Ethereum Blob Fees and ETH Burn in 2026: How L2 Data Actually Affects Supply
Site editorial attribution • Organizational attribution
Ethereum Blob Fees and ETH Burn in 2026: How L2 Data Actually Affects Supply
Short answer: Ethereum burns the blob base fee paid by blob-carrying transactions, but blob usage does not guarantee that ETH supply becomes deflationary. Burn depends on both the number of blobs and the blob base fee; issuance continues through proof-of-stake rewards. Fusaka increased blob capacity and introduced a reserve-price mechanism, while Glamsterdam was still being prepared in May 2026 rather than having launched in April. Analyze usage, price, issuance, and rollup economics separately.
An earlier version of this page claimed 1,450 ETH of daily blob burn, 1,200 ETH of additional burn, a 7.2% “real yield,” 80% data-capacity consumption, and an April Glamsterdam activation. It did not provide blocks, fee data, issuance data, or a calculation. Those figures have been removed. This revision publishes the protocol mechanics and a sensitivity model readers can reproduce.
The Core Mechanism in One Equation
EIP-4844 defines one blob as 131,072 units of blob gas. The fee for a blob is:
Blob fee in ETH = 131,072 × blob base fee in wei ÷ 10^18
The protocol deducts that fee from the transaction sender and burns it. The priority and execution-gas components of the same blob transaction follow their own rules. “Blob burn” is therefore not a second token or a special burn event; it is ETH removed through the separate blob fee market.
Daily blob burn can be approximated as:
Daily blob burn = blobs per block × blocks per day × fee per blob
Every term matters. More capacity can allow more blobs while reducing scarcity and fee pressure. A busy blob market at the minimum fee can burn little ETH; a short period of congestion can raise the base fee sharply.
Correction Ledger
| Earlier claim | Protocol and source audit | Correction |
|---|---|---|
| Glamsterdam activated in April 2026 | Ethereum Foundation May update described teams preparing for the next upgrade | Not active at the article date |
| Glamsterdam introduced dynamic blob pricing | EIP-4844 already introduced the self-adjusting blob base fee in 2024 | Mechanism predates Glamsterdam |
| Fusaka demand immediately exceeded new supply | No utilization or fee series was cited | Requires measured target utilization and fee data |
| Blob burn reached 1,450 ETH per day | No calculation or block sample | Removed; sensitivity math published below |
| Blob fees added 1,200 ETH of daily burn | No sustained base fee or blob count | Would require roughly 90.8 gwei under stated target assumptions |
| Rollups consumed 80% of Ethereum DA capacity | No numerator, denominator, or time window | Unsupported percentage removed |
| ETH became “double deflationary” | Both execution and blob fees burn the same ETH supply; issuance continues | Net supply can inflate or deflate |
| Stakers earned 7.2% real yield including burn | Burn is not paid to stakers and “real” requires an inflation/price definition | Yield framing corrected |
| Blobs store proof that data exists | Blobs carry rollup data; commitments/proofs serve distinct roles | Technical description corrected |
| L2 fees must stay low as blob fees rise | Rollup fees depend on DA cost, compression, margins, subsidies, and demand | Conditional, not guaranteed |
What a Blob Is
A blob is a temporary data object attached to an EIP-4844 transaction. Rollups can put batches of transaction data into blobs instead of permanent execution calldata. Ethereum consensus nodes make the blob data available for a defined period, while the execution layer can access commitments to the blobs rather than reading arbitrary blob contents inside the EVM.
The distinction is important:
The old explanation said rollups only post proof that data exists. That confuses data availability with execution validity. Rollups need the underlying transaction or state data available long enough for honest participants to reconstruct and verify the rollup state. Ethereum.org says current blob data is pruned from protocol nodes after roughly 18 days; long-term storage becomes the responsibility of rollup operators, users, indexers, and other services.
Why Temporary Data Is Cheaper
Permanent calldata burdens the execution chain's historical record. Blob data is designed for rollup verification and does not need to remain in every node's history indefinitely. Temporary retention reduces long-run storage pressure and gives blobs a separate resource market.
The separation creates two prices:
A blob-carrying transaction can pay both, because it still uses execution gas to submit commitments and transaction metadata. Calling this “double burn” is arithmetically understandable, but it can mislead investors into assuming a multiplier. Total ETH burned is simply the sum of all protocol-burned fee components.
How the Blob Base Fee Adjusts
EIP-4844 tracks blob gas used and excess blob gas. When usage remains above the target, excess accumulates and the blob base fee rises exponentially. When usage remains below target, excess declines toward zero and the fee falls.
This is a target, not a hard cap:
EIP-7918, included in Fusaka, added a reserve-price relationship tied to execution cost. Its purpose was to prevent blob fees from remaining at an economically negligible floor in conditions where the adjustment mechanism responded poorly to changes in demand and execution fees.
