Originally published at mrnasdog.com/research/arb/inflation by MrNasdog.
Arbitrum has never minted a single ARB — all 10,000,000,000 were created at the March 2023 genesis and the Arbitrum One supply counter has only ever fallen since — yet the Pressure Framework reads +6.33% net supply to market over the 90 days to Aug 23 2026. Sell pressure is 422.5M ARB, buy pressure is 0, because Arbitrum has no buyback, no burn and no lock: three monthly vesting tranches of 92.65M ARB plus 144.6M ARB released out of the Arbitrum DAO treasury and the Arbitrum Foundation's own vesting lock. Our supply monitor reads +6.76% — a gap of 0.43 percentage points, inside tolerance, so no data-conflict chip ships. Arbitrum's dilution is not an emission curve; it is a calendar and a treasury, and the calendar runs to Mar 16 2027.
The verdict, in one paragraph
For the 90-day window ending Aug 23 2026, the Pressure Framework reads ARB at +6.33% net. Sell pressure is 422.5M ARB, buy pressure is 0 ARB, and the circulating base is 6.68B ARB. Our supply monitor reads +6.76% — a gap of 0.43 percentage points, comfortably inside the half-point tolerance, so the framework and the monitor agree and no warning chip is rendered on the Arbitrum overview page. That agreement is unusually tight in absolute terms: the framework books 422.5M ARB of flow and the monitor's float grew 422.8M ARB over the same window, a difference of six hundredths of one percent. Arbitrum is best characterised as a zero-issuance token that is inflationary by schedule — nothing is created, but a fixed monthly calendar and a discretionary DAO treasury together push already-minted ARB into the tradable float faster than anything removes it.
Sell pressure: where new ARB comes from
Not from issuance. Sell #1, protocol inflation, is zero for Arbitrum. The entire ARB supply was minted once in March 2023, and the Arbitrum One token contract returns 9,999,998,977.63 today — 1,022.37 ARB BELOW the genesis figure, and it has never risen. We read the token's implementation contract directly rather than trusting a selector audit, and the wording matters: ARB has a live mint function. It is owner-restricted to the Arbitrum DAO's upgrade executor, capped at 2% of supply and callable no more than once a year, and it has never fired in the token's life. ARB is also an upgradeable proxy under DAO control, so this zero can never be recorded as permanent — a governance-controlled token can always be changed by governance. Because the supply read was identical at both ends of the window, we did not accept it at face value: nine evenly spaced interior samples were taken, each paired with a companion balance that had to move, and the Arbitrum DAO treasury duly stepped from 2,656.9M to 2,513.1M to 2,556.1M across them. The endpoint was live. The Arbitrum supply really is flat.
Sell #2, vesting unlocks, is 277.9M ARB and it is the metronome of this token. Arbitrum's own governance documentation puts 2,694M ARB with the team, contributors and advisors and 1,753M ARB with investors, all on a four-year lockup from Mar 16 2023, first unlocking Mar 16 2024 and then monthly on the 16th until the final tranche on Mar 16 2027 — 92.645M ARB a month. Three of those fell inside this window: Jun 16, Jul 16 and Aug 16 2026. We did not take the trackers' word for it. A complete block-by-block sweep of Aug 16 2026 — 342,093 of 342,093 blocks covered, zero failed chunks — found no 92.65M transfer anywhere; the largest ARB movement that day was 10.4M between two trading pools. The Arbitrum tranche is a paper entitlement spread across hundreds of individual beneficiary wallets, not one vault opening, so there is no single escrow to read a realised figure against and the published calendar governs.
Sell #3, foundation and unscheduled unlocks, is 144.6M ARB, and it is the row that decides this coin. The question that unlocks it is whether the Arbitrum DAO treasury sits inside the counted float or outside it — if inside, a grant payout is already-counted ARB changing hands and is not new supply at all. The subtraction settles it. Arbitrum's non-circulating stock is 10,000.0M total minus 6,678.1M circulating, or 3,321.9M ARB. Take out the DAO treasury at 2,556.1M and the Arbitrum Foundation's vesting lock at 124.7M and what remains is 641.1M — against 648.5M of scheduled vesting still to run over the seven remaining monthly tranches to Mar 16 2027. Those agree to 1.14%. The treasury is outside the float, the vesting calendar is real, and both are confirmed by the same arithmetic. Crucially, the Foundation's operating safe cannot also be subtracted — adding its 325.3M would overshoot the non-circulating bucket by roughly that amount — so that safe is INSIDE the float, and money moving from the treasury into it is genuinely supply arriving at market.
