Originally published at mrnasdog.com/research/leo/inflation by MrNasdog.
UNUS SED LEO has no issuance at all: every one of the 1,000,000,000 LEO was created and sold in May 2019, with no vesting calendar, and neither the Ethereum nor the Vaulta side minted a token in this window. The only flow that moves LEO supply is iFinex buying LEO back and taking it off the market — 633,733 LEO in the 90 days to Aug 31 2026, across 86 on-chain transfers. The Pressure Framework reads LEO at −0.07% net against a supply monitor at −0.13%, a gap of about 0.06 percentage points, well inside tolerance. LEO is a fixed-supply exchange token that shrinks slowly and verifiably.
The verdict, in one paragraph
For the 90-day window ending Aug 31 2026, the MrNasdog Pressure Framework reads UNUS SED LEO at −0.07% net supply change, and projects the same −0.07% forward. Total sell pressure was zero; total buy pressure was 633,733 LEO, against a circulating base of 919,864,410.9 LEO. Our supply monitor reads the same window at −0.13%, leaving a gap of about 0.06 percentage points — comfortably inside the half-point tolerance, so no monitor-gap chip ships on the LEO overview. The residual is ordinary noise in a market-cap-over-price supply series, not a disagreement about mechanism: LEO's float is a pure function of chain state, and both readings agree the number is falling. The cite-able label for LEO is a fixed-supply exchange token with a chain-verifiable daily buyback — deflationary, but at a pace measured in hundredths of a percent per quarter, not in headlines.
Sell pressure: where new LEO comes from
Nowhere, and that is the whole of it. Sell #1, protocol inflation, is zero because UNUS SED LEO is an exchange token rather than a chain: there is no consensus to pay for, no block reward denominated in LEO, and no staking emission. On Ethereum, the LEO ERC-20 reports a total supply of 660,000,000 LEO, and it read exactly that at both ends of this window — the same figure it has carried since the 2019 issuance. On Vaulta, the chain formerly known as EOS, the LEO contract reports a supply of 307,653,658.9 LEO, with zero issue, retire and create actions across the window on two separately operated history clusters. Together the two chains hold 967,653,658.9 LEO, and neither number moved.
One honest caveat belongs here, and it is the reason no LEO row is tagged as permanent. The mint authority on both chains is still live. The Vaulta contract still names an active issuer and still carries 692,346,341 LEO of unissued headroom beneath its own one-billion ceiling. The Ethereum token is a MiniMe-family contract whose own deployed bytecode carries a token-generating function, a token-destroying function and a controller-changing function, and whose controller address is a live 1,783-byte contract rather than the zero address. Neither power has ever been exercised, on either chain, in seven years — but a supply that is unchanged by choice is a different thing from a supply that is unchangeable by code, and the Pressure Framework records the difference rather than rounding it away.
Sell #2, vesting unlocks, is zero. The May 2019 sale delivered every token at once, at one dollar each, with no discount tier and no lockup, and no vesting or escrow contract exists on either chain. There is no cliff calendar left to run, so no future date can add LEO to the float. Sell #3, Foundation and unscheduled unlocks, is zero for the window, because nothing left company hands toward the market — the two identified Ethereum company wallets in fact grew by a combined 374,591 LEO, the opposite of sell pressure. Sell #4, long-term locked or bankruptcy, is also zero: the 2016 breach restitution returns bitcoin to iFinex, not LEO, so no estate or trustee holds LEO to distribute.
Buy pressure: where new LEO goes
Buy #1, the programmatic buyback, is the entire LEO ledger, and it is 633,733 LEO over the window. iFinex committed to spend no less than 27% of consolidated gross revenues repurchasing LEO on the open market and burning it, continuously, until none remain in commercial circulation. In practice this executes as roughly one transfer per day on Vaulta, from the operating account into the token issuer account, tagged with the memo burn. We counted 86 such transfers in the window, about 6,964 LEO a day: 218,999 from Jun 2 2026, 200,839 in Jul 2026 and 213,895 in Aug 2026. The stream paused for five days, from Jun 30 2026 to Jul 4 2026, and two independently operated history nodes confirm that pause is real rather than a hole in one index. Three readings — the exchange's own public burn feed and two separately operated Vaulta history nodes — return that same figure to the token.
What the word burn actually does here is worth being precise about, because it is not what most readers assume. The repurchased LEO is not destroyed. We check a burn two ways on every coin — the balance of the chain's unspendable dead address, and the token's own supply figure — and on LEO neither one moved, on either chain. The two Ethereum dead addresses held 0 LEO at both ends of the window, Vaulta has no dead account in the flow at all, and both supply figures are exactly what they were 90 days ago. The one surface that did move is the issuer's own balance, which is an accumulation rather than a destruction, so it is counted once and only once. The tokens are transferred into the issuer's account and parked. What makes that equivalent to a burn in practice is an accounting identity that closes exactly. The Ethereum supply of 660,000,000, plus the Vaulta supply of 307,653,658.9, minus the 47,789,248 LEO sitting in that account, equals 919,864,410.9 — the same circulating figure the wider market quotes for LEO, to the decimal. Every LEO that lands there leaves the float the instant the transfer confirms, and the account has never sent a token back out. The framework books the realised removal, not the announced one, and that same discipline kills the louder claims: the story that 18,000,000 LEO were burned in the first quarter of 2026 is false — the chain shows 781,820 LEO for January, February and March combined, about 23 times smaller.
