What Is Danksharding? Ethereum's Scaling Endgame

Danksharding is Ethereum's long-term data scaling design, intended to massively increase the amount of cheap, temporary data space available to rollups by having the network verify that large amounts of data are available without requiring every node to download and store it all. It's named after researcher Dankrad Feist, who proposed the core approach.

The Problem Danksharding Solves

Rollups scale Ethereum by executing transactions off the main chain and posting compressed data back to Ethereum so anyone can verify or reconstruct what happened. That data needs somewhere cheap to live. Historically, rollups posted this data as regular transaction calldata, which is expensive because Ethereum nodes store calldata permanently, forever, as part of the chain's history — even though rollups only need that data available for a short window to allow verification and fraud challenges.

Danksharding's goal is to provide a dedicated, much cheaper form of data space designed specifically for this temporary-availability use case, without forcing every single Ethereum node to store or download all of it.

Data Availability Sampling: The Core Idea

The key technique behind danksharding is data availability sampling (DAS). Instead of every node downloading an entire large data blob to confirm it's available, each node randomly samples small pieces of it. If enough nodes each sample different small pieces and all of their samples come back successfully, this gives strong statistical confidence that the entire blob is actually available — without any single node needing to download the whole thing.

This relies on erasure coding, a mathematical technique that expands data redundantly, such that even if a meaningful fraction of the pieces are missing, the original data can still be fully reconstructed from the remaining pieces. This means a malicious actor can't get away with hiding just a small piece of data — if enough of it is unavailable to fail multiple independent samples, the network will detect the problem.

Proto-Danksharding: The Stepping Stone

Full danksharding is a large, complex undertaking, so Ethereum shipped a simplified interim version first, often called proto-danksharding (EIP-4844), which introduced a new transaction type carrying "blobs" of data. Blobs are priced and gossiped separately from regular calldata, and are deleted from nodes after a set retention window (originally around 18 days) rather than being stored forever. This alone brought a significant reduction in rollup data costs, which flowed through directly to lower fees for layer 2 users.

Proto-danksharding introduced the blob transaction format and fee market but did not yet implement full data availability sampling across the validator set — that's the piece full danksharding still needs to add.

Full Danksharding vs. Proto-Danksharding

Aspect	Proto-Danksharding (EIP-4844)	Full Danksharding
Blob transaction format	Introduced	Same format, extended
Data availability sampling	Not yet implemented	Core feature
Node storage burden	Full blobs, briefly	Only small samples per node
Blob capacity	Limited	Substantially higher
Status	Live	In active research/rollout

Why This Matters for Rollup Costs

Since rollups make up the primary demand for blob space, expanding blob capacity through full danksharding is expected to further reduce the cost rollups pay to post data back to Ethereum, which in turn can lower the fees rollup users pay for transactions. This is central to Ethereum's broader scaling strategy, discussed at a higher level in our overview of Ethereum's roadmap, where danksharding sits under "the Surge."

Tradeoffs and Open Engineering Challenges

Data availability sampling requires careful network design to ensure sampling actually reflects true data availability and can't be gamed. It also depends on erasure coding implementations being correct — a subtle bug in that layer could undermine the safety guarantees the whole scheme relies on. Full danksharding remains a significant, ongoing engineering effort rather than a single flip-of-a-switch upgrade, and its rollout has been phased deliberately to reduce risk.

Bottom Line

Danksharding is Ethereum's plan to scale data availability for rollups by having nodes verify large data blobs are available through random sampling rather than full downloads, using erasure coding to make that sampling reliable. Proto-danksharding already delivered meaningful fee reductions for rollup users; full danksharding aims to extend that capacity much further. Track rollup activity and fee trends across layer 2 networks on MrDeFi's chains page.