Originally published at mrnasdog.com/research/ethfi/inflation by MrNasdog.
TL;DR. ether.fi's ETHFI cannot be minted. The deployed Ethereum contract carries no mint function in any form, it is not an upgradeable proxy, and the on-chain total supply read the identical 998,535,999 ETHFI at both ends of this 90-day window. What still moves is already-minted supply: 17.65M from the last vesting grant and 1.18M of ecosystem-fund grants that actually reached individual wallets — 18.83M in total, against a buy ledger of zero. Net +1.95% over the last 90 days and +1.95% projected for the next 90. The trap is the buyback: ether.fi really does spend revenue buying ETHFI, but the coins go to sETHFI stakers rather than to a burn address, so the tradable float never shrinks.
The verdict, in one paragraph
For the 90-day window ending Sep 5 2026, the MrNasdog Pressure Framework reads ETHFI at +1.95% net, both realised and forward, on a counted base of 965,350,000 ETHFI. Our supply monitor reads +9.23% for the same trailing window — a gap of 7.27 percentage points, far outside the framework's half-point tolerance, so a ⚠ monitor gap chip ships on the ETHFI overview page. The deep walk found the cause and did not fully close it. The counted float stepped up about 46M ETHFI around the 17th of each month, and three of those steps landed inside this window — Jun 17, Jul 17 and Aug 17 2026 — carrying the count to 1,018.2M, above both the 1,000,000,000 cap and the 998,535,999 that exists on-chain, before it was corrected down to 965.3M on Aug 27 2026. The chain's own total supply never moved by a single unit through any of it. ETHFI is best labelled a fixed-supply token in the last months of its vesting tail, with a buyback that is staker yield rather than scarcity.
Sell pressure: where new ETHFI comes from
Sell #1 — protocol inflation — is zero, and it is zero at the bytecode level rather than by policy. The ETHFI ERC-20 on Ethereum at 0xfe0c30065b384f05761f15d0cc899d4f9f9cc0eb contains no mint selector in any shape, no owner slot and no minter role, and it is not a proxy: all three standard upgrade slots read empty and neither upgrade entry point exists in the code. No key, no multisig and no governance vote can add a single new ETHFI. Read at both window ends through archival nodes, and at checkpoints between them, total supply was 998,535,999 ETHFI throughout. The same call returns 1,000,000,000 at an earlier height, which is how we know the archive is answering rather than repeating itself — the flat window is a real flat. The cap is 1,000,000,000; the outstanding total sits below it because 1,464,001 ETHFI was destroyed in a single event in 2025, more than a year before this window opened.
Sell #2 — vesting unlocks — is the bulk of the story at 17.65M ETHFI per 90 days. ether.fi split its billion into five buckets: Investors 33.74%, Treasury 21.62%, Core Contributors 21.47%, User Airdrops 19.27% and Partnerships and Liquidity 3.9%, all behind a one-year cliff from the March 2024 launch. Every bucket except one has reached its final figure — the investor block finished unlocking on Mar 18 2026, before this window opened — leaving only the core-contributor grant still releasing. That grant is 214,700,000 ETHFI on a three-year schedule, a flat drip of roughly 196,073 ETHFI a day with no cliff left, completing on Mar 18 2027 with 38,038,173 still to come. There is no on-chain escrow to read here — the allocation sits in wallets under contract — so under the framework's rules the published release curve governs. It was checked before it was trusted: 38,038,173 locked implies 961,961,827 circulating, and the market's own counted figure is 965,350,000, a difference of 0.35%. Two trackers model this grant faster — one at 318,184 a day, one at 442,714 a day — but the first rests on a team bucket that contradicts ether.fi's own allocation table on two lines, and neither reconciles to the counted float the way the project's curve does.
Sell #3 — Foundation and unscheduled unlocks — is 1.18M ETHFI, and it is measured at the point the coins actually reach holders rather than at the point they leave the treasury. The four ecosystem-fund wallets ether.fi publishes fell from 189,837,646 to 187,830,646 over the window, a drop of 2,007,000. But 2,000,000 of that went into a grants safe one hop short of the market, and that safe has since paid out 1,173,405 ETHFI across 83 transfers to individual wallets while still holding 829,718 undrawn — an identity that closes to the wei against the safe's measured balance. A further 7,000 was sold through a trading contract. Booking the gross treasury figure would have invented 826,595 of pressure that has not happened yet. Sell #4 — long-term locked or bankruptcy — is zero: ETHFI has no estate, no trustee and no court-ordered distribution.
Buy pressure: where new ETHFI goes
Every buy row is zero, and the most important of them is zero for a reason that catches most readers out. Buy #1 — programmatic buyback — records nothing even though ether.fi runs one of the more serious revenue buybacks in DeFi: 100% of eETH withdrawal-fee revenue funds a weekly purchase, a monthly purchase draws on the staking, vault and card businesses, the Aug 13 2026 product release widened the programme across every revenue line, and a separate DAO programme arms up to $50M of treasury buying while ETHFI trades under $3. The framework asks one question of any buyback — where do the coins end up? — and ether.fi's governance documentation answers plainly: repurchased ETHFI is redistributed to sETHFI holders. It is not burned. We traced it by sender rather than by a rising balance, because the staking contract also takes deposits: 579,730 ETHFI moved from the project wallet into the staking contract in exactly three transfers, on Jun 23 2026 and Jun 30 2026, and nothing has followed since. The destination sits inside the counted float — the staking contract holds 110,617,734 ETHFI against a counted supply of 965,350,000 — so money leaves the protocol, coins change hands, and the float ends exactly where it started.
