CRV Inflation Analysis · September 2026 · Supply growing, projected to keep growing
Curve DAO added 32.48M CRV to the tradable float over the last 90 days and removed exactly none. The Curve gauge emission minted 27.62M CRV, expiring four-year veCRV vote-locks returned 3.93M CRV, and a 2020 vesting escrow that every unlock tracker lists as finished paid out another 0.93M CRV. Against that, Curve runs no buyback, no burn, and no foundation bid — the buy side of the Pressure Framework ledger is zero on all four rows. The framework reads CRV at +2.09% net over the trailing window and +1.85% forward, against a supply-monitor reading of +1.97%. The Curve emission is genuinely shrinking — it stepped down 15.91% on Aug 12 2026 — but it is the only thing on this coin that is.
The verdict, in one paragraph
Over the last 90 days the MrNasdog Pressure Framework reads Curve DAO at +2.09% net: 32.48M CRV reaching the float against 0 CRV removed, on a circulating base of 1,554.03M CRV. The supply monitor reads the same window at +1.97%, a gap of 0.12 percentage points — comfortably inside tolerance, so this build ships no monitor-gap chip and needed no reconciliation walk. The two readings agree because the ledger and the classifier are measuring the same thing: subtracting the veCRV vote-escrow balance and the two live vesting escrows from CRV total supply reproduces the classified circulating figure to 0.0055%. Forward, the framework reads +1.85%, and that improvement comes entirely from the emission schedule rather than from anything new on the buy side. CRV is structurally inflationary with a decaying mint and no offset — a hard-ceilinged governance token whose issuance curve is honest arithmetic and whose demand side is deliberately left empty.
Sell pressure: where new CRV comes from
Sell #1 — protocol inflation — is 27.62M CRV, and it is by far the largest tap on Curve DAO. The CRV gauge emission is not a policy; it is arithmetic written into the ERC20CRV token contract in 2020 and running to 2049. The contract holds a per-second issuance rate and divides it by the fourth root of two once a year, which is a cut of 15.91% every August. That cut landed inside this window: Curve entered its sixth mining epoch on Aug 12 2026 and daily CRV issuance fell from 316,563 to 266,197 CRV, taking scheduled annual issuance from about 115.5M to 97.2M CRV — under a hundred million a year for the first time in Curve's history. This build read that rate off the contract at both ends of the window rather than off a documentation page, because a docs figure is usually the previous epoch. Because the change is dated and documented, the forward column is re-based onto the post-cut rate instead of the blended trailing average, which is 23.96M CRV for the next 90 days. One detail matters for anyone reconciling the number: CRV is only created when a Curve liquidity gauge claims it, so minting runs behind the schedule and then catches up. Measured issuance came in at 101.01% of what the schedule released, drawing down a mintable-but-unminted backlog that still stands at 10.30M CRV.
Sell #2 — vesting unlocks — is 0.93M CRV, and it is the row most trackers report as a zero. Curve's 2020 launch set aside just over five percent of supply for pre-CRV liquidity providers in an on-chain vesting escrow with a one-year schedule that expired in Aug 2021. On the calendar, therefore, CRV is fully unlocked and this window's scheduled entitlement is nothing. On the chain it is not nothing, because vested CRV only leaves the escrow when a beneficiary claims it, and five years later people are still claiming. The escrow balance fell from 11.23M to 10.30M CRV across the window, so 0.93M CRV genuinely reached the market. The framework books the realised outflow rather than the calendar entitlement, and the 10.30M CRV still sitting unclaimed behind it becomes an overhang rather than a forecast. The Curve investor and employee vesting escrows, by contrast, read empty at both ends of the window; those allocations are structurally spent and cannot contribute again.
Sell #3 — foundation and unscheduled unlocks — is zero, and it is a measured zero rather than an assumed one. The Curve DAO community fund escrow held 845.1K CRV at both ends of the window, identical to the last decimal, and the DAO's ownership, parameter and emergency agent wallets hold no CRV whatsoever. Governance confirms it from the other direction: a proposal to allocate 17.4M CRV to Curve's core development company was voted down 54.46% in Aug 2026, with the two largest vote-locked holders supplying most of the opposition. Sell #4 — long-term locked — is 3.93M CRV, and it is the tap no unlock calendar reports at all. Curve's veCRV vote-escrow lets holders lock CRV for up to four years in exchange for voting weight and fee revenue, and when a lock expires the principal comes back with no announcement and no schedule. The veCRV contract held 854.05M CRV at the start of this window and 850.12M CRV at the end, so 3.93M CRV came out of the lock and back into the float.
Buy pressure: where new CRV goes
Every buy row on Curve DAO is zero, and only one of them is zero by accident. Buy #1 — programmatic buyback — is zero because Curve has chosen not to have one. Curve's argument is that veCRV vote-locking already performs the function a buyback would perform: it removes CRV from the tradable float and pays the locker for doing it. Admin fees from Curve pools and from crvUSD borrowing are collected weekly, converted into a stablecoin, and distributed to vote-lockers in that stablecoin, so protocol revenue never converts back into CRV at any point in the pipeline. Buy #2 — protocol fee burn — is zero, checked from both directions this window because a burn is often invisible on one of them. The dead address held the same 169.86 CRV at both ends, so nothing was transferred to a sink; and CRV total supply rose rather than fell, so nothing was destroyed by a supply-reducing call either. Neither surface moved, which is what a coin with no burn path actually looks like.
