Fee Burn and Net Issuance: Reading Monetary Policy On-Chain
Ethereum's 2026 year-to-date ledger reads: gross issuance to validators of 796,623 ETH, fee burn across execution and blob fees of 16,525 ETH, consensus penalties of 1,686 ETH, and roughly 778,413 ETH net added to supply. The burn offset about 2.1 percent of issuance. That is the number most often described as monetary policy you can read off a chain, and the description is wrong in a specific way worth understanding before using it.
What Net Issuance Actually Is
The metric has four components and only one of them is set by a formula nobody can change quickly. Gross issuance is paid to validators and scales with the square root of total staked ETH, so it rises sublinearly as staking grows. Against that sit two separate burns: the execution base fee from EIP-1559, and the blob base fee introduced with EIP-4844. Consensus penalties are a small negative.
Both burns destroy ETH, and they are priced in two different markets. That separation is the whole story. Execution gas prices competition for mainnet computation; blob space prices competition for rollup data availability, and the two fee levels can and do diverge by orders of magnitude.
The live series are published at ultrasound.money and beaconcha.in, and both resolve the same way: issuance is smooth and predictable, the burn volatile and currently small. The split of the burn between execution and blob components is the part that repays attention.
Three of the Inputs Are Parameters
Here is what the mechanical reading misses. The blob base fee had a protocol-set minimum of one wei at launch, and with blob capacity heavily oversupplied relative to demand it sat there, so rollups consumed base-layer data for close to free and the blob burn was a rounding error.
Fusaka, which activated on December 3, 2025, changed that with EIP-7918, setting a reserve price for the blob base fee tied to the execution base fee at roughly a sixteenth of it. On activation the blob base fee rose by a factor of approximately 15 million. That is the largest move in this fee market's history and it contained no information about demand whatsoever. It was a constant being changed.
Two more parameters sit alongside it. EIP-7892 allows blob parameters to be adjusted through lightweight blob-parameter-only hardforks rather than full upgrades, which makes blob capacity a routinely tunable dial. And EIP-8361, proposed in August 2026, would reduce validator rewards toward zero once roughly half of all ETH is staked, which would rewrite the issuance side of the equation outright. Three of the four inputs are therefore governance choices, not market outcomes.
What 10x Adoption Would Actually Do
Which makes the scaling question answerable, and the answer is not encouraging for anyone treating this as an adoption proxy.
Take the arithmetic. A sampled base fee of about 0.335 gwei was observed on October 9, 2026, and at the prevailing 60 million gas limit the base fee would need to average roughly 13.85 gwei for the burn to cancel issuance. That is a move of about forty-one times, not ten. Meanwhile the blob side only burns above its floor when blob demand exceeds the target, and the target is adjustable downward in difficulty by the same governance that sets it. Raise blob capacity ahead of demand and the blob fee returns to the reserve price regardless of how much rollup activity arrives.
So ten times the adoption does not produce ten times the burn. It produces whatever the parameters permit, and they have consistently been set to keep data cheap because cheap data is the point of the roadmap. The base layer earning less when it succeeds at scaling is the designed consequence of moving demand onto a resource the protocol prices deliberately. An allocator cannot forecast net issuance from an adoption forecast; the forecast required is a governance forecast.
Distortions, and Four Better Numbers
The most common distortion is window selection. Published net inflation figures for this period run from roughly 0.2 to 0.8 percent annualized depending on the window chosen, a spread wide enough to support opposite narratives from one chain, so any figure quoted without its window is not a measurement. The established decoupling of base-layer fees from transaction counts compounds it: a quiet mainnet coexists with record activity one layer up.
Volatility also flatters the burn asymmetrically. Because the blob reserve price is pegged to the execution base fee, a congestion event raises both burns together, so a single turbulent day can lift a monthly average materially. Judge the burn on a median rather than a mean, and look at the distribution of daily base fees rather than their average.
Four companions carry more information than the headline. The staking ratio, near 30 percent of supply, because issuance scales with it and because it determines when a proposal like EIP-8361 would bind. Blob utilization against the current target, which says whether the blob burn is at the floor or above it. The split of the burn between execution and blob components. And issuance per unit of staked ETH, which isolates the formula from the staking decision.
Healthy Versus Unhealthy, and the Constructive Read
An unhealthy pattern is net issuance falling only because staking participation fell, since that is security weakening rather than demand strengthening. A healthy one is a rising blob burn while blob capacity is also rising, because that means data demand is outrunning a target that is being raised, which is the only configuration in which scaling and burn grow together.
The honest comparison is less dramatic than either camp claims. Net issuance of a few tenths of a percent sits below pre-Merge Ethereum's four to five percent and near Bitcoin's current schedule, making the asset neither the deflationary instrument of the 2022 framing nor inflationary in any meaningful sense. Scarcity in 2026 reads better from the share of supply locked in staking than from the burn counter.
The constructive signal is that the problem is now being addressed with explicit policy rather than hope. Fidelity's modeling of the EIP-7918 counterfactual estimated roughly 24,641 ETH, about $78.6 million at the ETH price used, in additional blob revenue had the floor existed since blobs launched, across 93 percent of days. A parameter that recovers that is a real improvement. It is also the clearest evidence that this series is a policy instrument, and the right way to read it is as the output of decisions that can be tracked, argued with and anticipated.
For informational purposes only. Not an offer to buy or sell any security. Available only to accredited investors who meet regulatory requirements.
