Calldata is the bottleneck. Before EIP-4844, rollups like Arbitrum and Optimism posted transaction data directly to Ethereum's expensive calldata, which constituted over 90% of their operating costs.
the-ethereum-roadmap-merge-surge-verge
Blog
EIP-4844 and the End of Calldata Bottlenecks
EIP-4844 (Proto-Danksharding) introduces blob-carrying transactions, decimating L2 costs by moving data off the expensive calldata path. This is the foundational step for Danksharding and the key unlock for the Surge.
introduction
THE DATA BOTTLENECK
Introduction
EIP-4844 redefines Ethereum's data economics by introducing a dedicated, low-cost data channel, directly addressing the primary cost driver for rollups.
Proto-danksharding is the fix. This upgrade introduces blob-carrying transactions, creating a separate, ephemeral data market that decouples data availability from execution gas fees.
The impact is immediate. Rollups now post data via blobs for ~$0.01 instead of calldata for ~$1.00, reducing L2 transaction fees by an order of magnitude and enabling new data-intensive applications.
Evidence: Post-EIP-4844, Base and zkSync Era saw transaction fees drop by over 80%, validating the economic model shift from a monolithic to a modular data layer.
key-trends
EIP-4844 AND THE END OF CALLDATA BOTTLENECKS
The Calldata Crisis: Why Rollups Were Bleeding Money
Rollup scaling was hamstrung by the exorbitant cost of posting data to Ethereum. EIP-4844 introduced a dedicated, cheap data channel, solving the core economic inefficiency.
01
The $1M+ Per Day Problem
Before EIP-4844, rollups like Arbitrum and Optimism were forced to post all transaction data as expensive mainnet calldata. This was their single largest operational cost, consuming $1M+ daily across major L2s and creating a direct scaling tax.
- Cost Bottleneck: Data costs grew linearly with adoption, capping scalability.
- Fee Volatility: User fees spiked with mainnet congestion, breaking the L2 value proposition.
$1M+
Daily Burn
~80%
Cost Share
02
Blobs: The Dedicated Data Lane
EIP-4844 introduced blob-carrying transactions, creating a separate, high-capacity data market orthogonal to EVM execution. Blobs are large (~128KB), cheap, and automatically pruned after ~18 days.
- Separate Pricing: Decouples data costs from gas auction volatility.
- Purpose-Built: Optimized for rollup data, not smart contract execution.
~128KB
Per Blob
100x
Cheaper Data
03
The New L2 Economic Model
With data costs reduced by ~100x, the economic model for rollups fundamentally changed. Fees are now dominated by L2 execution, enabling sustainable micro-transactions and new use cases.
- Profitability Shift: L2 sequencers transition from data-posting entities to execution profit centers.
- Fee Predictability: User experience improves with stable, ultra-low transaction costs.
< $0.01
Avg. L2 Tx
100x
Cheaper Data
04
The Full Danksharding Path
EIP-4844 (Proto-Danksharding) is not the endgame. Full Danksharding will scale blob capacity to ~1.3 MB per slot, supporting 100+ rollups concurrently through data availability sampling.
- Horizontal Scaling: More blobs per block enable an ecosystem of specialized rollups.
- Secure Scaling: Data availability sampling allows light nodes to verify terabytes of data.
~1.3 MB
Per Slot Target
100+
Parallel Rollups
deep-dive
THE DATA PIPELINE
Blob Mechanics: How EIP-4844 Cuts the Cord
EIP-4844 introduces a dedicated data channel for rollups, decoupling execution from data availability to slash costs and scale.
Calldata is a bottleneck. It forces rollups like Arbitrum and Optimism to compete with every L1 transaction for expensive, permanent storage on Ethereum's execution layer.
Blobs are ephemeral data. They provide a separate, low-cost data channel for 128KB data packets, which nodes discard after ~18 days, matching the fraud proof window.
This is a pricing revolution. Blob fees are priced by a separate EIP-1559 mechanism, isolating rollup data costs from volatile mainnet gas wars, enabling predictable, ultra-low fees.
Evidence: Post-Dencun, L2 transaction fees collapsed. Base and zkSync Era saw fees drop 99%, from dollars to fractions of a cent, proving the model works.