Buy #3 — foundation buy — is zero for the same structural reason as Buy #1: Curve DAO revenue is denominated in a stablecoin and stays that way, and there is no treasury mandate to buy CRV on the open market. Buy #4 — new long-term lock — is where the framework has to be careful, because the veCRV lock is a genuine multi-year lock and it is tempting to book the 850.12M CRV sitting inside it as buy pressure. That would be wrong twice over. A standing balance is capacity, not a flow, and the flow this window went the other way: the veCRV balance fell. That fall is already counted on the sell side as Sell #4, so booking it again here would double it. New Curve vote-locks are still being created every day — just fewer of them than are expiring. Buy #4 is zero by measurement, and if the veCRV balance rises in a future window this row is where that will appear.
Foundation and overhang
Curve DAO's team-controlled overhang is unusually small for a token of this age, and it is fully enumerable on-chain. The first item is the Curve community fund vesting escrow, which held 845.1K CRV at both ends of the window; it has been drawn down heavily over the years from a starting balance north of 26M CRV, but nothing moved in this window and any future draw carries a mandatory one-year re-vest. The second is the undrawn remainder of the 2020 pre-CRV liquidity escrow, 10.30M CRV, which is fully vested and claimable at any moment by its original beneficiaries — it is the reason Sell #2 is not structurally zero. The third is the mintable backlog: 10.30M CRVthat the emission schedule has already released but that no Curve gauge has claimed yet, which is why measured issuance can run slightly above schedule. The Curve DAO's own governance wallets hold no CRV, and the investor and employee escrows are empty. If any of these balances falls between refreshes, the outflow enters Sell #3 at the next refresh; the vesting escrow drain is already booked in Sell #2 and will keep being measured there.
How CRV compares to other vote-escrow DeFi tokens
CRV is the origin of the vote-escrow model, and almost every comparison worth making is a comparison to a token that copied it. The mechanism is the same everywhere: lock the token for years, receive voting weight, direct emissions to the pools you care about. The axis on which these tokens differ is what happens to the revenue that the voting weight controls. Curve pays it out in a stablecoin. A large and growing share of the vote-escrow cohort — the perpetual venues and the newer decentralised exchanges — routes the same revenue into open-market buybacks instead, which converts fee income directly into a bid for the token. That single design choice is the whole difference between a buy side of zero and a buy side that can outrun issuance, and it is why CRV scores where it does despite having one of the most disciplined issuance curves in DeFi.
Against uncapped continuous-emission tokens, CRV looks strong: its ceiling of 3.03B CRV is protocol-encoded, its rate is a public constant anyone can read out of the contract, and that rate falls 15.91%every year on a fixed date with no vote required. There is no discretionary emission committee and no governance lever that can accelerate it. Against fee-burning exchange tokens, CRV looks weak for the mirror-image reason: those tokens destroy supply out of revenue every quarter and can print a negative net number, while CRV structurally cannot go below zero net until either the emission ends or a buy mechanism is voted in. And against the halving-model chains that hard-cap supply, the comparison is closer than it appears — CRV's fourth-root-of-two annual reduction is a gentler, more frequent halving, but the tail runs to 2049 rather than 2140, and unlike a proof-of-work chain CRV has two additional supply taps, the vote-lock unwind and the escrow drain, that no issuance schedule describes.
What to watch in the next 90 days
The first thing to watch is the veCRV balance, because it is the largest swing factor and the least predictable one. It fell 3.93M CRV this window; if the trend reverses and locking outruns expiry, Sell #4 goes to zero and Buy #4 goes positive, which would take the forward reading close to +1.5%. The second is the pending Curve governance proposal to raise the DAO's share of protocol fees from 10% to 30%: as written it is a stablecoin revenue change and touches no CRV row, but a larger DAO revenue pool is exactly the precondition for a buyback proposal, and a buyback is the only thing that could move CRV's buy side off zero. The third is the pre-CRV escrow drain rate, which has been steady but is finite — 10.30M CRV remains, and if claiming accelerates Sell #2 grows before it disappears for good. The fourth is the next scheduled emission step-down, due in Aug 2027, which is outside this window and cannot help the forward number. There is no CRV unlock cliff of any kind between now and Nov 2026.
Summary
The MrNasdog Pressure Framework reads Curve DAO at +2.09% net over the trailing 90 days and +1.85% forward, against a supply-monitor reading of +1.97% — a gap of 0.12 percentage points and no data conflict. The structural mechanism is a contract-encoded gauge emission of 27.62M CRV a quarter that cuts itself 15.91% every August, supplemented by two quieter taps that no unlock calendar reports: 3.93M CRV released by expiring veCRV vote-locks and 0.93M CRV claimed out of a 2020 vesting escrow. The key risk is that the buy side is not merely small but structurally absent — Curve runs no buyback, no burn and no treasury bid, so CRV cannot print a negative net figure no matter how good its revenue quarter is. The ceiling is real at 3.03B CRV and the schedule is honest, but a shrinking mint against a buy side of zero is still a growing supply.
MrNasdog Pressure Framework analysis of CRV, Metric 1 — Inflation. Data + explanation only. Not financial advice. Updated Sep 5 2026.