With the destination settled, the flows were traced by sender rather than by watching a balance rise. The Arbitrum DAO treasury sent 230,000,000 ARB to the Foundation's operating safe on Jun 28 2026, the executed Arbitrum Foundation operating budget, and took back 86,183,334 on Jul 7 2026 and 42,999,990 on Jul 21 2026 as surplus capital was returned from wind-down wallets — a real net release of 100.7M ARB. Alongside it the Arbitrum Foundation's own vesting lock released 43,896,209 ARB into the same safe across May 29, Jun 30 and Jul 31 2026, a per-second vest that exhausts in April 2027. One thing is deliberately NOT added: the safe's standing 10M ARB a month onward transfer to an exchange deposit path. That is a move between two addresses both already inside the float — it is where the 144.6M ends up, not a separate 30M on top of it, and booking it twice would invent sell pressure. Sell #4, long-term locked or bankruptcy, is zero: Arbitrum has never been through an insolvency and there is no estate, trustee schedule or court-ordered ARB distribution.
Buy pressure: where new ARB goes
Nowhere. All four buy rows read zero, and that is the sharpest fact about ARB. Buy #1, programmatic buyback, is zero because Arbitrum has no buyback contract and no approved buyback programme; a bond-issuance buyback proposal was floated on the Arbitrum governance forum and never carried. Arbitrum does earn real revenue — sequencer fees, Timeboost priority auctions, and under the Arbitrum Expansion Program a 10% share of net revenue from every Orbit chain, announced on Jul 9 2026 — but every unit of it accrues as ETH or stablecoins into the Arbitrum DAO treasury. That announcement is a treasury-building event, not token value accrual; converting any of it into ARB purchases would need its own governance vote, and none has passed.
Buy #2, protocol fee burn, is zero and structural rather than incidental: Arbitrum charges gas in ETH, not in ARB, so there is no ARB fee stream that could be destroyed, and the Arbitrum One supply counter did not fall by a single unit across the window. Buy #3, foundation buy, is zero — no Arbitrum entity has disclosed an open-market ARB purchase, and every identified Arbitrum address was a net sender this quarter. Buy #4, new long-term lock, is zero, and here we read the contract instead of the announcement: ARB staking is a liquid design in which the receipt token is itself tradable and rewards are funded from sequencer revenue in ETH, so it neither mints ARB nor removes it, and the staking contract we located holds 0.0000 ARB at both ends of the window with fewer than 200 receipt units in existence. Nothing is absorbing Arbitrum's monthly unlock.
Foundation and overhang
Arbitrum's team-controlled overhang is the largest of any Layer 2 and it is unusually readable. First and dominant, the Arbitrum DAO treasury at 2,556,141,992 ARB — the single largest ARB holding in existence, roughly 38% of the tradable float, unscheduled and reaching the market only on a discretionary governance vote. Second, the Arbitrum Foundation operating safe at 325,327,067 ARB, funded by the 230M budget transfer on Jun 28 2026 and paying 10M a month onward to an exchange; it sits inside the counted float, so its balance is a queue of already-counted supply rather than an unlock, but it is the address to watch for acceleration. Third, the Arbitrum Foundation vesting lock at 124,672,924 ARB, releasing linearly into that safe and exhausting around Apr 17 2027. Fourth, the unvested team and investor stock of roughly 641.1M ARB, dispersed across per-beneficiary wallets with no single escrow contract, fully scheduled out by Mar 16 2027 and already booked in Sell #2.
Deliberately excluded from that list: exchange custodial wallets, which belong to depositors rather than to Arbitrum, and the Arbitrum One bridge escrow holding 231.3M ARB against the Ethereum representation of the token — that escrow mirrors an existing supply rather than adding to it, and treating it as an overhang would double-count. All four real overhangs are read on-chain at every refresh, and the trigger is the same for each: if any of these balances falls between refreshes, the outflow enters Sell #3 at the next refresh. That is exactly how this quarter's 144.6M ARB was found, and it is worth stressing that reading balances alone would have distorted it — the Foundation safe's balance ROSE by 243.9M across the window while it was paying 30M out to an exchange. Only the transfer log, read by sender, gives the right answer.
