Originally published at mrnasdog.com/research/ray/inflation
Raydium paid out about 0.43M RAY over 90 days from a fixed mining reserve, while the Raydium buyback funded by 12% of every pool's trading fee took 2.60M RAY off the open market and parked it in a public Solana wallet. RAY minting is permanently switched off on-chain, so the buyback outpaces the emission roughly six to one and the Pressure Framework reads −0.81% net on the active float. Our supply monitor reads +0.46% because it still counts the held buyback wallet as circulating RAY — a 1.27-point gap that is mechanical, not an error.
The verdict, in one paragraph
For the 90-day window ending Aug 25 2026, the MrNasdog Pressure Framework reads RAY at −0.81% net: 0.43M RAY of sell pressure against 2.60M RAY of buy pressure on a circulating base of 269.51M RAY. Our supply monitor reads the realised change over the same 90 days at +0.46%, a gap of 1.27 percentage points. That exceeds the framework's half-point tolerance, so a monitor-gap chip ships on the RAY overview. The gap resolves cleanly and arithmetically: the Raydium buyback does not burn RAY, it accumulates it in a public wallet that the monitor's upstream classification still treats as circulating supply. The 2.60M RAY absorbed is worth 0.97 points of the gap on its own, and the rest is a single-day wobble in the monitor's market-cap-over-price supply estimate. Raydium is best read as structurally deflationary on the active float, with the held buyback stack sitting behind it as a governance-reversible overhang.
Sell pressure: where new RAY comes from
Sell #1, protocol inflation, is the only live sell row on RAY and it is small: 0.43M RAY over 90 days. Raydium's RAY mint authority is renounced — a direct on-chain read of the RAY mint returns a null mint authority and a null freeze authority — so no new RAY can ever be created. On-chain supply sits at 554.998M RAY against a 555M hard cap and cannot rise. What reads as inflation is the fixed mining reserve paying already-issued RAY out as staking and farm rewards. This build measured that flow rather than assuming it: every reward vault the reserve funds was read at both ends of the window, and the reserve's own emission wallet released exactly 200,000 RAY on Jun 24 2026 and another 200,000 RAY on Aug 10 2026 into the RAY staking vault, plus 136,000 RAY into four short concentrated-liquidity farm campaigns. Netting off what is still sitting undistributed inside those vaults, 431,000 RAY actually reached holders.
Two independent cross-checks bracket that figure. The live protocol parameter on the RAY staking pool pays 0.0418 RAY per second, or about 1.32M RAY a year, and Raydium's own documentation quotes approximately 1.9M RAY a year from the mining reserve. Over 90 days those imply 0.33M and 0.47M respectively — the realised read sits between them, which is the shape you want when a vesting-style reserve drips rather than fires. Emission is thin but genuinely live: the staking vault was refilled on Aug 10 2026 and still holds roughly five weeks of rewards.
The other three sell rows are zero, and each for a structural reason. Sell #2, vesting unlocks, is zero because Raydium's team and seed vesting concluded on Feb 21 2024; RAY is fully unlocked and no cliff of any size falls inside the window. Sell #3, foundation and unscheduled unlocks, is zero because no release from any project reserve into the market was observed — the reserve wallet only moved to fund reward vaults, which is already booked in Sell #1. Sell #4 is zero because RAY has no bankruptcy estate and no trustee release schedule attached to it.
Buy pressure: where new RAY goes
Buy #1, the programmatic Raydium buyback, is the whole story on the buy side: 2.60M RAY over 90 days. The mechanism is fixed in the protocol's fee split — concentrated-liquidity and constant-product pools send 84% of fees to liquidity providers, 12% to the RAY buyback and 4% to the treasury, while the legacy AMM v4 pools send 88% to liquidity providers and the same 12% to the buyback. The framework does not model that split into a number; it reads the destination wallet directly at both ends of the window across the six RAY accounts that wallet owns, and the balance rose from 81.90M RAY to 84.51M RAY with not a single outgoing transaction on any account.
The destination matters more than the size, and it is the most commonly mis-reported fact about RAY. Bought-back RAY is held, not burned. Raydium's documentation names the accumulation address, the chain confirms the balance, and the on-chain supply figure confirms the negative: about 2,400 RAY have been destroyed in the token's entire history, against a buyback stack now worth 84.51M RAY. Media copy describing Raydium as "buy and burn" is wrong. That is why Buy #2, protocol fee burn, is zero: RAY has no burn path at all, and the buyback parks rather than destroys. Buy #3, foundation buy, is zero because the fee-funded buyback is the only protocol buying and no separate treasury purchase was observed. Buy #4, new long-term lock, is zero because no new lockup contract or staking cap was announced and RAY staking can be withdrawn at will.
