting maximum validator computation cycles and adjusting synchronization costs for the zero-knowledge EVM, and optimizing memory billing to clearly define the maximum EVM consumption cap. Fusaka's success not only addressed current pain points but also laid the foundation for Ethereum's rapid development pace. The major upgrades, previously held annually (such as Shapella in 2023 and Dencun in 2024), have now become semi-annual, demonstrating the foundation's strong execution capabilities despite personnel changes. Next year's major upgrade, "Glamsterdam" (Gloas + Amsterdam), is expected to focus on parallel processing technologies such as Block-Level Access Lists (BAL) to further improve performance.