A second number gets misread the same way, in the opposite direction. The roughly 79.9 million LEO that gets quoted as burned is a lifetime total, not a quarterly one, and it reconciles cleanly: 1,000,000,000 issued minus 919,864,410.9 circulating is 80,135,589 LEO removed since 2019, made of about 47.79M parked in the buyback account and about 32.35M retired from the Vaulta side before the current practice began. Divided across seven years that is a slow, steady drip — which is exactly what the daily stream shows.
Buy #2, protocol fee burn, is zero, and it cannot be anything else — LEO is not the gas token of either chain it lives on, so no network fee is ever denominated in LEO and none can be destroyed that way. Buy #3, Foundation buy, is zero for now, though it is armed: iFinex pledged 80% of net proceeds from the 2016 breach recovery and 95% of any recovered Crypto Capital funds to additional LEO buybacks, and the seized bitcoin began moving in Apr 2026. Nothing has reached the burn stream yet — the daily pace runs at about 6,964 a day now, and the quarterly totals sit flat at 781,820, 687,526 and 414,734 LEO for the first, second and part-third quarters of 2026 — no step up anywhere. Buy #4, new long-term lock, is zero: there is no staking contract to lock LEO into, and repurchased LEO is parked permanently rather than escrowed for later release.
Foundation and overhang
Four company-controlled holdings are worth tracking, and only one of them is small. First, an Ethereum cold-storage safe holding 648,000,000 LEO — about 98% of the Ethereum-side supply and roughly 70% of the circulating figure the market quotes. It read the same at both ends of the window, has never made a single outgoing transfer, and is counted inside the quoted float. That last detail is the important one: the headline 920M LEO is not what actually trades. Second, a company cold account on Vaulta holding 257,791,296 LEO, which did not move at all during the window. Third, the buyback account itself, now holding 47,789,248 LEO and rising by about 6,964 a day; it is an overhang only in the formal sense, since it is already excluded from circulating. Fourth, the Vaulta operating account at 2,069,561 LEO, which drains only into the buyback account.
The rule we apply to all four is the same: if any of these balances falls between refreshes, the outflow enters Sell #3 at the next refresh. Today none of them has, so every overhang row carries a value of zero — enumerated because we are watching it, not because it is moving. Each of the four is read from its own chain on every refresh, so a move would surface within a day. The unissued 692,346,341 LEO of Vaulta headroom is read the same way.
How LEO compares to other exchange tokens
Exchange tokens split into two mechanism families, and LEO sits at the far end of one of them. The first family issues on a chain of its own and then burns some of it back — the model where a token is simultaneously gas for a network and the subject of a quarterly buyback. Those tokens can show real deflation, but the burn has to outrun live issuance, and the issuance never stops. LEO has no issuance to outrun. Its supply was fixed in a single 2019 event, so every unit the buyback removes is a permanent reduction, and the direction of the supply curve cannot reverse unless iFinex breaks its own commitment and uses a mint authority it has never touched.
The second distinction is between a burn you can verify and a burn you are told about. Many exchange-token burn programmes are disclosed quarterly by the company, sized from revenue figures nobody outside can audit. LEO's is different in kind: the buyback lands as a dated on-chain transfer with a memo, into an account whose balance is arithmetically excluded from circulating supply. You can reconstruct the entire quarter from the chain without reading a single company statement — which is exactly how we established that the widely repeated quarterly burn figure was more than twenty times too large.
Where LEO is weaker than its peers is concentration, and the multi-chain shape makes it worse rather than better. A gas token's float is distributed across the users who need it to transact. LEO's is dominated by a dormant Ethereum safe holding roughly 70% of the quoted supply, with most of the remaining Vaulta side sitting in company custody. The deflation is real but small — a 0.07% quarterly reduction is a rounding error next to a holding of that size, and a decision to move even a fraction of that safe would swamp several years of buyback in a single transaction.
What to watch in the next 90 days
First, the daily burn transfer itself: it has run between 200,000 and 220,000 LEO a month all quarter, and any sustained move away from about 7,000 LEO a day is the cleanest possible signal that iFinex revenue has changed, in either direction. Second, the 2016 breach restitution — the seized bitcoin began transferring in Apr 2026, and the pledge commits 80% of net proceeds to LEO buybacks within eighteen months; if it lands, it arrives as a visible step up in the same daily stream, not as a separate announcement. Third, the Ethereum cold-storage safe: the first outgoing transfer it has ever made would be the single most important supply event in LEO's history. Fourth, any use of the mint authority on either chain, which has never happened and would change the reading immediately. Fifth, whether any LEO ever leaves the Vaulta buyback account — it never has, and the identity that makes the float honest depends on it.
Summary
UNUS SED LEO is one of the few tokens where the supply question has an arithmetic answer. Nothing mints LEO on either chain, no vesting remains, and the sole flow is a buyback that removed 633,733 LEO in the 90 days to Aug 31 2026 — a net of −0.07%, matched by a supply monitor at −0.13%. The structural mechanism is a revenue-funded, chain-verifiable daily buyback against a supply fixed once, in 2019. The key risk is not inflation but concentration: 648,000,000 LEO sit dormant in one Ethereum safe inside the quoted float, roughly 70% of it, and a single outgoing transfer there would matter more than years of buyback. The ceiling is 1,000,000,000 LEO issued, about 8.0% of it already removed — and the only thing that could add more is a mint authority both chains still carry and neither has ever used.
MrNasdog Pressure Framework analysis of LEO, Metric 1 — Inflation. Data + explanation only. Not financial advice. Updated Aug 31 2026.
Top comments (0)