Foundation and overhang
Three team-controlled overhangs sit behind the RAY float and all three are tracked. The largest is the buyback accumulation wallet, now holding 84.51M RAY — roughly 31% of circulating supply — spread over six token accounts, every one of which only grew across the window. Raydium states that the eventual use of the accumulated RAY rests with governance, which is precisely why it is an overhang and not a burn. The second is the mining-reserve emission wallet, holding 0.58M RAY at the moment of this build; it is the hopper that tops the reward vaults up, and its outflow is what Sell #1 measures. The third is the residual project allocation — 200.40M RAY of reserve and ecosystem supply that is not circulating and has no published release plan beyond the reward drip.
All three are read from the chain on every rebuild, and the rule is the same for each: if any of these balances falls between refreshes, the outflow enters Sell #3 at the next refresh. That is the single event that would flip RAY's reading, because it is the only way a hard-capped, mint-renounced token can still surprise on the supply side.
How RAY compares to other exchange-fee tokens
RAY belongs to the fee-funded-buyback class rather than the emission class, and inside that class the split that matters is burn versus hold. Tokens whose buyback ends at a burn address — the BNB auto-burn model, or an EIP-1559-style base-fee burn — retire supply permanently, so the reduction is irreversible and every supply tracker agrees on the number. Raydium sits on the other side: the 12% fee slice buys RAY on the open market and then holds it. Economically the effect on the tradable float is the same in the moment, but it is reversible by governance, and it creates exactly the reporting divergence this page carries — a supply tracker counts held tokens as circulating, while the Pressure Framework counts them as off the market.
Against the other Solana decentralised exchanges, RAY's distinguishing feature is that its supply question is essentially closed. It is hard-capped at 555M, fully unlocked since Feb 21 2024, and mint-renounced on-chain — three properties most of its Solana peers do not hold simultaneously, since several still run multi-year unlock calendars or retain a live mint authority. Compared with uncapped continuous-emission layer-1 chains, where staking issuance is a permanent structural sell row, RAY's emission is a finite reserve draining at roughly 1.3M to 1.9M RAY a year against a buyback currently running near 10M RAY a year. The risk on RAY is therefore not dilution; it is fee revenue, because the buyback is only as large as trading volume makes it.
What to watch in the next 90 days
First, the buyback run rate: a research desk logged roughly $150,000 of RAY bought in the three days to Aug 8 2026 and called it the fastest pace of 2026, so the trailing figure used here is a conservative one — if Solana trading volume falls back, Buy #1 shrinks with it. Second, any governance decision on the 84.51M RAY held in the buyback wallet; a decision to burn it would remove the monitor gap outright, while a decision to deploy it would open Sell #3 for the first time. Third, the reward-vault refill rhythm — the next top-up after Aug 10 2026 tells you whether the mining reserve is still paying at the same rate. Fourth, new farm campaigns: four short concentrated-liquidity campaigns funded inside this window added about 79,000 RAY to the sell side, and a heavier campaign season would lift Sell #1. Fifth, product changes that shift where fees are collected, such as the Aug 17 2026 LaunchLab quote-asset expansion, which have no direct supply effect but do move the revenue that funds the buyback.
Summary
The MrNasdog Pressure Framework reads Raydium's RAY at −0.81% net supply over 90 days, and projects the same for the next 90: a thin 0.43M RAY mining-reserve emission against a 2.60M RAY fee-funded buyback, on a 269.51M RAY float that is hard-capped at 555M and can never grow, because the mint authority is renounced on-chain. The structural mechanism is the 12% trading-fee buyback, and the key risk is that it is a hold, not a burn: 84.51M RAY — about 31% of circulating supply — now sits in a protocol wallet whose future use is a governance decision, and that same held stack is why our supply monitor reads +0.46% where the framework reads −0.81%. The ceiling is fixed and the emission is finite; the variable is fee revenue, and with it the size of the buyback.
MrNasdog Pressure Framework analysis of RAY, Metric 1 — Inflation. Data + explanation only. Not financial advice. Updated Aug 25 2026.
Top comments (0)