The mechanism does not promise a specific burn. It prices a scarce network resource.
Fusaka Changed Capacity, Not the Basic Burn Rule
The Ethereum Foundation scheduled Fusaka for mainnet activation on December 3, 2025. Its headline scaling feature was PeerDAS, or peer data-availability sampling, which changes how nodes obtain and verify portions of blob data so the network can support more blobs without every node downloading every blob in full.
The Foundation then scheduled Blob Parameter Only forks:
| Event | Target blobs per block | Maximum blobs per block | Scheduled date |
|---|---|---|---|
| Fusaka activation | 6 | 9 | Dec. 3, 2025 |
| BPO1 | 10 | 15 | Dec. 9, 2025 |
| BPO2 | 14 | 21 | Jan. 7, 2026 |
These target and maximum increases create more data capacity for rollups. More capacity can reduce user costs, but it can also lower the market-clearing blob fee if demand does not keep pace. Scaling and fee burn can move in opposite directions over some periods.
The Capacity-Burn Tradeoff
Suppose rollup demand is fixed at six blobs per block:
That is not a protocol failure. Ethereum's scaling objective is affordable, secure data availability, not maximizing fees at any cost.
Glamsterdam Had Not Launched in May 2026
The Ethereum Foundation's May 11, 2026 protocol update said teams were hardening and preparing for Glamsterdam. Earlier Foundation roadmap material described a target of “some time in 2026” and named enshrined proposer-builder separation and block-level access lists as headline features.
The earlier article's April launch and postmortem were therefore impossible on its May 7 publication date. A roadmap name, devnet, testnet, client release, and mainnet activation are separate milestones.
When reviewing any Ethereum upgrade claim, require:
Do not infer mainnet behavior from a conference roadmap or draft EIP.
Reproducible Blob-Burn Sensitivity Model
After BPO2, use a target assumption of 14 blobs per block. Ethereum slots are approximately 12 seconds, giving a theoretical 7,200 slots per day before missed slots or other effects. At target utilization:
Blob gas per day = 14 × 7,200 × 131,072 = 13,212,057,600 blob gas
The table converts sustained blob base fees into daily burn.
| Blob base fee | Fee per blob | Daily burn at 14 blobs and 7,200 slots |
|---|---|---|
| 1 wei | 0.000000000000131072 ETH | 0.0000000132 ETH |
| 1 gwei | 0.000131072 ETH | 13.21 ETH |
| 10 gwei | 0.00131072 ETH | 132.12 ETH |
| 50 gwei | 0.0065536 ETH | 660.60 ETH |
| 90.83 gwei | about 0.011905 ETH | about 1,200 ETH |
| 100 gwei | 0.0131072 ETH | 1,321.21 ETH |
These are sensitivity values, not observed May averages. Real daily burn uses actual blobs in each block, actual base fee, missed slots, and parameter changes.
Why the 1,200 ETH Claim Needed a High Fee
At target blob count, 1,200 ETH per day divided by 13.212 billion blob gas implies about 90.83 gwei per blob gas. An article claiming that burn should show evidence that this fee level and utilization persisted across the stated day. A brief spike cannot be annualized as a permanent tokenomic regime.
Maximum Versus Target Utilization
If every block carried the 21-blob maximum, daily blob count would be 50% above the 14-blob target. At a fixed 10 gwei base fee, the arithmetic burn would be about 198.18 ETH per day. But the fee would not stay fixed under sustained above-target use; excess blob gas would push it higher. A static table helps intuition but does not replace block-by-block simulation.
Net ETH Supply: Burn Minus Issuance
ETH supply changes through at least two broad protocol flows:
Simplified:
Net issuance = consensus-layer issuance - execution fee burn - blob fee burn
If burn exceeds issuance over a period, supply contracts. If issuance exceeds burn, supply expands. EIP-1559 explicitly says Ethereum cannot guarantee a fixed or always-deflationary supply because user demand for blockspace changes.
Worked Net-Supply Example
Assume, only for illustration:
Then:
2,000 - 1,700 - 132 = +168 ETH net issuance per day
Supply would still expand despite meaningful blob burn.
If execution burn rose to 2,100 ETH while the other assumptions stayed constant:
2,000 - 2,100 - 132 = -232 ETH per day
Supply would contract. Blob burn contributes at the margin; it does not determine the sign alone.
Fee Burn Is Not Staking Yield
Stakers receive consensus rewards, priority fees in eligible block-production circumstances, and potentially MEV depending on setup. They do not receive burned ETH. Burn may reduce aggregate supply, which can affect every holder economically, but it is not cash paid into a validator account.
