For everyone / Fee sharing
Buyback and burn
Two flows fund a buyback of PKL that is then burned. Everything on this page reduces total supply; it does not reduce circulating supply during the vesting years, and that sentence goes wherever the burn does.
The two flows
Two streams of money end up buying PKL and destroying it: a share of the chain's own fees, and a share of what apps pay in. Destroying PKL means there is less of it in total, forever. But for the first few years far more PKL is being unlocked from the team, investors and incentive pools than the buyback destroys, so the amount actually out there keeps going up. Anyone who tells you the burn makes PKL scarcer during those years is leaving that out.
Two flows fund the buyback: 40 percent of net gas revenue, and 60 percent of the ecosystem share collected from applications. The ecosystem share accumulates per epoch and is allocated 60 percent to buyback, 20 percent to treasury, 10 percent to security and 10 percent to incentives. Everything the buyback acquires is burned through the token's holder-only burn, so total supply decreases. During the vesting years, scheduled unlocks add more PKL to circulation than the buyback retires.
The chain-layer flow is the 40 percent buyback share of the net gas split, in ETH. The application-layer flow is the vault's per-epoch allocation of the ecosystem share. Both land in a revenue vault that acquires PKL under the execution rules below and hands it to a burner contract holding only its own balance, which calls the token's holder-only burn and emits the amount, the cumulative total and the epoch. Because the first application bracket is free, the ecosystem share depends on the distribution of application sizes and not only on the aggregate; the whitepaper models it per application and sums, never by applying brackets to an aggregate.
| Ecosystem share | Allocated to |
|---|---|
| 60% | PKL buybackreduces total supply |
| 20% | Treasury |
| 10% | Security |
| 10% | Incentives |
How the buyback executes
The buyback is designed around three constraints. PKL liquidity may be thin, so any mechanism assuming depth would execute badly. Pickle runs a single sequencer, so a predictable on-chain buyback is front-runnable by the operator, which has to be designed around rather than disclosed away. And revenue arrives in many assets, some too small to convert economically.
- Primary mechanism: a sealed-bid, batch-cleared auction. The revenue vault offers a fixed amount of ETH or USDC for PKL. All bids within a window clear at a single price, so the power to order transactions confers no advantage, and there is no market order for anyone to sandwich. If nothing clears above the floor, no PKL is acquired and the funds stay in the vault. The auction sets no price. Filling is refused for the sequencer, batcher and messenger keys, the treasury, the vault, the market-making reserve and every vesting contract, through a timelocked on-chain denylist, and no team, contributor or treasury address trades PKL in a published window around each auction.
- Secondary mechanism: time-weighted execution on the first-party exchange, under hard caps: at most 100 basis points of price impact and 0.5 percent of pool depth per slice, slices timed at random within a published window rather than on a fixed block, execution reverting above 102 percent of a 24-hour on-chain reference, and no execution at all in an epoch below $25,000 of accumulated value. The keeper is paid gas plus a fixed fee, never a proportion of size.
- Small and illiquid fee assets accumulate to a per-asset threshold before conversion under the same caps. An asset still below threshold after four epochs is marked dormant, disclosed and left in place, neither written off nor dumped. Every conversion emits input, output and reference price.
The vault accumulates and does not buy. That period is disclosed as accumulation, and no burn is claimed for it.
What burn means
Burn means total supply decreases. The burner contract holds only its own balance, calls the token's holder-only burn function, and emits the amount, the cumulative total and the epoch. It has no privileged role and no upgrade path. There is no allowance-gated burn anywhere: one would let any approved spender, including routers holding routine unlimited approvals, destroy a holder's balance, which on a token with no mint is irreversible in a way theft is not. Nobody can burn tokens they do not hold, and nobody can create more.
| Word | Meaning |
|---|---|
| Burned | total supply reducedverifiable from the token's supply |
| Removed from circulation | held at an inaccessible address or lockedsupply unchanged |
| Accumulated | held in a vault under governancenot a reduction of anything |
These three words are never used interchangeably on this site.
The float caveat
The sentence that goes with every burn figure
During the vesting years, scheduled unlocks add more PKL to circulation than the modelled buyback retires. The buyback reduces total supply. It must never be described as reducing circulating supply during vesting, and cumulative burn is always published alongside float added over the same period. A buyback denominated in a fixed currency retires more PKL when the price is lower; that is a mechanical property and not a price floor.
The whitepaper publishes the modelled combined buyback per year and calls it a real number and a small one, published because a mechanism whose scale is concealed cannot be evaluated. It adds, unsoftened: the free first bracket means a majority of applications contribute nothing to the buyback. That is a deliberate choice in the builder's favour, and this is its cost. The unlock schedule it is measured against is on the allocation page.