Buy #2 — protocol fee burn — is zero, and this is the row where the burn-versus-transfer test earns its place. A burn can show on either of two surfaces, so ETHFI was read on both: the dead address held the same 0.66 ETHFI at the start and the end of the window, and total supply held the same 998,535,999. Neither surface moved by the size of the buyback, which means the repurchased coins were moved, not destroyed — they remain in the float and are carried as an overhang instead of a burn. Buy #3 — Foundation buy — is zero: the two largest ecosystem-fund wallets held identical balances at both ends and the other two only sent, and the project wallet that handles repurchases ended smaller at 68,659 ETHFI because it forwards what it receives. Buy #4 — new long-term lock — is zero: the staking receipt moves freely between wallets, staked ETHFI is already counted as tradable, and the contract's value per share fell across four checkpoints before recovering, which is a deposit-and-withdrawal queue, not a lock.
Foundation and overhang
ETHFI's overhang is unusually large and unusually still. The biggest single piece is one ecosystem-fund wallet holding 173,200,600 ETHFI that did not move a token in either direction this window — a log sweep across both ends returned zero inflows and zero outflows, so the flat balance is genuinely flat rather than an in-equals-out artefact. The remaining three published ecosystem-fund wallets bring the treasury total to 187,830,646 ETHFI. Behind them sit three more tracked pools: the grants distributor safe with 829,718 staged for release, the core-contributor grant with 38,038,173 still locked on the schedule to Mar 18 2027, and the buyback's destination — the sETHFI staking contract holding 110,617,734 ETHFI, where repurchased coins accumulate on behalf of stakers rather than being destroyed. Every one of these is an on-chain address read at every rebuild. If any of these balances falls between refreshes, the outflow enters Sell #3 at the next refresh.
How ETHFI compares to other fixed-supply DeFi governance tokens
ETHFI belongs to the class of hard-capped DeFi governance tokens that mint nothing and dilute anyway. Mechanically it is the opposite of an uncapped proof-of-stake L1, where a validator subsidy adds supply every block and the only question is the rate. ETHFI has no issuance at all — the mint function is absent from immutable code — so its entire sell side is a redistribution schedule working through already-minted units. That makes it closer in shape to Pendle or Lido DAO than to Kaia or Starknet: the dilution has an end date, and after Mar 18 2027 the vesting engine simply stops.
The sharper comparison is on the buy side, because "this protocol buys back its token" is the single most over-read fact in DeFi. There are two kinds of buyback and they are not close. Lido DAO's treasury buyback delivers into an agent that sits outside the counted float, so bought LDO genuinely leaves circulation and the framework books it. ether.fi's buyback delivers into a staking contract that sits inside the counted float, and the governance documentation is explicit that the coins are redistributed to sETHFI holders. Pendle runs the same shape and gets the same zero. The revenue is real, the purchases are real, and the supply effect is nil — what the holder receives is yield, not scarcity. A burn address would change the answer overnight; a distribution to stakers never does.
The third comparison is with tokens whose counted supply is trustworthy. ETHFI's was not, for most of this window: the classified float ran past its own 1,000,000,000 ceiling to 1,018.2M before correcting to 965.3M on Aug 27 2026. That is the sort of failure a chain read catches instantly and a dashboard never does, and it is why the framework reads wallets and contracts rather than supply feeds.
What to watch in the next 90 days
First, the core-contributor grant keeps dripping about 196,073 ETHFI a day with no cliff, adding roughly 17.65M between now and Dec 4 2026, and finishing entirely on Mar 18 2027. Second, the grants distributor safe still holds 829,718 ETHFI that has already left the treasury and not yet reached holders; that balance draining is the cheapest early warning of a heavier Sell #3. Third, the ecosystem fund's largest wallet at 173,200,600 ETHFI has been silent — its first outbound transfer would be the single biggest change on this page. Fourth, the programmatic buyback announced on Aug 13 2026 has not delivered anything to the staking contract since Jun 30 2026; a resumption would still score zero on supply, but a decision to point any part of it at a burn address instead would flip the buy ledger for the first time. Fifth, the counted-supply correction of Aug 27 2026 should keep the monitor gap shrinking as the monthly steps age out of the trailing window.
Summary
ETHFI is a fixed-supply token that still dilutes. Nothing can mint it — the contract has no mint function and no upgrade path to add one, and on-chain total supply held at 998,535,999 ETHFI across the entire 90-day window — but 18.83M of already-minted supply reached the market, 17.65M from the last core-contributor grant and 1.18M from ecosystem-fund grants, for a net of +1.95% in both the trailing and the forward window. The key risk is that the buy side offsets none of it: ether.fi's revenue buyback is genuine and lands inside the counted float, so it pays stakers rather than reducing supply, and the dead address did not gain a single token. The ceiling is real and close — 38,038,173 ETHFI remain locked and the schedule ends on Mar 18 2027, after which ETHFI's only supply lever is treasury discretion.
MrNasdog Pressure Framework analysis of ETHFI, Metric 1 — Inflation. Data + explanation only. Not financial advice. Updated Sep 5 2026.
Top comments (0)