Calling staking APR plus burn percentage “real yield” combines unlike quantities:
A Better Staker Return Decomposition
Net staker return in ETH = consensus rewards + priority fees + MEV - validator/provider fees - penalties/slashing - downtime loss
Net staker return in fiat = ETH-denominated return plus change in ETH market price, with tax and execution effects
Supply contraction can be shown separately as an ETH-wide tokenomic measure. Do not add it directly to staking APR as though both accrue to the same account.
Ethereum's validator documentation also notes that reward rates change with total stake. More staked ETH generally lowers per-validator consensus reward rates under the protocol's issuance curve.
Rollup Economics: Who Pays and Who Captures Value?
A rollup typically collects fees from users, pays costs to execute/sequencer infrastructure and post data or proofs, and may retain a margin. The exact stack differs by optimistic versus validity rollup and by governance design.
Simplified Rollup Income Statement
User fee revenue
minus Ethereum blob and execution costs
minus prover/sequencer infrastructure
minus other chain, bridge, oracle, and operating costs
equals rollup operating margin before incentives and governance distributions
Cheaper blobs can lower user fees, increase rollup margin, subsidize growth, or some combination. The result depends on competition and pricing policy.
Worked Batch Example
Suppose a rollup batch contains 10,000 user transactions and uses two blobs.
Per included transaction, the posting cost is:
0.00562144 ÷ 10,000 = 0.000000562144 ETH
If ETH trades at $2,500 in this illustrative example, that is about $0.00141 per transaction before rollup execution, proving, overhead, margin, failed transactions, and other costs.
If the same batch compresses only 2,000 transactions, posting cost per transaction is five times higher. Compression and batch utilization can matter as much as blob price.
Why More L2 Activity May Burn Less ETH Per User
Ethereum scaling seeks to reduce cost per transaction. Several trends can increase L2 transactions while reducing burn per transaction:
That can still benefit Ethereum by expanding use, settlement demand, liquidity, developer activity, and long-run blockspace demand. But the tokenomic thesis must specify whether it depends on transaction count, total data demand, fee price, ETH working balances, collateral demand, or broader network effects.
“Millions of L2 transactions” is not enough to calculate ETH burn.
Blobs Are Data Availability, Not Permanent Storage
Ethereum guarantees blob availability for the protocol retention window, not forever. This design creates responsibilities beyond L1:
If historical data disappears from every external archive after pruning, users may have difficulty reconstructing old rollup state even though commitments remain. Data availability at settlement time and historical data persistence are distinct properties.
PeerDAS and the Path Beyond Proto-Danksharding
Proto-Danksharding went live with Dencun in March 2024. It introduced the blob transaction format and temporary data availability but initially required broad blob propagation among consensus nodes.
Fusaka's PeerDAS uses sampling and custody distribution so individual nodes need not download every blob in full. This makes higher aggregate blob throughput more practical while preserving confidence that data was available.
Full Danksharding remains a multi-stage roadmap. Ethereum.org lists data-availability sampling, proposer-builder separation, and other protocol work as prerequisites. Roadmap claims of more than 100,000 transactions per second are long-term architecture goals, not current production throughput.
Risks to Monitor as Capacity Expands
Scaling should be judged by reliability and decentralization as well as capacity.
Rollup Security Does Not Come From Blobs Alone
Posting data to Ethereum can give a rollup strong data-availability guarantees, but users still depend on rollup-specific systems:
| Component | Question |
|---|---|
| Sequencer | Can users submit or force-include transactions if it fails or censors? |
| Proof system | Are fraud or validity proofs permissionless and active? |
| Provers | How concentrated is proof generation? |
| Upgrade keys | Can a council or company change contracts immediately? |
| Bridge | What assumptions govern withdrawals to Ethereum? |
| Data | Is all state-reconstruction data posted and archived? |
| Exit | How long and under what conditions can users withdraw? |
| Governance | Who can pause, upgrade, or redirect funds? |
Ethereum.org explicitly notes that many rollups retain centralized sequencers and small prover sets. Cheap blobs improve one layer of the system; they do not automatically decentralize the rollup.
An ETH Value-Capture Dashboard
Track quantities that connect L2 growth to ETH economics.
| Metric | Why it matters | Common mistake |
|---|---|---|
| Blobs per block | Measures data demand | Ignoring target changes |
| Blob base fee | Prices scarcity | Looking only at blob count |
| Blob ETH burned | Direct supply removal | Annualizing spikes |
| Execution base-fee burn | Larger burn channel in many regimes | Calling all burn “blob burn” |
| Consensus issuance | Offsets burn | Assuming supply always contracts |
| L2 user fees | User cost and rollup revenue | Treating as Ethereum revenue |
| L2 posting cost | Demand paid to Ethereum | Ignoring execution overhead |
| Transactions per blob | Compression and batch efficiency | Comparing raw transaction counts |
| Sequencer/prover concentration | Operational risk | Treating rollup as fully decentralized |
| ETH held for fees/collateral | Non-burn demand channel | Assuming every transaction buys new ETH |
Use the same source and observation window. Parameter changes can make pre- and post-upgrade utilization percentages incomparable.
