Blob Transactions on Ethereum Explained
Master blob transactions (EIP-4844): Type-3 transactions, 128KB ephemeral blobs, KZG commitments, and blob gas pricing. Understand how blobs reduced L2 fees from $0.50 to $0.01 and the roadmap to full Danksharding.
What are Blob Transactions?
Blob transactions are Type-3 Ethereum transactions introduced in EIP-4844 (Dencun upgrade, March 2024). A blob is a 128KB chunk of data carried in a transaction but kept separate from standard calldata. Blobs are: (1) Included in blocks and verified by consensus. (2) Stored by full nodes for ~18 days. (3) Pruned from validator storage after 18 days. (4) Accessed only by data availability sampling (future feature), not by Ethereum state execution.
Understanding this concept is a prerequisite for making informed decisions in DeFi. Most losses in crypto come from misunderstanding the fundamentals.
Blobs solve the "data availability problem" for rollups. L2 rollups post transaction data to Ethereum to enable full node recovery. Calldata was the only option pre-blobs, costing 16 gas/byte. Blobs cost 0.125 gas/byte, enabling 128x cost reduction for L2s. Ethereum mainnet benefits by increasing throughput (6 blobs/block vs ~130KB calldata) while maintaining node storage requirements (blobs auto-prune).
Blob Transaction Structure
Type-3 Transaction Format
Type-3 transactions extend Type-2 (EIP-1559) with blob-specific fields: (1) blob_versioned_hashes: array of KZG commitments (one per blob). (2) max_fee_per_blob_gas: price user willing to pay for blob space. (3) blob_data: actual 128KB data (included in blob sidecar, not main tx). Unlike calldata (part of tx), blobs are in a separate gossip channel for efficiency.
Blob Sidecar & Gossip
Blobs are transmitted separately as "blob sidecars" in the consensus layer P2P network. When proposer includes blob commitment in block, beacon nodes gossip blob sidecar independently. This prevents large blobs from bloating the main tx mempool. Nodes can ignore sidecars (if storage-constrained) and still validate block (via commitment).
Field Elements & Encoding
Blob data is encoded as 4,096 field elements (each ~32 bytes in BLS12-381 field). This enables polynomial commitments: KZG commitment (48 bytes) proves the polynomial (blob data). Field encoding ensures cryptographic operations (commitments, proofs) are efficient. Users see blobs as 128KB byte arrays; internals are field elements.
Transaction Inclusion & Block Space
Max 6 blobs per block (consensus rule). Each blob takes ~768KB of block space (externally). Blobs don't count toward execution gas limit (separate blob gas limit). This means 1 normal tx (21k gas) + 6 blobs can fit in same block, maximizing throughput. Validators are incentivized to include all 6 blobs (via fees) regardless of tx demand.
Calldata vs Blobs Comparison
| Aspect | Calldata | Blobs | Advantage |
|---|---|---|---|
| Cost Per Byte | 16 gas/byte (~$0.32/KB) | 0.125 gas/byte (~$0.003/KB) | 128x cheaper |
| Storage Duration | Permanent (all nodes, forever) | 18 days (~325 slots) | Reduces node burden |
| Access Pattern | Fully indexed, queryable | Sampled, verification-only | Data availability focus |
| Block Impact | Counts toward gas limit | Separate blob gas limit | No execution bottleneck |
| Max Per Block | ~130KB (gas-limited) | ~768KB (6 blobs max) | 5.9x more throughput |
| Use Case | Permanent state, execution | Temporary rollup data | Purpose-built layers |
Blob Gas Pricing & Market
Dynamic Blob Gas Mechanism
Blob gas uses EIP-1559 dynamic pricing: base_blob_fee increases if avg_blobs_used > 3, decreases if < 3. Target is 3 blobs per block. Starting base fee: 1 wei per blob. With 6 blobs per block over time, fee increases exponentially (doubling every ~50 blocks if demand stays high). During low demand, fees decay toward 1 wei.
