On June 20, during the 214th Ethereum All Core Dev Execution (ACDE) call, core developers agreed to retain the Fusaka upgrade’s final scope with only one addition—EIP-7939—bringing the total number of EIPs included to 12. With this decision, Fusaka officially transitioned from the planning stage to the implementation phase.
As the largest bundled hard fork since The Merge, Fusaka is expected to unlock a significant order-of-magnitude increase in Layer 2 (L2) data capacity, assuming it is deployed by the end of 2025. Substantially lower L2 fees over the next one to two years would help solidify Ethereum’s competitive edge.
1. The Continuous Scaling Logic of Ethereum's Roadmap
It’s well known that Ethereum’s scalability challenges were once a major bottleneck, resulting in high mainnet transaction costs and hindering the broader adoption of decentralized applications.
According to data shared publicly by Vitalik Buterin in April this year, Ethereum Layer 1 currently processes around 15 transactions per second (TPS). The gas limit has recently increased to 36 million—roughly a sixfold rise over the past decade.
Meanwhile, the most significant gains have come from Layer 2 (L2) networks, where throughput has now reached approximately 250 TPS, reflecting real progress in scalability. These improvements are not just theoretical: many users have experienced substantially lower fees and faster transaction speeds.
Over the past year, transfer fees on L2s like Arbitrum, Optimism, and Base have dropped to around $0.01—or even lower—representing a one or multi-order-of-magnitude decrease. Gas costs on the Ethereum mainnet have also become noticeably more user-friendly, although they still fluctuate with market activity and on-chain demand.
This transformation is no accident. It’s the result of Ethereum’s methodical adherence to its long-term roadmap and continuous iteration. Here’s a brief look at some of the key network upgrades over recent years:
- 2022: The Merge — Ethereum transitioned to a Proof-of-Stake (PoS) consensus mechanism, drastically reducing energy consumption and freeing up execution-layer capacity for future improvements.
- 2024: Dencun — Introduced blob-based data availability, providing temporary low-cost storage for L2s and significantly reducing rollup costs, thereby opening a new path to scalability.
- 2025: Pectra— Successfully activated on May 7, this upgrade improved validator operations and enhanced participation flexibility within the PoS system.
- Upcoming: Fusaka— As the next major step, Fusaka is expected to go live on mainnet in Q3 or Q4 2025 (pending final confirmation). According to Tomasz Kajetan Stańczak, Co-Executive Director of the Ethereum Foundation, Fusaka will include several important EIPs—most notably PeerDAS (Peer-to-Peer Data Availability Sampling)—further improving Ethereum’s scalability and moving it closer to mainstream adoption.
From The Merge → Dencun → Pectra → Fusaka, Ethereum has been steadily advancing toward its long-term vision: to build a secure, scalable, decentralized, and sustainable global network.
2. A Comprehensive Look at the Fusaka Upgrade
The 12 core EIPs included in the Fusaka upgrade span multiple technical domains, including data availability, node lightweighting, EVM optimization, and improved coordination between the execution and data layers.
Among them, the most highly anticipated proposal is EIP-7594 (PeerDAS), which introduces a Data Availability Sampling (DAS) mechanism. This allows network validators to verify data by downloading only a small portion of blob data, rather than the full dataset.
This dramatically reduces the network’s bandwidth and storage requirements while improving verification efficiency—laying the groundwork for scalable, high-throughput Layer 2 (L2) transaction processing. The “blob” concept originates from EIP-4844, introduced as part of the 2024 Dencun upgrade.
As one of the most significant Ethereum milestones in 2024, EIP-4844 enabled blob-carrying transactions, allowing L2s to move away from traditional calldata-based data storage. This substantially lowered transaction fees on L2s by providing a more cost-efficient data handling model.
So what exactly are blob-carrying transactions? In short, they embed large volumes of data into “blobs” that are attached to transactions. This reduces the storage and processing burden on the Ethereum mainnet, as blob data does not count toward Ethereum’s state. It directly addresses L1 data availability costs, enabling L2s to offer faster, cheaper transactions—without compromising Ethereum’s core security or decentralization.
Blob-based scaling was further extended in the Pectra upgrade, which went live in May 2025 and increased the blob capacity from 3 to 6 per block. Notably, Vitalik Buterin has suggested that the Fusaka upgrade could further raise the limit to 72 blobs per block, with an initial increase to 12–24 blobs. If DAS is fully implemented, theoretical capacity could reach up to 512 blobs per block.
Once deployed, this upgrade is expected to boost L2 throughput to tens of thousands of transactions per second (TPS), vastly improving performance and cost structures for high-frequency use cases like DeFi, social networks, and on-chain games. This direction aligns with one of the core goals in Vitalik’s proposed “A simple L2 security and finalization roadmap.”
Fusaka also aims to streamline Ethereum’s state and node structures through the introduction of Verkle trees. This will significantly reduce the size of state proofs, enabling light clients and stateless verification, while promoting greater decentralization and broader mobile adoption across the network.
In addition, Fusaka addresses long-standing flexibility and performance bottlenecks in the Ethereum Virtual Machine (EVM) through several targeted proposals:
- EIP-7939 (CLZ opcode): Introduces an efficient bitwise operation to accelerate cryptographic computations, improving overall EVM performance.
- EIP-7951 (precompile for secp256r1 curve support): Enhances Ethereum's compatibility with Web2 infrastructure and enterprise environments.
- EIP-7907: Expands the maximum contract size limit, enabling the deployment of more complex application logic and offering greater flexibility to developers.
To ensure that scaling enhancements don’t compromise network stability, Fusaka also introduces mechanisms for resource control:
- EIP-7934: Establishes block size limits to prevent blocks from becoming excessively large due to blob expansion.
- EIP-7892 / EIP-7918: Adjust blob-related fee structures to prevent resource abuse and enable dynamic pricing based on supply and demand conditions.
3. A Watershed for Ethereum Scaling and Experience
Stepping back to view the broader picture, it's clear that Fusaka is more than just a technical upgrade—it’s poised to serve as a bridge from scalability to usability, impacting multiple layers of the Ethereum ecosystem.
- For Rollup developers, it means lower data availability costs and greater design flexibility.
- For wallets and infrastructure providers, it enables support for more complex interactions and heavier node workloads.
- For end users, it promises cheaper, faster, and more responsive on-chain experiences.
- For enterprises and compliance-focused users, EVM extensions and simplified state proofs will ease integration with regulatory systems and large-scale infrastructure.
That said, cautious optimism remains appropriate. As of this writing, Fusaka is still undergoing testing across multiple devnets, and its final launch timeline remains subject to change. In a best-case scenario, Fusaka could go live on mainnet by the end of 2025—potentially becoming Ethereum’s most significant milestone since The Merge.
Ultimately, Fusaka is not just about improving scalability—it represents a critical step in Ethereum’s evolution toward serving mainstream businesses and everyday users. It’s expected to lay the technical groundwork for the next generation of rollup ecosystems, enterprise-grade DApps, and enhanced on-chain user experiences.
With Fusaka, Ethereum’s long-anticipated watershed moment—the tipping point for true mainstream adoption—may finally be approaching.