Scenario Matrix for Blob Economics
| Scenario | Usage | Blob fee | User cost | ETH burn implication |
|---|---|---|---|---|
| Capacity grows faster than demand | More total blobs, below target | Low | Lower | Burn may remain small |
| Demand matches new target | High and stable | Moderate | Controlled | Meaningful but not necessarily deflationary |
| Short congestion event | At/above max | Sharp spike | Temporary increase | Large short burn, poor basis for annual forecast |
| Sustained DA scarcity | Above target pressure | Persistently high | Compression/migration response | High burn until demand adapts |
| Rollups move to other DA | Ethereum usage falls | Low | Depends on alternative | Lower burn and weaker settlement demand |
| Compression improves rapidly | More L2 transactions per blob | Lower data per user | Lower | Burn per transaction falls |
The healthiest ecosystem outcome need not maximize burn. Low costs with growing secure settlement can be economically valuable even if ETH supply expands modestly.
Frequently Asked Questions
Are Ethereum blob fees burned?
Yes. EIP-4844 specifies that the actual blob fee is deducted from the sender and burned. It is separate from execution gas and is not paid to validators as a blob base fee.
Did Glamsterdam activate in April 2026?
No official source cited here supports that claim. The Ethereum Foundation's May 2026 update described preparation for Glamsterdam, indicating it had not already activated on mainnet.
Did Fusaka introduce blobs?
No. Dencun introduced EIP-4844 blobs in March 2024. Fusaka activated PeerDAS, EIP-7918, and other changes, then used parameter-only forks to increase blob target and maximum capacity.
Does every blob burn the same amount of ETH?
No. Each blob uses the same amount of blob gas under EIP-4844, but the blob base fee changes with excess demand and protocol parameters. Fee per blob equals 131,072 times the current blob base fee.
Does high blob utilization make ETH deflationary?
Not automatically. Net supply depends on consensus issuance plus execution and blob fee burn. Blob burn can rise while total issuance still exceeds total burn.
Is burned ETH part of staking APR?
No. Burn removes supply and affects all ETH holders indirectly. Staking APR comes from validator rewards and fee/MEV components available to the validator setup, minus costs and penalties.
Do blobs contain rollup proofs?
Blobs generally carry rollup data. KZG commitments and proofs verify blob-data commitments, while rollups use separate fraud or validity mechanisms to verify state transitions. The concepts should not be conflated.
Why can L2 transactions get cheaper while ETH burn rises?
Rollups batch and compress many transactions into blobs. Total data demand and blob price can rise even while the cost is spread across more user transactions. The opposite can also happen after capacity increases.
Can rollups stop using Ethereum blobs?
Yes. A rollup can choose calldata, another data-availability system, or a hybrid, depending on its contracts and security model. That choice changes cost and trust assumptions.
Conclusion
Ethereum's blob market is a real fee-and-burn mechanism, but it is not a perpetual deflation switch. Dencun created the market, Fusaka expanded its scaling architecture and capacity, and future upgrades continue the roadmap. Demand determines whether the fee rises above its floor and how much ETH is burned.
The correct investor framework is simple: measure actual blobs, actual blob base fees, execution burn, and issuance over the same period. Then examine whether rollup growth produces durable Ethereum settlement demand without hiding sequencer, prover, bridge, or data-retention risk.
What to Read Next
Continue with <a href="/insights/ethereum-layer-2-evolution-2026">Ethereum Layer 2 Evolution in 2026</a>. It extends this fee-market model into rollup security, sequencer economics, bridge risk, and the conditions under which L2 growth can accrue value to ETH.
CryptosEyes publishes general protocol and market research, not individualized investment advice. Ethereum parameters, upgrade schedules, and rollup designs can change.
Source & Review Basis
This article is reviewed against the source types below. Source links are provided to help readers verify primary documents, market context, and methodology independently.
Canonical blob transaction, fee, burn, gas accounting, data retention, and fee-adjustment specification.
Official Fusaka activation, PeerDAS, EIP-7918, and blob parameter-only fork schedule.
Official status evidence that teams were preparing and hardening for Glamsterdam in May 2026.
Specification and rationale for the blob reserve-price mechanism included in Fusaka.
Canonical execution base-fee burn and explicit explanation that net ETH supply can be inflationary or deflationary.
Current rollup, blob, data-availability, sequencer, prover, and Danksharding roadmap context.