PeerDAS has operated successfully for nearly a year on Ethereum

Ethereum’s most consequential networking upgrade in years has now been running without major incident for nearly twelve months. PeerDAS, the protocol that lets validators confirm data availability without downloading full blocks, has held up through a sustained increase in blob capacity and a dramatic collapse in Layer 2 transaction costs.

What PeerDAS actually does

Each node only needs to check roughly one-eighth of the extended data to be confident the full dataset is there. If enough randomly selected samples come back intact, the network reaches consensus that the data is available, even though no single participant held all of it.

The technical machinery underneath involves erasure coding similar to Reed-Solomon. Extended blob data gets divided into 128 columns, which are then distributed across the network using gossip subnets. Each node subscribes to a subset of around eight columns, making the workload light enough that validators don’t need upgraded hardware to participate.

The formal name is EIP-7594, and it shipped as the headline feature of the Fusaka hard fork on December 3, 2025.

Before mainnet activation, the Ethereum development community ran extensive testing across hundreds of nodes on a dedicated devnet.

Blob capacity and the fee collapse

At launch, the target sat at six to nine blobs per block. By December 9, 2025, six days after the Fusaka upgrade, the target climbed to 10 blobs. Less than a month later, on January 7, 2026, it reached 14 blobs.

The fee impact has been striking. In the weeks before PeerDAS activated, Layer 2 users were paying between $0.20 and $0.50 per transaction. By the first quarter of 2026, those costs had compressed to a range of $0.005 to $0.02. That is a reduction of more than 95% in a matter of months.

The reduction is not purely a function of PeerDAS. Ethereum’s prior EIP-4844 upgrade, which introduced blob-carrying transactions in early 2024, had already separated blob fees from execution gas fees. PeerDAS extends that foundation by allowing far more blob data to flow through the network without proportionally increasing the burden on individual validators.

Why decentralization is the real story

Because no single node is required to download or store the complete blob dataset, the protocol achieves what researchers call decentralized data availability consensus. A validator running on modest consumer hardware can participate fully in confirming data availability by sampling only its assigned column subset and relying on peer requests to reconstruct anything missing.

Vitalik Buterin has described the protocol as a meaningful advance in blockchain scalability. The researchers whose names are most closely tied to the work, including Danny Ryan and Dankrad Feist, spent years on the theoretical groundwork before the implementation took shape. Feist’s original danksharding proposal is the larger architecture that PeerDAS represents an early phase of.

Full danksharding would push blob capacity significantly higher still, potentially enabling thousands of blobs per block.

What comes next for Ethereum’s scaling roadmap

The progression from 6 to 14 blobs over roughly five weeks in late 2025 suggests the Ethereum developer community is comfortable moving quickly once a baseline of stability is established.

Disclosure: This article was edited by Estefano Gomez. For more information on how we create and review content, see our Editorial Policy.