Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
What Is the Blob Market?
The blob market is Ethereum's hidden fee auction for L2 data. Most users never see it directly — but it controls how much Arbitrum, Base, and Optimism charge you.
The blob market is Ethereum’s separate, independent fee auction for the temporary data storage that Layer 2 rollups use to post transaction batches to the mainnet — introduced by EIP-4844 in March 2024.
Most Ethereum users never interact with the blob market directly. But if you use Arbitrum, Base, or Optimism, the blob market quietly determines a large part of what you pay in fees. When blob demand spikes — during a major airdrop, a token launch, or a surge in rollup activity — the cost of posting data to Ethereum rises, and that cost can pass through to you as higher L2 transaction fees. The blob market is the pricing engine sitting underneath all of Ethereum’s Layer 2 scaling.
Key takeaways
- The blob market is Ethereum’s separate fee auction for rollup data, created by EIP-4844 and running on its own pricing curve independent of regular gas fees.
- Blob fees can swing from fractions of a gwei to over 100 gwei in minutes when rollup demand spikes — the EIP-1559-style mechanism adjusts every single block.
- Three upgrades since launch — Pectra (May 2025), Fusaka (December 2025), and BPO2 (January 2026) — have progressively expanded blob capacity and added a minimum fee floor to prevent indefinite 1-wei collapse.
—
What the Blob Market Is
Before March 2024, rollups had to post their transaction data to Ethereum using calldata — permanent storage space that lives in the chain’s execution history forever. It worked, but it was expensive. Every byte cost real gas, and because calldata is permanent, validators, nodes, and anyone syncing the chain had to carry that data indefinitely.
EIP-4844 changed that. It introduced blobs: a new type of data attachment for Ethereum transactions, stored separately from the execution layer and automatically pruned after roughly 18 days. Because blobs are temporary and handled by consensus clients rather than the EVM, they can be priced much more cheaply. They get their own fee market — the blob market — that runs alongside Ethereum’s regular gas market but is completely independent from it.
The blob market sets the price of blobspace on a block-by-block basis. Rollup operators bid into this market when they submit a blob transaction. Regular users do not bid directly. Instead, the rollup absorbs the blob cost and factors it into the fees you pay on the L2 side. Think of it like a wholesale fuel market: the airline buys jet fuel at commodity prices, and you buy a ticket. You feel the price indirectly, but you are not at the commodity exchange yourself.
How Blob Market Transactions Work on Ethereum
A blob is a chunk of binary data, 128 KB in size, attached to a special type-3 transaction on Ethereum. Rollups use blob transactions to post compressed batches of their users’ activity — swaps, transfers, NFT mints — back to the L1 in bulk. The blob data itself is not executed by the Ethereum Virtual Machine. Ethereum’s consensus clients verify that the data was available, a KZG cryptographic commitment anchors that proof on-chain, and then the raw blob data is pruned from nodes after approximately 18 days.
That 18-day window is intentional. Rollups only need the data to be available long enough for fraud provers (in optimistic rollups) or validity provers (in ZK rollups) to do their work. Once the proofs are settled, the raw batch data no longer needs to live on every node forever. This is what makes blobs so much cheaper than calldata: Ethereum is not asking validators to store your data in perpetuity.
The distinction between blobs and calldata is fundamental. Here is how they compare:
| Characteristic | Blobs vs. Calldata |
|---|---|
| Storage location | Blobs: consensus layer, pruned after ~18 days. Calldata: execution layer, permanent. |
| EVM execution | Blobs: not executed, not accessible to smart contracts. Calldata: read by the EVM. |
| Pricing | Blobs: blob gas market (separate). Calldata: regular gas market. |
| Cost per byte | Blobs: significantly cheaper at moderate demand. Calldata: fixed per-byte cost in execution gas. |
| Data commitment | Blobs: KZG polynomial commitment proves availability. Calldata: no separate proof needed. |
The KZG commitment is the cryptographic glue that makes this work. Even after the raw blob is gone, the commitment remains on-chain. Validators can confirm the data was present when posted — which is all a hybrid rollup needs to verify its state transitions safely.
