APT Inflation Analysis · September 2026 · Supply growing, projected to keep growing
Aptos has done almost everything a chain can do to slow its own supply — a protocol cap of 2.1B APT, a staking reward rate cut to a 2.60% floor it cannot fall below, and 100% of every gas fee destroyed — and the Pressure Framework still reads APT at +4.46% over the trailing 90 days and +3.67% over the next 90. The reason is not the mint. Aptos minted 4.92M APT for stakers over the window while the October 2022 allocation calendar released 33.93M APT — about seven times more — and the gas burn took back only 0.51M APT. The one date that changes the arithmetic is Oct 12 2026, when the four-year core-contributor and investor vest finishes and the monthly release drops from 11.31M APT to 4.54M APT.
The verdict, in one paragraph
Against a circulating base of 858.72M APT, the framework books 38.85M APT of sell pressure and 0.51M APT of buy pressure over the trailing 90 days — a net of +4.46% — and projects +3.67% for the next 90 days, the difference being the vesting step that lands on Oct 12 2026. The inflation monitor reads +4.70% for the same window, a gap of 0.24 percentage points, comfortably inside the framework's tolerance, so no data-conflict flag ships on the overview page. That gap is denominator convention and nothing else: the monitor divides its supply change by the 90-day-old base, and running the framework's own 38.33M APT of net float growth over that older base gives +4.67% — 0.03 percentage points from the monitor. The label for Aptos is a capped, burning chain whose launch calendar is still unwinding: the protocol side of the ledger is genuinely tight, and it is not what moves the number.
Sell pressure: where new APT comes from
Only one mechanism on Aptos creates an APT that did not exist before, and it is Sell #1: staking rewards, paid once per epoch to whoever is bonded. The rate is 2.60% a year, and it is worth being precise about what that means, because the chain publishes two of them. The live rate on the Aptos staking rewards module works out to 5.936073e-6 per epoch, and its own floor value is identical to it — the decay curve that has been cutting Aptos emissions since launch has reached its bound and has nothing left to give. A deprecated field on the same module still advertises 7.00%; reading that one instead would have overstated this row by 2.7 times. Against a staked base averaging 765.18M APT across nine reads spanning the window, the mint came to 4.92M APT. The epoch count was measured rather than assumed: Aptos ran 1,083 epochs in the window against the 1,080 a 7,200-second target implies, the reward is paid per epoch rather than per day, and nothing in the protocol rescales for the drift — so reading the calendar would have quietly under-counted the mint.
Sell #2, vesting unlocks, is where the Aptos supply story actually lives, at 33.93M APT. The October 2022 genesis allocated exactly 1,000,000,000 APT across four buckets, and all four release on the 12th of every month: 3.21M to the community, 3.96M to core contributors, 2.81M to investors and 1.33M to the Aptos Foundation — 11.31M APT a month, every month, since launch. Three of those firings landed inside this window, on Jun 12 2026, Jul 12 2026 and Aug 12 2026. It matters that this is not inflation in the minting sense: no APT is created by an unlock, the coins already existed and were already inside the total, and what changes is that they stop being unspendable and become tradable float. That is precisely the quantity the Pressure Framework measures, and precisely the quantity a supply cap does not protect a holder from.
Sell #3, Foundation and unscheduled unlocks, is 0, and on Aptos that can be demonstrated rather than asserted. Aptos publishes no wallet addresses for its locked buckets, so the framework closed the question on an identity instead of inventing a wallet map: the four published buckets say 349,749,982 APT should still be locked after the Aug 12 2026 tranche, and the chain reports 349,749,934 APT not yet counted as tradable — 48 APT apart on a 349.75M base. Every held-back APT is on the published calendar; there is no discretionary pool with its own release decision to make. Sell #4, long-term locked or bankruptcy, is 0 as well: APT has no bankruptcy estate, no trustee and no court-ordered distribution.
Buy pressure: where new APT goes
Buy #1, programmatic buyback, is 0. Aptos runs no programme that spends treasury money repurchasing APT on the open market, there is no contract and no accumulation address to read, and nothing was announced or executed inside the window. The Aptos Foundation has said publicly that it would explore a buyback funded from licensing revenue and ecosystem investments; an intention to explore is not a programme, and the framework does not book it. Buy #3, Foundation buy, is 0for the mirror reason — the Foundation is a holder of a scheduled allocation and a staker of its own coins, not a buyer of anyone else's.
Buy #2, the protocol fee burn, is the one row on the buy side that moves, at 0.51M APT, and it is the row most worth explaining. Aptos destroys 100% of every gas fee — none of it is paid to validators — and the gas price itself was raised roughly tenfold in early 2026. There is no burn address to look at, because the coins are destroyed in place rather than sent to a keyless sink, so the framework proved the number by identity instead of by an empty read: had gas fees merely moved rather than been destroyed, the count of APT in existence would have risen by the full 4.92M that was minted; it rose by 4.40M, and the 0.51M difference is what was burned. That figure was then checked against an independent surface, the project's own supply dashboard, which reports 155,091 APT burned in the last 30 days and 1.7M APT since mainnet launch — a 90-day equivalent about a tenth away from the identity figure. The two surfaces are independent, because supply moves on the mint alone and the burn counter moves on gas alone, but they measure one burn, and the framework books it once.
