# Ethereum Blob Space — How It Changed L2 Economics

When EIP-4844 ("Proto-Danksharding") activated in March 2024, it introduced a new transaction type: blob transactions. Two years later, the impact is clear — Layer 2 rollups are 95% cheaper, Ethereum L1 is less congested, and the blob gas market has created a new economic dynamic.

What Are Blobs?

Blobs (Binary Large Objects) are data chunks attached to Ethereum blocks that:

Store up to 128 KB of data per blob (6 blobs per block max)

Are cheaper than calldata (~10x reduction in cost)

Are pruned after ~18 days (not stored forever like calldata)

Have a separate gas market (blob gas vs execution gas)

Why blobs matter for L2s:

Rollups post transaction data to Ethereum L1 for security. Before blobs, this data went into calldata (expensive). Now it goes into blobs (cheap).

The Economic Impact

Transaction Cost Comparison (Pre vs Post-Blobs)

Arbitrum One (example rollup):

|--------|-------------------------|---------------------------|-----------|

| Cost to post 128KB | ~$50,000 | ~$500 | 99% |

| Average L2 tx fee | $0.80 | $0.03 | 96.25% |

| Transactions per L1 blob | ~500 | ~5,000 | 10x |

Result: L2s went from expensive to nearly free.

Blob Gas Market Dynamics

Blob gas uses the same EIP-1559 fee mechanism as execution gas, but with its own base fee.

Current blob gas market (March 2026):

Target blobs per block: 6 (increased from 3 in Pectra upgrade)

Max blobs per block: 9

Base fee adjustment: ±12.5% per block (same as execution gas)

Blob base fee over time:

March 2024: ~1-5 gwei (cheap, low usage)

July 2025: Spike to 200 gwei (usage surge)

March 2026: Stabilized at 10-30 gwei

Why the July 2025 spike?

Multiple new L2s launched simultaneously (Base, Linea, Scroll all ramping up). Demand for blob space exceeded supply for ~2 weeks. Blob base fee spiked 40x.

Lesson learned: Blob space is still limited. During high demand, costs can spike.

L2 Revenue Model Shift

Before blobs, L2s operated on thin margins. Now they're highly profitable.

Arbitrum revenue breakdown (Q1 2026):

Revenue from users: $24M (tx fees)

Cost to post to L1: $1.2M (blob fees)

Profit margin: 95%

Impact: L2 sequencers are printing money. This enables aggressive user acquisition (gas subsidies, airdrops).

Technical Deep Dive

Blob Transaction Structure

// Blob transaction (type 0x03)
struct BlobTx {
    uint256 chainId;
    uint256 nonce;
    uint256 maxPriorityFeePerGas;
    uint256 maxFeePerGas;
    uint256 gas;
    address to;
    uint256 value;
    bytes data;
    bytes[] blobVersionedHashes; // KZG commitments to blobs
    uint256 maxFeePerBlobGas;
    bytes signature;
}

Key field: maxFeePerBlobGas — max willing to pay for blob gas.

How L2s Use Blobs

Batch transactions: L2 sequencer collects 1000s of user txs

Compress data: Use zlib/gzip to compress tx data

Create blob: Package compressed data into 128KB blob

Post to L1: Submit blob tx to Ethereum L1

Verify: Ethereum nodes verify blob commitments (KZG proofs)

Prune: After ~18 days, blob data is pruned (but commitment stays)

Example: Optimism batching

// Sequencer batches L2 transactions
const batch = compressTransactions(l2Txs); // ~100KB compressed

// Create blob transaction
const blobTx = {
    to: L1_BATCH_INBOX,
    data: encodeL2Batch(batch),
    blobs: [batch], // Actual blob data
    maxFeePerBlobGas: 50n * 10n**9n, // 50 gwei max
};

// Submit to L1
await l1Signer.sendTransaction(blobTx);

Blob Gas Estimation

Blob gas price is independent of execution gas. To estimate total cost:

// Execution gas (for the transaction itself)
const executionGasCost = gasUsed * (baseFee + priorityFee);

// Blob gas (for blob data)
const blobGasCost = blobCount * 131072 * blobBaseFee; // 131072 gas per blob

// Total cost
const totalCost = executionGasCost + blobGasCost;

Current typical costs (March 2026):

Execution gas: ~$0.50 (21000 gas at 25 gwei ETH)

Blob gas (1 blob): ~$0.30 (131072 gas at 20 gwei)

Total: ~$0.80 to post 128KB to L1

Compare to pre-blobs: ~$50,000 for 128KB. 62,500x cheaper.