How Blob Market Pricing Works
The blob fee market works on the same logic as EIP-1559 — but for a completely separate resource bucket called blob gas. Each block has a target blob count. When the number of blobs in a block exceeds the target, the blob base fee rises by up to 12.5% for the next block. When blobs fall below target, the fee drops by the same proportion. The market self-adjusts every single block.
This is why blob fees can look almost mystical to outside observers. At low demand, fees fall toward their theoretical minimum — which was 1 wei for most of 2024 and into 2025, because the number of blobs being posted regularly stayed well below the 3-blob-per-block target set at the Dencun launch. The protocol kept cutting fees each block until they hit the floor. That floor, until Fusaka, was 1 wei: essentially free.
Then demand surges — a major protocol airdrop, a surge of new rollup activity, a popular NFT mint that floods multiple L2s simultaneously — and the blob base fee can climb from near-zero to over 100 gwei within minutes. Each block where demand exceeds target raises the next block’s fee by 12.5%. Stack a dozen of those blocks in a row and you have a 3x or 4x fee spike before most users even notice. For more detail on how the blob fee is calculated per transaction, CryptoProcent has a dedicated explainer.
A key technical detail: each blob transaction pays 131,072 blob gas units per blob. The type-3 transaction also pays regular execution gas in the normal ETH gas market, so a rollup’s total cost per submission has two components — one in each fee market.
The Pectra upgrade in May 2025 raised the per-block blob target from 3 to 6, effectively doubling the capacity headroom available to rollups. Fees fell again as demand grew into the new space. Then Fusaka in December 2025 introduced PeerDAS and the first Blob Parameter Only (BPO1) fork, which raised the blob target to 10 per block and the maximum to 15. BPO2 followed on January 7, 2026, raising the target to 14 and the maximum to 21.
Fusaka also introduced EIP-7918, a blob fee floor that prevents the base fee from collapsing to 1 wei indefinitely. Under EIP-7918, the minimum blob base fee is proportional to the L1 execution base fee. When Ethereum mainnet is active, blob fees can no longer fall below a meaningful minimum — addressing a key concern that near-zero blob fees were eroding ETH’s value accrual from fee burns.
The upgrades by milestone:
| Milestone | Blob Target / Max Per Block |
|---|---|
| Dencun / EIP-4844 (March 2024) | Target: 3 / Max: 6 |
| Pectra (May 2025) | Target: 6 / Max: 9 |
| Fusaka / BPO1 (December 2025) | Target: 10 / Max: 15 |
| BPO2 (January 7, 2026) | Target: 14 / Max: 21 |
How the Blob Market Affects L2 Costs
The concrete numbers tell the story. Before EIP-4844, Coinbase’s Base network was spending approximately $9.34 million on L1 calldata costs in Q1 2024. By Q3 2024 — after blobs launched in March — that number had dropped to approximately $42,000 in the same quarter, per BlockEden’s post-launch analysis. The cost of posting data to Ethereum fell by roughly 95%. Those savings passed directly through to users as dramatically lower transaction fees on Base, Arbitrum, and Optimism.
But the connection between the blob market and your L2 fees is indirect — and that indirection trips people up. The rollup operator pays blob fees, not you. The operator then prices L2 transactions to cover its costs plus margin. When blob fees spike, the operator absorbs the first hit. Prolonged spikes eventually force operators to raise L2 fees to stay solvent. So you do not feel a blob fee spike the moment it happens — you feel it days or weeks later if it persists.
There is also a common misconception worth correcting: a spike in Ethereum mainnet gas prices does not automatically spike blob fees. The two markets are structurally independent. A congested L1 mempool and a congested blob space can happen at the same time, but they do not cause each other. In fact, some of the worst blob fee spikes in 2024–2025 happened during periods when mainnet gas was relatively calm, because the congestion was driven purely by rollup activity rather than L1 DeFi demand.
Understanding Ethereum blobspace — the data lane itself, separate from the fee mechanism — helps explain why capacity expansions like Pectra and BPO2 have such a direct effect on your L2 bill. More blobspace means more supply, which flattens the fee curve during demand surges.