Buy #4, new long-term lock, is 0, and staking cannot change that on Aptos. There is 752.58M APT bonded across 84 validator pools, which sounds like a large removal from the float until you check whether the float ever counted it — and the same 48 APTidentity that closes the locked bucket proves that the tradable figure equals vested genesis plus every net staking reward, with nothing subtracted for bonding. Aptos's own writing is explicit that locked and unlocked APT can both be staked. So staking removes nothing from this reading.
Foundation and overhang
The APT overhang is very large, completely enumerated, and almost entirely non-discretionary. It is the 349.75M APT still locked across the four genesis buckets, and it decomposes exactly: 237.55M APT still owed to the community bucket, 98.67M APT to the Aptos Foundation, and 13.53M APT across the two insider buckets — core contributors and investors — which finish entirely on Oct 12 2026. Every one of those releases on a published monthly date, so the release calendar is not a decision anyone gets to make each quarter; it is already written. There is no buyback accumulation wallet to track, because there is no buyback, and there is no bankruptcy estate residual.
One item sits outside that arithmetic and is watched rather than booked: the Aptos Foundation has announced it will permanently lock 210M APT and fund operations from staking rewards rather than token sales. Those are already-circulating coins, no address has been published for them, and a removal of that size would show up in the identity above and does not — so it moves neither the sell nor the buy side of this page. The trigger sentence applies to all of it: if the locked balance falls between refreshes by more than the published calendar accounts for, or if a Foundation lock address is published and its balance moves, that outflow enters Sell #3 at the next refresh.
How APT compares to other capped proof-of-stake chains
On the pure issuance axis, Aptos is now stricter than most of its peers. A typical uncapped Cosmos-family or delegated-proof-of-stake Layer 1 runs a staking-linked emission somewhere between 4% and 10% a year with no ceiling in the code at all, and its inflation reading is dominated by that mint. Aptos mints at 2.60% a year, has hit the floor of its own decay curve, publishes a ceiling of 2.1B APT, and destroys every unit of gas rather than paying it to validators. Measured against a halving-model chain like Bitcoin, the comparison is subtler: Bitcoin still mints on every block at a decaying rate on a known clock, and Aptos mints less as a percentage while adding a burn Bitcoin does not have.
And yet APT reads +4.46% while a mid-cycle Bitcoin reads a fraction of a percent, and that is the whole lesson of the comparison. A supply cap constrains total supply; it does not constrain tradable float, and on a chain not yet four years old those are very different quantities. In shape, APT is far closer to a recently-launched token still working through its four-year investor and team vest than it is to a mature capped chain — the difference being that Aptos's vest is smooth and monthly rather than cliff-based, so there is no single dated event to trade around, only a relentless 11.31M APT arriving on the 12th.
The other structural comparison is to exchange tokens that run quarterly buybacks and burns. Those chains offset issuance with a demand-linked removal that scales with usage, and their inflation readings can go genuinely negative. Aptos has the burn half of that mechanism and none of the buyback half, and the burn is not yet large enough to matter: 0.51M APT a quarter is about 1.3% of what reaches the market over the same period. A perpetuals venue built on Aptos is projected to burn "over 32M APT a year" at scale; the entire chain is currently burning at roughly 1.9M APT a year. That is the distance between the mechanism existing and the mechanism mattering.
What to watch in the next 90 days
First, Sep 12 2026: a full 11.31M APT monthly unlock, unchanged in size and composition. Second, and by far the most important, Oct 12 2026: four years to the day after mainnet launch, the core-contributor and investor legs of the vest pay their forty-eighth and final tranche, retiring 6.77M APT of the monthly figure. Third, Nov 12 2026: the first month of the new regime, when the release is the community and Foundation buckets alone at 4.54M APT — a step down of about 60% that will pull the framework reading below +3% from the following window onward, with no protocol change involved at all. Fourth, the gas burn, which is the only lever that could bend this page's sign: it would have to grow roughly seventy-fold to offset the calendar, so what matters is the trend rather than the level, and it is currently softening — the last 30 days annualise slightly below the trailing 90. Fifth, the staking rate review due around Oct 11 2026, which is a formality this year because the rate already sits on its floor, but is the place any future emission change would appear.
Summary
The MrNasdog Pressure Framework reads APT at +4.46% over the trailing 90 days and +3.67% projected forward: supply growing, projected to keep growing. The mechanism is not inflation but unlock — Aptos mints only 4.92M APT a quarter for stakers at a rate sitting on its own floor, burns 0.51M APT of gas, and holds a published ceiling of 2.1B APT, while the October 2022 allocation calendar releases 11.31M APT on the 12th of every month out of 349.75M APT still locked. The key risk is that this is mechanical: it runs at the same rate every month regardless of price, and nothing on the buy side is within an order of magnitude of offsetting it. The genuine comfort is the date — Oct 12 2026 ends the insider vest and cuts the recurring monthly release by about 60%, which is the largest single improvement available to this page and it arrives on a schedule nobody has to vote for.
MrNasdog Pressure Framework analysis of APT, Metric 1 — Inflation. Data + explanation only. Not financial advice. Updated Sep 10 2026.