CALLDATA COST COMPARISON
Before & After 4844: The Fee Impact Matrix
Quantifying the direct impact of EIP-4844 on transaction costs for key L2 operations, replacing historical calldata with ephemeral blobs.
| Cost Metric / Feature | Pre-4844 (Calldata) | Post-4844 (Blobs) | Impact |
|---|---|---|---|
Dominant Cost Component | L1 Calldata Gas | L1 Blob Gas | New market |
Typical Cost per Byte | 16 gas (4 gas after discount) | ~0.125 gas (1 blob ~= 128k gas) | ~97% reduction |
Data Availability Guarantee | Permanent on-chain | Ephemeral (~18 days), then pruned | Trade-off for cost |
L2 Batch Submission Cost (Example) | $2,000 - $10,000 (high volatility) | $200 - $800 (stable, low) | ~90% cost reduction |
Throughput Limit per Block | ~80 KB (calldata limit) | ~768 KB (6 blobs * 128 KB) | ~9.6x increase |
Supports ZK Proof Data | Inefficient, high cost | Native, optimal for large proofs | Enables ZK-Rollup scaling |
Fee Market Coupling | Tightly coupled with L1 execution | Decoupled via separate blob gas fee market | Predictable L2 fees |
counter-argument
THE DATA BLOB REALITY
The Skeptic's Take: Is This Just Kicking the Can?
EIP-4844's data blobs are a temporary fix that shifts, rather than solves, the core data availability bottleneck.
Blobs are a capacity upgrade, not a solution. EIP-4844 introduces a new, cheaper data lane for rollups, but it's still finite. The data availability bottleneck moves from calldata to a new, albeit larger, blob market. This is a classic scaling strategy: increase the pipe diameter until the next constraint emerges.
The fee market will re-emerge. As adoption by Arbitrum, Optimism, and Base grows, blob demand will saturate the fixed per-block supply. Blob gas fees will spike during congestion, recreating the cost volatility EIP-4844 was meant to solve for rollups. The can is kicked further down the road, not into a ditch.
Evidence: Ethereum's current target is 3 blobs per block (0.375 MB). Full data sharding (Danksharding) proposes 64 blobs (8 MB), a 17x increase that remains years away. Until then, blob supply is a hard cap that rollup growth will inevitably hit.
takeaways
EIP-4844 AND THE END OF CALLDATA BOTTLENECKS
TL;DR for Builders and Investors
EIP-4844 (Proto-Danksharding) is a fundamental upgrade that decouples L2 scaling from mainnet execution by introducing a new, cheap data layer.
01
The Problem: L2s Are Paying for Mainnet's Expensive Storage
Rollups like Arbitrum and Optimism were forced to post data to Ethereum's calldata, a premium-priced resource. This created a hard floor for L2 transaction costs and limited throughput.
- Cost Bottleneck: Up to 80-90% of an L2's operational cost was data posting.
- Throughput Ceiling: Congestion on L1 directly limited L2 finality and capacity.
80-90%
Cost Was Data
~10x
Calldata Premium
02
The Solution: Blobs as a Dedicated Data Commodity
EIP-4844 introduces blob-carrying transactions, a separate data lane with a distinct fee market. Blobs are large (~128KB), cheap, and automatically pruned after ~18 days.
- Cost Reduction: L2 data posting costs drop by >100x vs. calldata.
- Throughput Unlocked: Enables ~100k+ TPS across the L2 ecosystem without congesting L1 execution.
>100x
Cheaper Data
~100k+
Ecosystem TPS
03
The New Frontier: Hyper-Scalable App-Specific Chains
With data costs near-zero, the economic model for app-specific rollups (via Arbitrum Orbit, OP Stack, zkSync Hyperchains) becomes viable. This enables new design spaces previously crushed by overhead.
- Micro-Economics: Viable for games, social, and DeFi with <$0.001 per tx.
- Architectural Shift: Moves competition from shared L2s to sovereign execution environments.
<$0.001
Per Tx Goal
1000s
Viable Rollups
04
The Investment Thesis: Infrastructure for the Blob Economy
The next wave of infrastructure winners will be those that optimize the blob lifecycle: sequencers (Espresso, Astria), DA layers (Celestia, EigenDA), and bridges (LayerZero, Across) that leverage cheap data for cross-chain intents.
- New Stack: Creates demand for modular DA and shared sequencers.
- Market Expansion: Enables mass-market dApps that were previously impossible.
New Stack
Modular DA
Mass-Market
dApp Viability
ENQUIRY
Get In Touch
today.
Our experts will offer a free quote and a 30min call to discuss your project.
NDA Protected
Confidentiality24h Response
Time to ValueDirectly to Engineering Team
No Sales Fluff10+
Protocols Shipped
$20M+
TVL Overall