Fusaka goes live on December 3! Ethereum L2 fees drop by 60%, challenging 100,000 transactions per second.

On December 3, 2025, Ethereum will launch the Fusaka upgrade on the Mainnet, marking the second major Hard Fork of the year following May's Pectra. The core feature of Fusaka, PeerDAS, allows validators to verify summarized data Blocks without the need to download all data, thereby dropping bandwidth and storage demands, while significantly enhancing data throughput to support a modular stack of 100,000 transactions per second.

What is Fusaka? Ethereum's Second Annual Hard Fork

(Source: Dune Analytics)

Rollups currently carry most of the Ethereum transactions and fee income, but they are still limited by the amount of data that can be published back to Layer 1 and the associated costs. Fusaka aims to alleviate this pressure. Its core feature, PeerDAS (Peer Data Availability Sampling), allows validators to verify aggregated Block data without the need to download all the data, thereby dropping bandwidth and storage requirements while significantly increasing data throughput.

At the same time, the fork that only includes Block Parameter Object (BPO), new gas and block size limits, as well as historical record expiration adjustments, enables the blockchain to adapt to repeated capacity increases rather than a one-time capacity leap. This incremental upgrade strategy allows Ethereum to flexibly adjust according to actual demand, avoiding the risks and complexities that may arise from excessive upgrades.

The name Fusaka comes from the two internal upgrade code names Osaka (execution layer) and Fulu (consensus layer), merged into “Fusaka.” This naming convention has become a tradition in the history of Ethereum upgrades, such as Dencun (Deneb + Cancun), Shapella (Shanghai + Capella), etc. In the final coordination call of Fusaka, the launch slot for the mainnet was set to 13,164,544, expected to launch around December 3 at 21:49 UTC.

PeerDAS and Blob Capacity 8x Growth Path

The core extension of Fusaka has changed to EIP-7594, namely PeerDAS. PeerDAS does not require every full node to download the entire Rollup data block, but instead breaks it into smaller cells and uses sampling and erasure coding techniques, allowing validator nodes to only retrieve random fragments. If there are enough available fragments, the network can be confident that the complete data exists. This can reduce the bandwidth and storage requirements for each node and lays the foundation for eventually achieving an 8-fold increase in blob capacity over time, without forcing home users to use data center hardware.

Blob is a temporary data package uploaded to Ethereum by Rollup. They are cheaper than calling data and will be automatically deleted after about 18 days, so they do not bloat the chain. Dencun (March 2024) introduced EIP-4844 “blobs”, which is a cheaper temporary data channel for rollup, also known as protodanksharding. Fusaka further optimizes on this basis, paving the way for future capacity expansion.

To make this growth more flexible, EIP-7892 introduces a Blob Parameter Only (BPO) fork, which is a minor Hard Fork that only changes three blob-related parameters: target, maximum value, and base fee adjustment factor. After Fusaka, Ethereum can increase blob capacity in a smaller and more frequent manner as L2 demand grows, without having to wait for several years for a major explosive fork.

Core Technological Innovations of Fusaka

PeerDAS (EIP-7594): Validators only need to download data fragments instead of the whole, dropping hardware requirements.

BPO fork (EIP-7892): Flexibly adjust blob parameters to achieve incremental capacity expansion

Gas and Block Size Optimization: Increase gas target value and raise block size limit to 10 MB.

Historical Record Expiration (EIP-7642): Nodes can discard old data to save hundreds of GB of space.

Fusaka's Strategic Position in the Ethereum Roadmap

To understand the position of Fusaka, it is necessary to review the history of Ethereum upgrades. The Merge in 2022 transitioned Ethereum from a proof-of-work mechanism to a proof-of-stake mechanism, thereby reducing energy consumption by approximately 99.9%. Shapella (2023) supports the withdrawal function for staked ETH, transforming the one-way staking system into a liquidity system. Dencun (March 2024) introduced blobs. Pectra (May 2025) added EIP-7702 account abstraction features and improved staking parameters.

These upgrades align with Vitalik Buterin's concise roadmap: Merge, Surge, Verge, Purge, and Splurge. Surge aims to scale Ethereum through Rollups and better data availability, while Verge and Purge focus on lighter clients and clearing old historical records.

Fusaka is the first upgrade to simultaneously advance all these features. It expands the data scale of Rollup through Surge and optimizes the expiration of historical records and lighter synchronization mechanisms through Verge and Purge. In addition, it sets clear goals for the modular Ethereum stack, increasing Layer-2 throughput based on L1 settlement to achieve over 100,000 transactions per second (TPS).

Hidden Upgrades to User Experience and Developer Tools

Not all content from Fusaka focuses solely on the original capacity. Some EIPs also emphasize user experience, security, and developer convenience. EIP-7951 adds secp256r1 precompiled contracts, enabling Ethereum to natively support P-256 signatures, which are adopted by protocols such as Apple's Secure Enclave, Android Keystore, FIDO2, and WebAuthn. This allows wallets to rely on device-level biometric recognition and keys instead of mnemonic phrases, bringing the first layer closer to mainstream fintech login processes.

The significance of this feature is profound. The mnemonic phrase has always been one of the biggest user experience barriers in crypto wallets, as ordinary users find it difficult to securely store 12 or 24 words. If they can use their phone's Face ID or fingerprint recognition to manage the wallet, it will greatly drop the access threshold and attract more mainstream users.

Developers can obtain EIP-7939, which is the leading zero count opcode, used for calculating the number of leading zeros in a 256-bit word. It reduces the cost and difficulty of bit-level mathematical operations, large integer computations, and the implementation of some zero-knowledge proof circuits. EIP-7917 (Deterministic Proposer Front-running) makes the proposer scheduling for the next epoch completely deterministic and can be accessed on-chain via the beacon root. This is crucial for mechanisms based on Rollup and pre-confirmation.

Glamsterdam 2026 and the ultimate goal of 100,000 TPS

The next upgrade named Glamsterdam is expected to be released in 2026, and it has two main features: proposer-builder separation (ePBS) and block-level access lists (BAL). ePBS aims to strengthen the Maximum Extractable Value (MEV) supply chain by separating block construction and proposals at the protocol level, rather than relying solely on external relays. BAL is designed to improve execution efficiency and better handle state access, including the future increase in blob capacity.

If Ethereum can maintain its current progress, Fusaka will no longer be just an event, but more like a turning point. It marks a shift in Ethereum's roadmap towards a coherent and value-focused scaling solution. This solution aims to support a modular stack of 100,000 TPS while not sacrificing the decentralized characteristics that initially gave value to the Ethereum network.

A target of 100,000 TPS is extremely ambitious in the blockchain industry. In comparison, the theoretical peak of the Visa network is around 65,000 TPS, with an actual average usage of about 1,700 TPS. If the Ethereum ecosystem can truly achieve 100,000 TPS, it will have the capacity to handle global-level payments and financial applications. The key is that this throughput is primarily provided by Layer 2, while Layer 1 focuses on settlement and data availability guarantees, which is the core concept of modular blockchains.