The Future of Ethereum: Sharding, Danksharding, and EIP-4844 Explained

By
Bakhodir
-
0
106

Ethereum has evolved from an expensive, congested smart-contract chain into the settlement layer for an expanding Layer-2 (L2) ecosystem. However, base-layer throughput remains limited, and fees spike during periods of high demand. The roadmap to address these challenges centers on three key innovations: sharding (reframed as data sharding), Danksharding, and the already-implemented EIP-4844 (Proto-Danksharding). Together, these upgrades aim to make L2 transactions significantly cheaper and scale Ethereum to support global usage while maintaining decentralization.

TL;DR: Ethereum’s Scaling Milestones

EIP-4844, deployed through the Dencun upgrade on March 13, 2024, introduced blobs—affordable, temporary data storage for rollups—resulting in L2 fee reductions of approximately 50-99% shortly after launch. The Pectra upgrade, activated on May 7, 2025, increased blob throughput by raising the target from 3 to 6 blobs per block and the maximum from 6 to 9, further expanding L2 data capacity.

Looking ahead, Danksharding will scale from current proto-danksharding limits to 64 blobs per block, with proposer-builder separation (PBS) and data availability sampling (DAS) as essential prerequisites, massively boosting L2 capacity. PeerDAS (EIP-7594), part of the upcoming Fusaka upgrade, enables nodes to sample blob data rather than downloading all of it and introduces staged “BPO” (Blob-Parameter-Only) blob increases, with testnets already scheduled.

The long-term target outlined in Vitalik’s roadmap is approximately 100,000+ transactions per second (TPS) across L2s and L1 combined, though this is a goal rather than a guarantee.

Understanding Ethereum’s Shift: From Execution Sharding to Data Sharding

Early “Eth2” plans envisioned multiple execution shards, each capable of running smart contracts. With rollups proving their effectiveness, Ethereum pivoted its strategy: concentrate execution on L2s while optimizing L1 for data availability. In simpler terms, L2s handle the computational heavy lifting while L1 guarantees that data is published so anyone can verify it. This represents the core principle of Ethereum’s rollup-centric roadmap.

Why this matters: L2 solutions need to post batches of transaction data to L1. When posting data is inexpensive and abundant, L2 fees remain low and throughput scales effectively. This approach differs from other blockchain architectures, such as NEAR Protocol’s sharding implementation, which focuses on execution sharding across multiple chains.

EIP-4844 (Proto-Danksharding): The Blob Revolution

Shipped through the Dencun upgrade on March 13, 2024, EIP-4844 introduced a new transaction type that carries blobs—opaque data chunks utilized by rollups. These blobs exist on the consensus layer temporarily (approximately 4096 epochs, or roughly 18 days), are committed using KZG cryptography, and are priced in a separate fee market to avoid competing with regular gas. Typical blob size is approximately 128 KiB, with initial limits setting a target of 3 blobs per block and a maximum of 6.

Impact: Following the Dencun upgrade, numerous L2s reported fee reductions ranging from 50-99%, making activities like swaps and mints dramatically more affordable, as reported by The Defiant and other industry sources.

Interesting fact: The KZG ceremony that secured blob proofs attracted approximately 141,000 contributions, making it the largest ceremony of its kind.

Pectra Upgrade (May 7, 2025): Expanding Blob Capacity

The Pectra upgrade didn’t fundamentally redesign the system but rather fine-tuned key parameters to enable further L2 scaling. Through EIP-7691, Ethereum doubled the target blob count from 3 to 6 and increased the maximum from 6 to 9 per block, representing a significant expansion in daily L2 data capacity and providing relief for blob pricing pressures.

To be precise: the target is 6 blobs with a maximum of 9, compared to the pre-Pectra configuration of target 3 and maximum 6, as detailed in the EIP-7691 specification.

While Pectra included other enhancements such as EIP-7702 for smart-account style authorizations, the blob increase represents the primary scaling improvement.

Danksharding: The Ultimate Scaling Solution

Danksharding extends proto-danksharding and dramatically scales blobs—from 6 (proto limits) to 64 blobs per block—while maintaining node requirements light enough for widespread participation. Two critical components enable this:

PBS (proposer-builder separation): Specialized builders assemble blocks and proposers select them, reducing MEV centralization pressure and enabling larger blocks safely. Learn more about PBS on ethresear.ch.

DAS (data availability sampling): Nodes randomly sample small pieces of blob data for verification instead of downloading everything, preserving security while dramatically reducing bandwidth requirements. This is similar in concept to how modular blockchains like Celestia approach data availability.

With 64 blobs at approximately 128 KiB each, this translates to roughly 8 MiB of blob data per block available for rollups, though exact numbers may evolve as outlined in Ethereum’s roadmap documentation.

PeerDAS & Fusaka: The Next Implementation Phase

PeerDAS (EIP-7594) represents the practical DAS networking upgrade arriving with the Fusaka release. It distributes blob data custody across subnets and enables regular nodes to verify availability by sampling columns rather than full blobs, delivering an order-of-magnitude data availability boost without requiring every node to operate as a datacenter.

Importantly, Ethereum will increase blob counts gradually using BPO (Blob-Parameter-Only) forks to avoid overwhelming the peer-to-peer layer. Testnets have scheduled BPOs that incrementally step the target and maximum beyond 6/9 in measured stages (for example, 10/15 then 14/21) before pursuing much higher targets, as detailed in Ethereum Foundation blog posts.

Practical Implications for Users and Developers

More affordable L2s today: Thanks to EIP-4844 and Pectra, rollups now have access to more (and cheaper) blob space. Users can expect fees in the cents range under normal conditions, sometimes even lower.

Expanded capacity tomorrow: PeerDAS combined with staged blob increases will expand capacity further, with full Danksharding pushing capabilities significantly beyond current limits. This infrastructure may also enable Layer 3 solutions to flourish.

Long-term ambition: Vitalik’s publicly stated target is 100,000+ TPS across L2s and L1 combined—an ambitious goal contingent on continued protocol and L2 improvements including compression, proofs, and interoperability. This should be understood as an aspirational goal rather than a guaranteed outcome.

Correcting Common Misconceptions

Misconception #1: “Sharding means multiple execution shards”

That was the original plan. Today’s approach focuses on data sharding for rollups, where L2s handle execution while L1 ensures data availability. For a different take on sharding, see how NEAR Protocol implements execution sharding.

Misconception #2: “Pectra set blobs to 6”

More precisely, Pectra set the target to 6 and the maximum to 9 (previously 3 and 6). This distinction matters for accurate capacity calculations.

Misconception #3: “Full Danksharding is coming soon”

Danksharding is a multi-year effort. Proto-danksharding shipped in 2024, blob targets increased in 2025, and PeerDAS is next on the roadmap. The 64-blob full Danksharding implementation will arrive after additional groundwork, including in-protocol PBS and stable peer-to-peer networking.

The Bottom Line

Ethereum’s scaling strategy is deliberately iterative: deploy safe, immediately helpful solutions (EIP-4844), increase limits prudently (Pectra, BPOs), implement smarter networking (PeerDAS), and then execute the major leap (Danksharding). The outcome is an L1 that maintains credible decentralization while empowering L2s to deliver user experiences competitive with web-scale systems. This is how Ethereum transitions from a congested highway to an intelligent superhighway without sacrificing its foundational principles.

Key Resources and Further Reading

Post Views: 32