UNUS SED LEO (LEO) Burn

LEO issued 1B tokens in May 2019 to plug the $850M hole left by the Crypto Capital affair. Current supply per the official API (stats leo.burn.supply) is 919,857,852 LEO, meaning 80,142,148 LEO — exactly 8.01% of genesis — have been destroyed. This ledger does not take supply from CoinGecko: CoinGecko's circulating figure is correct at 919.86M, but its "total supply" still sits at 985.24M, 65.4M above reality, which inflates the listed FDV about 7% above the true market cap.

The 80.14M splits in two: 77,696,619 LEO across 17,578 EOS burn transactions the API returns, spread over 2,469 burn days from 14 Jun 2019 to 1 Sep 2026, and 2,445,529 LEO burned before the API's record begins (kept in extra.legacy_burn, not smeared across the daily series). The two parts reconcile to the supply figure to the unit — meaning this ledger has no "missing" burns, something rarely verifiable for an exchange token.

One chain-level clarification: the LEO ERC-20 contract on Ethereum sits frozen at 660M; all burning happens on the EOS leg. Anyone counting Ethereum dead-wallet balances sees zero and wrongly concludes LEO does not burn — this ledger verified exactly that on 27 Aug before switching to the API.

Summed by year from the API series: 2019 (from 14 Jun) 10.08M over 199 burn days, averaging 50.7K LEO/day; 2020 16.14M, 44.8K/day; 2021 32.41M, 92.9K/day — the Bitcoin bull-run year and the year of record burn days: 19 Apr 2021 784,665 LEO, 25 May 2021 672,171, 18 Jan 2021 557,219, 10 May 2021 539,282 (the second-largest day ever is 13 Mar 2020, 777,280 LEO, right after the Covid crash).

Then the pace broke: 2022 6.32M (18.7K/day, −80% vs 2021); 2023 4.53M (12.9K/day); 2024 3.67M (10.1K/day); 2025 2.66M over 296 burn days (9.0K/day); 2026 to 1 Sep just 1.89M over 209 days (9.0K/day). Measured on burn days, the 2026 rate is under 10% of the 2021 peak — even though LEO's price is many times higher, so the share of supply destroyed per year shrank from ~3.4% (2021) to ~0.3% (2026).

Two readings are both true: Bitfinex revenue measured in LEO falls because LEO's price rose from ~$1 (2019) to $9.32 — the same dollars buy fewer tokens; and Bitfinex's trading share shrank versus 2019–2021. This ledger cannot separate the two from burn data alone; it records the outcome: 3,184,248 LEO burned in the trailing 12 months, 632,692 LEO in the trailing 90 days.

The whitepaper promise is to repurchase and burn "at least daily". The API series shows that held roughly true through 2024 — 2 idle days in 2019, 6 in 2020, 16 in 2021, 26 in 2022, 13 in 2023, and just 1 idle day out of 366 in 2024. In 2025 idle days jumped to 69; 2026 already counts 35 by early September.

The cause is two long pauses for which this ledger found no public announcement. First: from 18 Jul 2025 to 14 Sep 2025, 59 consecutive days without a single burn; on 15 Sep 2025 one transaction of 384,919 LEO — roughly two months of burning at the prior pace, i.e. a catch-up. Second: 8 Jan 2026 to 3 Feb 2026, 27 empty days, then 228,754 LEO burned at once on 4 Feb 2026. Those two catch-ups are the largest days since 16 Jan 2022 (265,286 LEO).

Execution changed too. From 2019 through 17 Dec 2025 each burn day consisted of 8 transactions: one large 3–5K LEO and seven small ones of tens to hundreds — most likely purchases across several trading pairs bundled for burning (this ledger's inference; Bitfinex does not explain). From 18 Dec 2025 there is exactly 1 transaction per day, recently 6,300–7,200 LEO. Hence 17,578 transactions over 2,469 days, 7.1 per day on average.

The 2019 whitepaper rule: iFinex uses at least 27% of the previous month's consolidated gross revenue to buy LEO on the market and destroy it. Because the ratio is fixed, burn value ÷ 0.27 is a rare public estimate of revenue for a private exchange that publishes no financials. This ledger holds burn quantities, not the actual daily purchase prices, so every dollar figure below is this ledger's estimate at the early-September-2026 price of $9.32.

August 2026 burned 213,895 LEO ≈ $2.0M → implied revenue ≈ $7.4M/month. Trailing 12 months 3.18M LEO ≈ $29.7M → ≈ $110M/year. For comparison, the 2021 peak burned 32.41M LEO at prices around $2–3, i.e. $65–95M burned → implied revenue of $240–350M that year (rough estimate; 2021 prices taken as a range). If the whitepaper's 27% is executed faithfully, Bitfinex revenue has fallen by roughly two-thirds from its peak.

Two caveats for the division: "at least 27%" means iFinex may burn more (as on the 500–780K LEO days of 2020–2021, which likely included above-quota buying), so true revenue could be LOWER than the estimate; and the 2025–2026 pauses show the burn cadence no longer tracks revenue day by day, only month or quarter.

LEO has no unlock schedule: all 1B tokens were issued in May 2019, and whatever iFinex and partners hold has counted as circulating from day one — the "Unlocks" column on this page is a true zero, not missing data. Circulating = total = 919.86M, so every LEO burned is a net reduction.

The biggest burn machine has not run yet: the whitepaper pledges at least 80% of Bitcoin recovered from the 2016 Bitfinex hack (the US seized ~94,636 BTC in 2022) and 95% of Crypto Capital recoveries to buy back and burn LEO within 18 months of receipt. As of 25 Feb 2026 that BTC remains frozen pending court allocation between Bitfinex and other claimants; in April 2026 the US government moved only a small ~$606K hack-linked sum to Coinbase. K33 estimated (early 2026) that LEO's then ~$8B market cap sat ~60% above the implied value of that buyback scenario — the market has prepaid for a burn that has not happened.

ATH $10.61 on 4 May 2026, ATL $0.80 on 24 Dec 2019 (CoinGecko). At the early-September-2026 price of $9.32 the 80.14M LEO burned equals $747M — but most was bought at $1–3, so iFinex's actual outlay is far below that figure.

The sole source is api-pub.bitfinex.com: the leo/burn/hist endpoint returns each burn (timestamp, chain, tx hash, amount), capped at 500 per call so the ledger pages by date window; the stats leo.burn.supply endpoint returns current supply. The daily series on this page sums transactions per UTC day; the "burned" row = 1B − API supply. CoinGecko/CMC supply is not used for this token because their "total supply" is 65M off.

There is no revenue column: Bitfinex publishes none. No daily purchase price: the API returns LEO quantities only. USD in the state card = quantity × price at the time the page was last updated, and any revenue inference via 27% is this ledger's estimate. Empty days in the series are days Bitfinex genuinely did not burn (cross-checked against the API twice for Jul–Sep 2025), not collection gaps.