Shortly after Ethereum finished switching from Proof-of-Work to Proof-of-Stake in late 2022, the supply charts began doing something that no one in the cryptocurrency community had ever seen from a significant network. They were falling. Not drastically, as in the case of a crisis, with fear and noise, but quietly and steadily, as in the case of a structural process operating as intended.
ETH was being burned more quickly than it was being produced. The “ultrasound money” theory was generating real-time data to bolster its claim that Ethereum would be a more difficult asset than Bitcoin since its supply might actually decline. The charts then leveled off, the burn rate progressively decreased, and by early 2026, the supply line was once again pointing slightly upward. The ultrasound had been silent.
| Concept | “Ultrasound Money” — Ethereum community term describing ETH as a harder asset than Bitcoin due to its fee-burn mechanism (EIP-1559), which can destroy more ETH than the network creates during periods of high demand |
|---|---|
| Supply at the Merge (Sept 2022) | Approximately 120,520,000 ETH — the baseline from which the deflationary narrative was measured following the switch from Proof-of-Work to Proof-of-Stake |
| Supply (April 2026) | Approximately 120.7–121.5 million ETH — net growth of over 950,000 ETH since the Merge, reversing the brief deflationary period of 2022–23 |
| Current Annual Inflation Rate | ~0.23% — mildly inflationary as of early 2026; daily ETH issuance to stakers (~1,700 ETH/day) is outpacing the burn rate during low-activity periods |
| Key Cause: Dencun Upgrade | March 2024 upgrade that reduced Layer-2 transaction fees by 90–98% — dramatically cutting the volume of ETH burned per transaction as activity migrated off the mainnet |
| Proposed Fix: EIP-7918 | Part of the Fusaka upgrade — introduces a minimum burn floor for L2 data, ensuring a baseline level of ETH destruction even during low-activity periods on the mainnet |
| Staked Supply | Over 30% of all ETH is currently staked and locked from circulation — a structural factor limiting available supply despite the mild inflationary trend in total issuance |
| Institutional Catalyst Watched | BlackRock’s BUIDL fund and real-world asset tokenisation on Ethereum L1 — expected to drive high-value transaction volume that would push burn rates back above issuance |
Without getting bogged down in the technical details, it is worthwhile to comprehend the mechanics of what transpired. A portion of each transaction fee paid on the network is automatically destroyed by Ethereum’s fee-burn mechanism, EIP-1559, which was introduced in 2021. The burn is high when there is a lot of activity. The burn decreases with decreased activity. By introducing Proof-of-Stake, which substituted staking for energy-intensive mining, the Merge significantly reduced the annual rate of new ETH issuance from roughly 4-5% to a fraction of that.
Because of the combination, Ethereum was actually deflationary for a while. The issue is that the burn mechanism is only as strong as the on-chain activity feeding it, which became evident following the Dencun upgrade in March 2024. Additionally, Dencun diverted massive quantities of Ethereum traffic off the mainnet and onto less expensive networks created on top of it by cutting Layer-2 transaction costs by 90 to 98%. Less ETH was burned when there was less mainnet congestion. By early 2026, the network was burning significantly less and issuing about 1,700 ETH to stakers every day. The yearly rate of inflation had stabilized at approximately 0.23%.
It’s easy to interpret that figure as a failure, but it’s more difficult to understand it as a structural state brought about by success. By most counts, Dencun’s supercharged Layer-2 environment is exactly what Ethereum needed: more users, more apps, cheaper transactions, and wider accessibility. Ironically, the scaling approach was so effective that it depleted the activity that gave the base layer’s burn mechanism its strength. It’s similar to constructing an efficient highway system surrounding a city and then asking why the major downtown thoroughfare is quieter. The roads are in decent condition. The traffic moved in a different direction.
Investors who fell for the ultrasound story are now dealing with a more conditional situation than they were initially offered. Although the deflationary argument was always contingent on utilization, it was partially obscured by the charts depicting declining supply during the post-Merge frenzy.
Since over 30% of all ETH is staked and kept away from circulation, there is a structural ceiling on the amount of supply that is actually available to the market, and the current ~0.23% yearly inflation is not concerning because it is a small portion of what Ethereum inflated at before to the merger. However, when the supply is gradually increasing, it becomes more difficult to maintain the “supply squeeze” narrative, which holds that there won’t be enough ETH for buyers to find.
The Fusaka upgrade’s EIP-7918, which would set a minimum burn floor for Layer-2 data fees, is the suggested remedy. The goal is to avoid the burn rate from completely dropping to zero during market lulls by making sure that a baseline amount of ETH is destroyed even during calm times.
Beyond the burn floor, issues such as whether high-value institutional activity on Ethereum’s base layer starts up in a significant way will determine if it will be sufficient to restore the deflationary dynamic. The market is keeping an eye on BlackRock’s tokenized fund on Ethereum L1, not because of its current size but rather because of what it indicates about the potential pipeline of real-world asset settlement. The rate at which that pipeline translates into actual burn volume is currently unknown. At least the direction appears correct.
