2.07%. That's the number that should be screaming at every ETH holder this morning. Not 20%. Not 15%. Two percent. For the 282-day stretch tracked through ethsupply.fyi, Ethereum's famed fee-burn engine swallowed just 16,524 ETH against 796,623 ETH of fresh issuance. The network printed roughly 778,000 new coins net. And the market is still trading as if EIP-1559 turned ETH into digital gold.
I've been staring at supply ledgers since the ICO days, and I'll tell you straight: the crowd moves fast, but the ledger moves faster. This one is not a price signal. It's a structure signal. And the structure is saying something the euphoria doesn't want to hear.
Let's set the table. EIP-1559 landed in 2021 as Ethereum's great deflationary promise — a base fee that gets burned on every transaction, theoretically shrinking supply as network activity rises. For two years, it worked well enough to fuel the "ultrasound money" meme. Then came the L2 migration, the blob era, and a slow bleed of high-value activity off the L1 execution layer. By 2026, the burn had quietly become a rounding error.
Now Glamsterdam is in the pipeline, dangling a conditional 200 million gas target — a five-fold-plus expansion from the current 60 million ceiling. The pitch is simple and seductive: more block space, more throughput, more fees, more burns. More scarcity. Except the data in front of us says the opposite happens first.
Here's the mechanical reality that gets buried under the marketing. When you raise the gas limit, you spread the same fee demand across more units of gas — which pushes the per-unit base fee down, not up. More room doesn't create more demand. It dilutes the price of the room you already have.
Run the math on the current configuration. To make fee burns fully offset issuance at the present 60 million gas ceiling, the execution-layer base fee would need to sit at roughly 13.85 gwei. Where is it actually sitting? 0.335 gwei. That's a 41x gap. Forty-one times. Not a rounding miss — a chasm.
And here's the part that should make you sit up. Even under the Glamsterdam 200 million gas target, the break-even base fee only falls to about 4.16 gwei. That sounds better until you remember where we are: 0.335. The 200M upgrade doesn't close the gap. It barely dents it. You'd need a demand surge that nobody is forecasting to climb from a third of a gwei to four-plus gwei and hold there.
This is what I mean about the difference between a扩容 narrative and a ledger. The narrative is loud. The ledger is 16,524 ETH of burns versus 796,623 ETH of issuance. Ledger wins.
Let me get granular, because the honesty of this dataset is its best feature. ethsupply.fyi breaks the flows apart instead of lumping them into a headline. Validator rewards: the bulk of issuance. Execution-layer base fee burns plus blob fee burns: the 16,524 figure. Slashing: 1,685 ETH. SELFDESTRUCT and misc: negligible. Net increase: 778,412 ETH over the window — an annualized net rate of about 0.83%.
Read that again. 0.83% annualized net inflation. For an asset that spent 2021-2023 being pitched as a deflationary hedge against fiat debasement, that is a quiet, structural repricing. The "ultrasound money" thesis didn't die with a crash. It died with a spreadsheet.
And it gets sharper when you follow the staking thread. There are roughly 43.75 million ETH staked across 853,325 validators — about 35.8% of the total 122.1 million supply. Crude APR on that base: around 2.36% before node, VPS, and cloud costs. After those, a staker might net 1.5-2.0%. Where the yield is sweet, the risk is steep — but here the yield isn't even sweet. It's thin, and it's paid in newly minted ETH, not in protocol revenue.
That's the crux. Staking rewards on Ethereum are largely a transfer of inflation tax, not a distribution of network income. When burns cover 2% of issuance, the other 98% of your "yield" is dilution dressed up as reward. The validator gets paid. The holder gets diluted. Same coin, two ledgers.
Now the contrarian angle, the one nobody is publishing because it doesn't fit the roadmap slide deck.
EIP-1559, in a low-demand environment, behaves less like a deflation accelerator and more like an inflation amplifier. Think about it. PoS issuance scales with the amount staked. The more people stake, the more ETH gets minted. Burns, meanwhile, scale with L1 transaction demand — which has been structurally migrating to L2s. So you have an issuance curve that rises with participation and a burn curve that falls with migration. Those two lines diverge in exactly the wrong direction. The mechanism was designed for a world of L1 congestion. We live in a world of L1 abundance, and abundance breaks the deflation story.
There's a second blind spot: the Glamsterdam 200M gas target is not a protocol mandate — it's a validator configuration choice. The article's own framing flags it as conditional and validator-dependent. That means the network's future capacity, and therefore its future burn rate, rests on a coordination decision by node operators, not a hard fork guarantee. If validators chase priority-fee revenue and crank the limit up while demand stays soft, base fees sink further and the burn-to-issuance ratio deteriorates. If they hesitate, capacity stalls. Either way, the roadmap's clean "more space equals more scarcity" arrow bends.
And let's talk about the L2 relationship honestly. The current 0.335 gwei base fee means L2s are settling and publishing data to L1 for almost nothing. Blob burns are so small that in one sampled epoch the figure was 0.000442 ETH. That's not a revenue stream. That's a rounding artifact. The L2 ecosystem is thriving on top of an L1 that is charging it near-zero rent. The繁荣 is real — it's just not accruing to ETH's supply ledger. Value capture is leaking upward and sideways while the settlement layer subsidizes the growth above it.
Solana, meanwhile, has reportedly flipped Ethereum on fee revenue. Let that land. The chain with the cheaper, faster, higher-throughput pitch is now out-earning the "ultrasound" chain on the metric that was supposed to prove scarcity. Speed kills, but slow kills too in this game — and Ethereum is running neither fast nor scarce right now. It's running abundant and inflationary.
So what do you actually do with this?
First, stop pricing ETH as a deflationary asset. Price it as what the ledger shows: a structurally mild-inflation asset running near 0.8% net annual issuance. That reframes the comparison to Bitcoin post-halving, where BTC's issuance sits in a similar neighborhood. The scarcity edge ETH enjoyed in 2021 is gone. Hype is the fuel, but fundamentals are the engine — and the engine is printing.
Second, watch the validator gas-target decision as the real signal, not the press release. If operators push toward 200M and base fees slide below where they are now, the burn-to-issuance ratio gets worse, not better. The "扩容利好" could invert into a core-metric deterioration that most traders won't see coming until the next supply print.
Third, treat staking yield with cold eyes. A 2.36% gross that's mostly minted dilution is not income. It's a slow rebalancing against you. If rate environments shift or risk appetite turns, expect staking growth to stall or even unwind — a negative-feedback loop that quietly cuts issuance but also signals waning conviction.
The market is still trading the 2021 story. The 2026 ledger is telling a different one: burns at 2%, issuance climbing, validators holding the capacity dial, and an L1 collecting near-zero rent from the ecosystem it anchors.
I've seen the moon. Right now I'm looking for the exit — not because ETH is broken, but because the scarcity pitch that carried it here is. The next question isn't whether Glamsterdam ships. It's whether anyone re-reads the supply ledger before they re-price the narrative.