Impact on Ethereum L1

Reduced Congestion

Before blobs, L2s competed for block space with regular transactions. A large L2 batch could fill an entire block.

Result: High gas prices, congested network.

After blobs: L2 data goes into separate blob space. No competition with regular txs.

Evidence:

Average L1 gas price: 15 gwei (March 2026) vs 40 gwei (March 2024)

Block fullness: 60% (down from 95% pre-blobs)

Ethereum Revenue Shift

Ethereum's revenue comes from burned fees (EIP-1559). Pre-blobs, L2s were a major source of burn. Post-blobs, their contribution dropped.

Ethereum daily revenue (from fees):

March 2024: ~$15M/day (L2s contributed ~30%)

March 2026: ~$8M/day (L2s contribute ~5%)

Why the drop?

L2s pay 99% less for data posting

More activity moved to L2s (less on L1)

Long-term implication: Ethereum L1 becomes a settlement layer, not an execution layer. Revenue comes from blob gas + L1 DeFi, not from user transactions.

Pectra Upgrade — More Blob Space

The Pectra upgrade (March 2026) increased blob targets:

Target blobs per block: 3 → 6

Max blobs per block: 6 → 9

Impact:

2x more blob space

Lower blob base fees (more supply)

L2s can post more frequently (better UX)

But: Future upgrades (Danksharding) will increase this to 64 blobs per block, then eventually to full sharding.

Developer Implications

Building on L2s

Before blobs: L2s were expensive for high-throughput apps (games, social).

After blobs: L2s can handle millions of transactions per day at negligible cost.

Recommendation: If you're building a consumer app (gaming, social, payments), default to L2. There's no reason to be on L1 anymore.

Choosing an L2

All major L2s use blobs now:

Arbitrum

Optimism (+ Base)

zkSync

Starknet

Polygon zkEVM

Scroll

Linea

Key differences:

|----|------|----------|------|----------|

Blob Transaction Libraries

ethers.js v6+ supports blob transactions:

import { ethers } from 'ethers';

const provider = new ethers.JsonRpcProvider('https://mainnet.infura.io');
const wallet = new ethers.Wallet(privateKey, provider);

// Create blob transaction
const blobTx = await wallet.sendTransaction({
    to: recipient,
    value: ethers.parseEther('0.1'),
    blobs: [blobData], // Uint8Array (max 128KB)
    maxFeePerBlobGas: ethers.parseUnits('50', 'gwei'),
});

Note: Most developers don't send blob txs directly. L2 sequencers handle this.

The Road Ahead

Full Danksharding (2027+)

Current blobs are "Proto-Danksharding" (EIP-4844). The endgame is full Danksharding (EIP-4844 was just phase 1).

Future upgrades:

2027: 64 blobs per block (10x increase)

2028+: Full sharding (1000x increase)

Result: L2s will have effectively unlimited data availability at near-zero cost.

Data Availability Layers

Some projects are building alternatives to Ethereum blobs:

Celestia: Dedicated DA layer (modular blockchain)

EigenDA: DA using EigenLayer restaking

Avail: Standalone DA chain

Trade-off: Cheaper DA, but less secure than Ethereum.

Recommendation: For serious apps (DeFi, RWAs), stick with Ethereum blobs. For experiments (games, social), alt-DA is fine.

Resources

EIP-4844 Specification

Blobscan — Blob transaction explorer

Solingo L2 Development Course — Build on Arbitrum/Optimism

L2Beat — L2 metrics and comparisons

Conclusion

Blobs transformed Ethereum's economics. L2s went from expensive experiments to the default choice for developers. Ethereum L1 is becoming a settlement layer, optimized for security and decentralization, not throughput.

For developers:

Default to L2s for new projects

Understand blob gas market if building infra

Prepare for full Danksharding — data will be nearly free

The era of expensive on-chain transactions is over. Blobs made it happen.

Ethereum Blob Space — How It Changed L2 Economics