Ethereum’s highly anticipated Fusaka hard fork upgrade has successfully gone live on the Hoodi testnet, marking the final major testing phase before its mainnet activation.
This deployment occurred on October 28, 2025, at approximately 18:53 UTC, following successful activations on the Holesky testnet and Sepolia testnet. Hoodi, Ethereum’s newest testnet designed to closely mimic mainnet conditions, serves as the ultimate “dress rehearsal” to iron out any issues under high-stress simulations.
Fusaka is the next phase in Ethereum’s “Surge” roadmap, building on the Pectra upgrade from earlier in 2025. It focuses on enhancing scalability, reducing transaction costs, and improving security—particularly for layer-2 (L2) rollups like Optimism and Arbitrum.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
The upgrade incorporates a suite of Ethereum Improvement Proposals (EIPs) to make the network more efficient for DeFi, NFTs, gaming, and institutional adoption. Key goals include higher throughput: Increasing the block gas limit from 30 million to 150 million units, allowing blocks to process more transactions and complex smart contracts without congestion.
Expanding blob capacity introduced in Dencun and implementing EIP-7594 (PeerDAS), which lets validators sample smaller data segments from L2s instead of full blobs, slashing node resource demands.
Potential 90% drop in L2 fees by optimizing rollup economics. New node protections and a $2 million audit contest to catch vulnerabilities. These changes aim to support Ethereum’s growing ecosystem amid rising usage, with ETH currently trading approximately near $4000.
PeerDAS, short for Peer Data Availability Sampling, is a networking protocol proposed in Ethereum Improvement Proposal 7594 (EIP-7594).
It represents a critical advancement in Ethereum’s scalability roadmap, specifically within the “Surge” phase, which aims to enhance data availability (DA) for Layer 2 (L2) rollups without compromising decentralization or security.
Introduced in January 2024, PeerDAS builds directly on the blob infrastructure from EIP-4844 activated in the Dencun upgrade in March 2024 by allowing nodes to verify the availability of large amounts of data while downloading only a tiny fraction of it—typically around 1/8th, with potential reductions to 1/16th or 1/32nd as the system scales.
In essence, PeerDAS addresses a core bottleneck in Ethereum’s rollup-centric design: L1 data availability. Rollups like Optimism, Arbitrum, and Base process transactions off-chain but must post compressed data as blobs to Ethereum’s Layer 1 for security and verifiability.
Under EIP-4844, every full node had to download all blobs, limiting throughput to about 0.75 MB per block with targets of 3-6 blobs. PeerDAS decouples this by introducing data availability sampling (DAS), where nodes probabilistically confirm data exists without fetching everything.
This enables Ethereum to handle exponentially more data—potentially 400%+ increases in L2 capacity—while keeping node requirements low.PeerDAS is set to activate in the Fusaka hard fork, tentatively on December 3, 2025, following successful testnet deployments on Holesky, Sepolia, and Hoodi.
Ethereum co-founder Vitalik Buterin has highlighted it as “key to L2 scaling,” emphasizing how it lets validators sample data chunks instead of full blobs, slashing costs and boosting efficiency.
The Need for PeerDAS
Ethereum’s evolution toward a rollup-centric future, as outlined in Vitalik Buterin’s scalability trilemma solution balancing scalability, security, and decentralization, relies heavily on efficient DA.
Pre-Dencun, rollups used expensive calldata for DA, which clogged the network and drove up fees. EIP-4844 introduced blobs—temporary, non-executable data carriers optimized for rollups—reducing L2 fees by up to 90% initially.
However, blobs still required full downloads by all nodes, capping scalability at ~1 MB/s aggregate throughput.As L2 adoption surges with daily transactions now exceeding L1, this limit hinders growth.
PeerDAS, part of the path to full “Danksharding,” uses erasure coding and peer-to-peer sampling to scale DA bandwidth by orders of magnitude. It shifts from “complete” DA checks to “partial” ones, where nodes sample random “pieces” of data to infer the whole with high confidence.
PeerDAS operates at the consensus and networking layers, extending blob handling with mathematical and protocol innovations. Each blob from EIP-4844 is broken into cells—small, fixed-size data units (e.g., 31 bytes each, fitting KZG polynomial commitments). Cells are the atomic unit for sampling.
Blobs are extended using one-dimensional Reed-Solomon erasure coding a error-correcting code. This creates a data matrix rows: One per original blob + its parity extensions redundant data for reconstruction.
Vertical slices across all blobs in a block, each tied to a gossip subnet a P2P subgroup for efficient data flooding. The full matrix can be reconstructed from any 50% + 1 of its columns, providing redundancy without duplicating everything.
Proofs: Blob senders (e.g., rollup operators) precompute cell KZG proofs using the Kate-Zaverucha-Goldberg commitment scheme for every cell. These tiny proofs ~48 bytes are bundled into the transaction wrapper, allowing cheap verification.
Nodes join gossip subnets corresponding to their columns, ensuring data floods efficiently within subgroups (e.g., via Ethereum’s libp2p P2P stack). Nodes sample random columns from a diverse set of peers maintained via Ethereum’s peer discovery.
Nodes download ~1/8th of data scalable to 1/32nd by adding columns, enabling 10x+ blob capacity (e.g., from 0.75 MB to 8+ MB per block). L2 fees drop 90%+, supporting DeFi, gaming, and mass adoption. Light clients verify DA with tiny resources; no “supernodes” required.
Keeps Ethereum accessible to hobbyists and mobile users. Reduces bandwidth/storage by 87.5%+; validators sample fragments, not full blobs. Node operators save on hardware; faster sync times.
Probabilistic guarantees against data withholding. Rollups inherit L1’s robust DA without central points of failure. Directly aids ZK-rollups (e.g., via prover networks) and optimistic rollups. Enables “blob-only forks” for even cheaper L2s.
PeerDAS is Ethereum’s elegant solution to the DA trilemma, transforming blobs from a prototype into a scalable powerhouse. By leveraging erasure coding, P2P sampling, and cryptographic proofs, it paves the way for a hyper-scalable, decentralized L2 ecosystem—cheaper fees, faster txs, and broader access.
Sepolia is on mid-tier stress testing. Hoodi final validation; confirmed smooth by client teams like Nethermind. Tentatively scheduled for December 3, 2025, at the earliest at least 30 days post-Hoodi per Ethereum Foundation guidelines. Developers will monitor for any final tweaks during bi-weekly ACDC calls.
Post-Fusaka, Ethereum eyes the “Glamsterdam” upgrade in 2026, featuring enshrined proposer-builder separation (ePBS) for further decentralization. Analysts emphasize the upgrade’s role in cheaper L2 transactions and ecosystem growth.
Node operators are urged to update clients (e.g., Geth, Nethermind) in preparation.This milestone underscores Ethereum’s iterative approach to scaling without compromising decentralization.