Blob Market Alternatives: When Rollups Leave Ethereum
Not every rollup uses Ethereum blobs. When blob fees spike, or when a project needs higher throughput than Ethereum currently offers, some teams look to alternative data availability (DA) layers: Celestia, EigenDA, and Avail are the three most commonly used.
Celestia, the most established alternative, costs roughly 64% less per MB than Ethereum blobs at typical fee levels. But cheaper is not free of trade-offs. Celestia has its own validator set and does not inherit Ethereum’s mainnet security guarantees. A rollup that posts data to Celestia instead of Ethereum is making a deliberate security trade-off — trading Ethereum’s validator security for lower DA costs. For some use cases, that trade-off makes sense. For others, especially those handling high-value DeFi activity, it is a harder sell.
EigenDA takes a different approach. It uses restaked ETH via EigenLayer — where ETH staked through liquid staking protocols can be re-committed to secure additional services — to provide data availability with higher throughput than Ethereum blobs. The security model differs from native Ethereum DA, but EigenDA aims to inherit part of Ethereum’s economic security through the restaking mechanism.
| DA Layer | Key Trade-off |
|---|---|
| Ethereum Blobs | Highest security (mainnet validator set); cost rises with blob demand |
| Celestia | ~64% cheaper per MB at current rates; separate security model, own validator set |
| EigenDA | High throughput; security via restaked ETH (EigenLayer), not native Ethereum |
Hybrid strategies are also real. Some dedicated appchain deployments use Celestia or EigenDA as their primary DA layer during blob fee spikes, then fall back to Ethereum blobs when prices normalize. The net effect is fee volatility management at the operator level — the rollup hedges between DA options rather than being fully exposed to either.
The existence of these alternatives exerts soft competitive pressure on Ethereum blob fees. If blobs become too expensive for too long, rollups migrate. This is one of the driving forces behind Ethereum’s aggressive push to expand blob capacity through PeerDAS and BPO parameter forks — keeping blobs cheap enough that the migration incentive disappears.
The Blob Market Roadmap: Pectra, Fusaka, and Full Danksharding
The blob market that exists in mid-2026 looks nothing like the one launched in March 2024. It has been materially rebuilt through three major upgrades in barely two years — faster than almost any previous Ethereum protocol area.
At Dencun, EIP-4844 (also called Proto-Danksharding) launched with conservative parameters: 3 blobs per block as the target, 6 as the maximum. Blob fees collapsed immediately to near-zero because rollup demand was well below that target. For most of 2024, a common complaint in CT was that the blob market was essentially irrelevant — fees were too low to matter. That calm masked a structural reality: the market was functioning as designed, just at low utilization.
Pectra in May 2025 doubled the target to 6, giving rollups more headroom and briefly driving fees down again as the expanded supply absorbed existing demand. Then Fusaka in December 2025 was a step change: PeerDAS (EIP-7594) deployed, splitting blob data into 128 distributed columns across the validator network. Each node now stores and serves only a fraction of total blob data, cutting bandwidth requirements by roughly 8x while maintaining data availability guarantees. BPO1 raised the blob target to 10. BPO2, just three weeks later on January 7, 2026, raised it again to 14 with a max of 21.
The full timeline:
| Milestone | What It Changed |
|---|---|
| Dencun / EIP-4844 (March 2024) | Blob market launched; 3-blob target; fees collapsed to 1 wei |
| Pectra (May 2025) | Blob target raised from 3 to 6; fee headroom doubled |
| Fusaka / PeerDAS / BPO1 (December 2025) | PeerDAS deployed; blob target raised to 10/15; EIP-7918 fee floor |
| BPO2 (January 7, 2026) | Blob target raised to 14/21; further capacity expansion |
EIP-7918 is the other major Fusaka change worth flagging: the blob base fee floor, which ties the minimum blob fee to the L1 execution base fee. This addresses the “1-wei collapse” that let blobs be nearly free during low demand — ensuring that blob fee burns contribute meaningfully to ETH’s monetary policy even when the market is not congested.
The long-term target is full Danksharding: 128 blobs per block, roughly 16 MB of data per block, achieved through full implementation of DAS (data availability sampling) across the validator set. At that throughput, Ethereum’s blob capacity would make alternative DA layers largely redundant for most rollup use cases — and the blob market would likely stabilize at fees low enough to keep all major rollups on Ethereum without the need for hybrid DA strategies.