Practical Cost Examples
Low congestion (1-2 blobs/block): base blob fee ~1-10 wei. Cost per blob: ~0.001-0.01 LINK (~$0.0001-0.002). One blob = one L2 batch = 50k transactions. Per-tx cost: $0.000002-0.00004. High congestion (6 blobs/block): base blob fee ~10000 wei. Cost per blob: ~0.10-0.50 LINK (~$2-10). Per-tx cost: $0.00004-0.0002. Arbitrum batches 100 txs per blob, so: low demand = $0.001-0.02 per tx, high demand = $0.02-0.10 per tx.
Blob Space Economics
Daily blob throughput: 6 blobs/block * 7200 blocks/day * 128KB/blob = 5.5 GB/day. At low fees: $0.001/blob * 43,200 blobs/day = $43k/day MEV revenue. At high fees: $10/blob * 43,200 = $432k/day. Validators capture blob fees. Proposers select blobs based on fee bids; high bidders get included first.
Sequencer Impact & L2 Strategy
L2 sequencers post blobs at their own cost. During high blob congestion, sequencers may delay posting (wait for fees to drop) or batch aggressively (compress more txs per blob). Users experience slower finality but cheaper costs. This creates L2-level MEV: sequencers can batch favorably to high-fee txs, delay low-fee txs. Competition between sequencers incentivizes fair batching.
KZG Commitments & Verification
KZG in Action
When a blob is included, proposer includes KZG commitment (48 bytes) in block header. Consensus nodes verify: does this blob's polynomial hash match the commitment? Verification is ~10ms per blob, parallelizable. 6 blobs = 60ms total, easily within 12-second slot time. Verification proves blob wasn't tampered with.
Light Client Verification
Light clients (no full nodes) can verify blobs using commitments alone. Download commitment (48 bytes), run KZG verification (~10ms), confirm blob validity. No need to download full 128KB blob. This enables: (1) Mobile wallets verifying L2 blocks. (2) Cross-chain bridges proving Ethereum data. (3) Bridges with minimal trust (only commitment hash, no full data).
Trusted Setup & Security
KZG requires a trusted setup ceremony (parameters generated once, used forever). Ethereum's ceremony (Nov 2022-April 2023): 140k+ participants. If ceremony was compromised (all participants colluded), KZG could be forged. Practically impossible: would require coordinating 140k people globally, impossible without detection. KZG is considered cryptographically secure for Ethereum's use case.
Future: Quantum-Resistant Alternatives
Quantum computers could break KZG (ECDLP is vulnerable). Long-term: lattice-based commitments (MLWE, LWE) are post-quantum. Ethereum research explores migration paths, but not urgent (quantum threat is 10-20+ years away). Current KZG is secure against all known attacks.
Blob Space Economics
L2 Scaling Unlock
Pre-blobs: L2 fees were mostly DA costs (80-90% of fee). Arbitrum $0.50 tx = $0.40 DA + $0.10 compute. Post-blobs: DA cost dropped 95%, compute now matters. Arbitrum $0.01 tx = $0.001 DA + $0.009 compute. This shift enables: (1) Smaller tx economical. (2) More complex smart contracts worth gas. (3) DeFi, gaming, social apps all economical.
MEV & Sequencer Economics
With cheap blobs, sequencer incentives shift. Pre-blobs: sequencer profited from MEV (reordering) + blob batching. Post-blobs: batching is cheap, MEV is main profit. Sequencers now compete on fairness (fair ordering) rather than batching efficiency. Some L2s (Arbitrum, Base) use sequencer leasing: user pays flat fee, sequencer guarantees inclusion within N blocks. Blobs enable this model economically.
Storage & Node Costs
Validators save ~270GB/year from blob pruning. Without pruning: 100+ nodes would cost extra storage. With pruning: negligible cost for most operators. This enables: (1) More validators participate (lower hardware requirements). (2) Ethereum remains decentralized (no storage bottleneck). (3) Archive nodes optionally retain blobs for historical analysis (costly but possible).
Inscription Competition
Blobs enable cheap inscriptions (Bitcoin Ordinals on Ethereum). Blobscriptions cost 128x less than calldata inscriptions. Competition for blob space is now 2-way: L2s vs inscription users. During high blob demand, either L2 batch times increase or inscription costs rise. Long-term: full Danksharding increases blob space, keeping fees low for both.
Blobscriptions & Applications
What Are Blobscriptions?
