What Is EIP-4844? Proto-Danksharding and Blob Transactions Explained

EIP-4844, also known as proto-danksharding, was an Ethereum upgrade activated in March 2024 (as part of the Dencun upgrade) that introduced a new, cheaper type of transaction data called "blobs," specifically designed to reduce the cost that Layer 2 rollups pay to post their transaction data to Ethereum, and it produced one of the largest single drops in L2 fees to date.

The problem it solved

Before EIP-4844, rollups published their batched transaction data as regular Ethereum "calldata" — the same general-purpose data field any Ethereum transaction can use. Calldata is priced per byte and stored permanently as part of Ethereum's history, which made it relatively expensive, especially as more L2s scaled up and competed for that same limited, permanently-stored capacity. Since rollup data doesn't actually need to live on Ethereum forever — it just needs to be available long enough for anyone to verify or challenge a state transition — the ecosystem was effectively paying for permanent storage it didn't need.

What blobs actually are

EIP-4844 introduced a new transaction type that can carry "blobs" — large chunks of data (roughly 128 KB each) that are priced and processed separately from regular calldata, using their own fee market. Critically, blobs are not stored permanently by Ethereum execution clients — they're retained for a limited window (roughly 18 days) and then pruned, since that's enough time for anyone who needs to verify or dispute a rollup's data to do so. This is the core insight behind proto-danksharding: rollups don't need forever-storage, they need temporary, guaranteed availability, and pricing that separately means blob demand doesn't compete directly with regular transaction gas demand.

Why "proto" and what full danksharding would add

EIP-4844 is called "proto-danksharding" because it's an intermediate step toward a more ambitious future upgrade, full danksharding, which aims to scale the number of blobs Ethereum can support per block substantially further, distributing the burden of storing and verifying that data across the validator set more efficiently. Proto-danksharding introduced the blob transaction format and fee market first, without yet implementing the full data-sharding scaling mechanism, letting the ecosystem adopt the new format and start benefiting from it immediately while the harder scaling work continues.

The effect on Layer 2 fees

	Before EIP-4844	After EIP-4844
Data type used	Regular calldata	Blobs
Storage duration	Permanent	~18 days, then pruned
Fee market	Shared with all other Ethereum transactions	Separate blob fee market
Effect on L2 fees	Higher L1 data-posting cost passed to L2 users	Substantially lower L1 data-posting cost

Because the L1 data-posting cost is one of the two main components of any L2 transaction fee (the other being the L2's own execution fee, covered in understanding L2 gas fees), this upgrade directly lowered the cost of using nearly every major rollup, contributing significantly to why gas fees on L2 are so much lower than L1 today.

Why this mattered for Ethereum's broader scaling strategy

EIP-4844 is a direct expression of Ethereum's rollup-centric roadmap — rather than scaling Ethereum's base layer execution directly, the strategy is to make Ethereum an excellent, cheap settlement and data availability layer for rollups, and let those rollups handle the actual transaction execution and scaling. Proto-danksharding is arguably the most consequential single step taken toward that goal so far, since it directly targeted the single biggest cost driver for every rollup built on Ethereum.

What this means for you as a user

You don't interact with blobs directly — this is infrastructure-level plumbing handled entirely by rollup operators and Ethereum's protocol layer. What you experience is simply lower fees on the L2s you already use, and the effect showed up across essentially every major rollup simultaneously when the upgrade activated, since all of them share Ethereum as their common data availability layer.

What comes after proto-danksharding

Full danksharding, the eventual successor to this upgrade, aims to increase the number of blobs each Ethereum block can carry considerably further, alongside changes to how validators verify blob data without needing to download and store all of it themselves — a technique called data availability sampling. That combination is intended to let Ethereum support a much larger number of active rollups simultaneously without a proportional increase in the burden placed on any individual node operator. Full danksharding remains a multi-stage undertaking, and proto-danksharding was deliberately scoped as the piece that could be shipped soonest while delivering most of the near-term fee benefit to rollup users.

Bottom line

EIP-4844 (proto-danksharding) introduced blob transactions — cheaper, temporary data storage specifically designed for rollups — and it meaningfully reduced the cost of using nearly every major Ethereum Layer 2. It's an intermediate step toward full danksharding, part of Ethereum's broader strategy of scaling through rollups rather than by making the base layer itself execute more transactions directly.