How ARB compares to other Layer 2 governance tokens
Against a hard-capped, halving-model chain like Bitcoin, Arbitrum looks superficially better and behaves worse. Bitcoin issues new supply on a fixed, decaying schedule that anyone can price to the block; Arbitrum issues nothing at all, and its 10B cap is genuinely absolute short of a 2% governance mint that has never been used. But a cap is not a float. Bitcoin's dilution is roughly 0.2% a quarter and falling; Arbitrum's is +6.33% a quarter, because two thirds of its supply started outside the market and is still walking in. Capped and inflationary are not contradictory when the lock, not the mint, is doing the work.
Against the fee-burning Layer 1s — Ethereum with its base-fee burn, or BNB with its quarterly auto-burn — the structural difference is that those chains denominate gas in their own token, so activity mechanically removes supply. Arbitrum denominates gas in ETH. Every transaction Arbitrum processes burns someone else's token. That single design decision is why Arbitrum can post record throughput, real sequencer revenue and a genuine Orbit revenue share and still show a buy side of exactly zero: usage builds the Arbitrum DAO treasury in ETH and stablecoins, and none of it touches ARB. Optimism's OP shares this shape, and both are weaker on the buy side than any chain whose fees are paid in its own unit.
Against exchange tokens that run programmatic buybacks out of revenue, the contrast is starkest of all, and it is a governance contrast rather than an economic one. Arbitrum earns enough to fund a buyback; it has simply never voted for one. The DAO treasury is both the largest overhang and the obvious funding source for the only mechanism that would offset it, which makes ARB a token whose supply outcome is decided in forum threads rather than in code. Until a fee-conversion, burn or buyback proposal actually passes, the Arbitrum ledger has a sell side driven by a calendar and a buy side that is empty by construction.
What to watch in the next 90 days
Sep 16 2026, Oct 16 2026 and Nov 16 2026 each release another 92.65M ARB — that is 277.9M of the forward ledger's 355.5M, and it is the most certain number on this page. Watch the Arbitrum DAO treasury address for the next discretionary release: its history is lumpy rather than steady, with large deployments in Jun 2024, Oct 2024, Feb 2025, Jun 2025 and Jun 2026, so the forward ledger uses the trailing twelve-month net of 33.7M per quarter rather than repeating this window's 100.7M spike tail. Watch whether the Arbitrum Foundation operating safe's standing 10M-a-month exchange transfer changes size — twelve consecutive identical firings make any change in it meaningful. Watch for any Arbitrum governance vote that converts sequencer, Timeboost or Orbit revenue into ARB buybacks or burns; it is the single change that would move this coin from a 0 to a 4 on the inflation score. And watch Mar 16 2027 on the horizon, when the four-year vesting calendar finally ends and Arbitrum's largest recurring sell mechanism simply stops.
Summary
The MrNasdog Pressure Framework reads Arbitrum (ARB) at +6.33% net supply to market over the 90 days to Aug 23 2026, and projects +5.32% for the next 90 days. The structural mechanism is worth stating precisely, because it is easy to misread: Arbitrum has no protocol inflation at all — the Arbitrum One contract has never minted a coin beyond the 10B genesis supply and its counter today sits below it — so every unit of pressure comes from a four-year vesting calendar releasing 92.65M ARB on the 16th of every month, and from the Arbitrum DAO treasury releasing already-minted stock at its own discretion. The key risk is that nothing at all sits on the other side: no buyback, no burn, no lock, because Arbitrum charges gas in ETH and its revenue never becomes an ARB purchase. The ceiling is real and close — 10,000,000,000 ARB is the maximum that will ever exist without a 2% governance mint that has never been called, and 66.8% of it already trades — which means the dilution has an end date, Mar 16 2027, rather than a rate.
MrNasdog Pressure Framework analysis of ARB, Metric 1 — Inflation. Data + explanation only. Not financial advice. Updated Aug 23 2026.
Top comments (0)