Blobscriptions are inscriptions stored in blobs (similar to Bitcoin Ordinals on calldata). Users inscribe images, text, code, and metadata into blobs. Cost: ~0.001-0.01 LINK ($0.03-0.30) per inscription vs calldata inscriptions ($3-30). 128x cheaper enables: (1) Art/NFT projects storing metadata. (2) Domain name registries (Blob Name Service). (3) Decentralized storage prototypes (Ethereum as DA layer).
Technical Implementation
Smart contract requests blob submission with data payload. User submits Type-3 tx with blob containing metadata. Contract indexes blob commitment. Users retrieve data by querying commitment off-chain (indexer service). On-chain, only commitment is stored (48 bytes). Data itself is in blob (temporary), but commitment is permanent.
Use Cases & Adoption
Blob Name Service (BNS): register domains, store metadata in blobs. Per-domain cost: $0.50-2 vs ENS ($5-50+). ERC-404 NFTs: store trait data in blobs instead of IPFS. Cost: 1000x cheaper than pinning IPFS. Ethereum data availability protocol: rollups use blobs natively, builders use blobs for state proofs. Adoption growing as more projects realize blob cost advantage.
Limitations & Trade-offs
Data availability window: 18 days. After 18 days, inscribed data is no longer retrievable from consensus layer (only via archive nodes). This is fine for most use cases (metadata retrieval happens within 18 days), but not suitable for permanent storage. Long-term: users should archive blob data off-chain (IPFS, centralized storage) for permanent availability.
FAQ
Can I execute smart contracts with blob data?
No. Blobs are data-only, not execution. Smart contracts cannot directly read blob contents (data is not in state, only KZG commitment is). Blobs are designed for L2 data availability, not Ethereum state mutation. If you need to execute logic based on blob data, you must either: (1) Post data to calldata (expensive). (2) Post KZG commitment, have contract trust off-chain data (centralized). (3) Wait for future DA-integrated designs.
What if all 6 blob slots are full?
Users (L2s, inscription services) bid for blob slots. Highest bidders get included. Lower bidders either: (1) Wait for next block. (2) Increase bid. (3) Fall back to calldata (expensive). During extreme congestion (unlikely until adoption massively increases), blob fees can spike to $1-10 per blob, making blobs expensive again. Full Danksharding will increase blob slots, preventing this.
Can I retrieve blob data after 18 days?
No, from consensus nodes (they prune). Archive nodes can retain blobs indefinitely (optional, expensive). If you need permanent access, solutions: (1) Archive blob data off-chain (IPFS, S3, Arweave). (2) Run archival Ethereum node (~2TB/year extra storage). (3) Use calldata for permanent storage (expensive). For L2s, this is fine: L2 nodes download blobs within 18 days, state is finalized locally.
How do L2 light clients verify blob data?
L2 light clients download L1 block headers, extract blob commitments (KZG hashes). To verify a blob, they either: (1) Download full blob from L2 sequencer, verify commitment matches. (2) Download random samples of blob (Blob Availability Sampling, future feature), verify samples match commitment. BAS enables light clients to prove blob availability without full download.
Are blobs compatible with smart contracts?
Blobs don't exist in smart contract state, so contracts can't directly read or manipulate blobs. However, contracts can: (1) Store KZG commitments (via events or storage). (2) Verify proofs of blob inclusion (via Verkle tree proofs, future). (3) Accept blob commitments as valid state proofs from L2s. Blobs are meant for consensus layer (L1) + L2 data, not contract execution.
When will full Danksharding arrive?
Proto-Danksharding (EIP-4844) is live (March 2024). 2D Danksharding (intermediate step, more blob dimensions) planned 2025-2026. Full Danksharding (64+ shards, 16MB data per block) planned 2027+. Timelines are research-driven and may slip. Expect incremental improvements: more blobs per block, BAS for light clients, before full Danksharding.
Educational disclaimer: This guide is for informational purposes only and does not constitute financial advice. Crypto involves significant risk — do your own research before making any decisions. Learn more about our team.
Educational disclaimer: This guide is for informational purposes only and does not constitute financial advice. Crypto involves significant risk — do your own research before making any decisions. Learn more about our team.