PeerDAS is the stepping stone. It does not deliver 128 blobs yet, but it sets the architecture for getting there. Each BPO fork simply turns a parameter dial further — target blobs and max blobs — without requiring a full hard fork each time, making the upgrade path faster and more predictable than prior Ethereum scaling milestones.
FAQ
What is the blob market in crypto?
The blob market is the fee market Ethereum uses to price temporary data storage for Layer 2 rollups. It was created by EIP-4844, launched in March 2024, and operates independently from Ethereum’s regular gas market. Rollups bid into the blob market when submitting transaction batches to Ethereum mainnet. The price is set by an automated EIP-1559-style mechanism that adjusts the blob base fee up or down by up to 12.5% each block depending on whether demand is above or below the target blob count.
How is the blob market base fee calculated?
The blob base fee starts at 1 wei (the minimum floor before Fusaka) and adjusts each block based on how many blobs were included relative to the per-block target. If actual blob count exceeds the target, the base fee rises by up to 12.5% for the next block. If it is below target, it falls by up to 12.5%. After Fusaka introduced EIP-7918, the blob base fee can no longer fall below a minimum proportional to the L1 execution base fee, so the indefinite 1-wei floor is gone. Each blob transaction pays 131,072 blob gas units multiplied by the current blob base fee.
Why did blob market fees drop after the Dencun upgrade?
Blob fees dropped because rollup demand was well below the initial 3-blob-per-block target set at Dencun. When each block includes fewer blobs than the target, the blob market’s adjustment mechanism cuts the fee by up to 12.5% per block. With consistently low utilization, fees fell to the 1-wei floor and stayed there for most of 2024. The blob market was working exactly as designed — it simply had far more supply than demand at launch.
What is the difference between blob market fees and regular Ethereum gas fees?
Blob fees and regular Ethereum gas fees are priced independently and paid into separate resource pools. Gas fees cover execution — smart contract computation, token transfers, DeFi trades — and are priced in gwei per unit of computation. Blob fees cover temporary data storage for rollup batches and are priced in blob gas, using an EIP-1559-style mechanism with its own base fee and target. A spike in one market does not automatically spike the other. Both types of fees are burned under EIP-1559 and EIP-4844, contributing to ETH’s supply reduction, but blob fee burns are currently much smaller in volume than execution fee burns.
How long does blob data stay on the Ethereum network?
Blob data is retained by Ethereum consensus clients for approximately 18 days (around 4,096 epochs) after the block in which it was included. After that window, nodes prune the raw blob data. The KZG cryptographic commitment that was included in the block header remains on-chain permanently, providing proof that the data existed and was available. Rollups only need the 18-day window to complete their fraud proofs or validity proofs, after which the raw batch data serves no further purpose for Ethereum’s security model.
Where To Start
Whether you are an L2 user trying to make sense of fee spikes, or a developer weighing DA options for a new rollup, a few concrete steps will take you further:
- Check the current blob base fee on Blobscan (blobscan.com) before submitting large or time-sensitive L2 transactions. Blob fees update every block, and a five-minute delay can mean a 10x cost difference during congestion.
- If your fees on Base, Arbitrum, or Optimism just spiked unexpectedly, look at blob fee activity on Blobscan first. A blob fee surge explains most sudden L2 cost increases that are not tied to a mainnet DeFi event.
- If you are building on or evaluating rollup infrastructure, review the DA trade-off table in the Blob Market Alternatives section above. Hybrid DA strategies have become standard for cost-sensitive appchain deployments.
- Track Ethereum’s BPO parameter forks. Each one expands blob capacity without a full hard fork. The path from BPO2’s current 14-blob target to full Danksharding’s 128-blob target will play out in incremental steps over the next two to three years.
Blob fee data changes every 12 seconds. Bookmarking Blobscan before your next large L2 transaction costs nothing and can save you real money during congestion windows. And if you are evaluating rollup infrastructure, the DA trade-off table above is the fastest way to frame the security-versus-cost decision without getting lost in protocol whitepapers.
