Fee flywheel
Protocol fees split into two opposite-eligibility pools so that if everyone hides in tun, activity yield spikes and pulls specimens back awake — no admin required.
This is the most important design decision in the protocol, and it exists to answer one specific failure mode.
The failure mode: if tun is strictly better than hydrated, every rational holder enters tun. Nobody trades. Vault volume goes to zero. Fees go to zero. Distributions go to zero. The reward mechanic eats the revenue that funds it. A single-pool design walks into this within two months.
The fix: split the distribution into two pools with opposite eligibility.
Protocol fee revenue
│
├─ 70% → swapped into tokenized stock, then split:
│ ├─ 35% CRYPTOBIOSIS POOL → tun specimens, weighted by tun tier
│ └─ 35% ACTIVITY POOL → hydrated specimens that participated
│ this epoch (vault trade, event entry,
│ loan open/close)
│
├─ 20% → $DSUP buyback → treasury (NOT burned, see below)
│
└─ 10% → operations
Why this self-regulates. Both pools are fixed fractions of revenue, but the number of claimants in each floats. If 80% of the collection sits in tun, the Cryptobiosis pool splits 35% across 960 specimens while the Activity pool splits an equal 35% across the ~240 that stayed awake. Per-capita activity yield goes up 4×, and specimens start waking. If everyone stays hydrated and churns, tun becomes the underserved side and the pull reverses.
The system finds its own equilibrium without an admin ever touching a parameter. That is the entire point.
Delivery. Distributions land directly in each specimen's TBA. No claim transaction, no merkle proof, no gas from the holder.
Cadence. Weekly epoch, settled by a permissionless distribute() that anyone can call once the epoch closes. A small caller bounty comes out of the ops slice so it always gets called, by someone, forever, without us.
Buyback, not burn. Burning is theatre at this size. Treasury $DSUP is the only strategic asset the protocol has — it funds liquidity pairing for incubated projects (§9). Burning it means throwing away the balance sheet to produce a chart annotation.
Reflexivity, stated plainly, on every page that shows a number: distributions are funded by protocol fee revenue. Volume dies, distributions die. Week-two yield is not month-three yield. These are rewards-program airdrops — not dividends, not equity, not shareholder rights.
The 4× figure in the equilibrium example is exact for that split: 35% of revenue divided across 240 activity-pool claimants versus the same 35% divided across 960 in a fully-tun scenario is a 4× per-capita gap at that specific 80/20 split, and it scales with whatever the actual hydrated/tun ratio is at settlement — it is not a fixed multiplier, it's the mechanism itself. The 70/20/10 top-level split and the 35/35 sub-split are both parameters, not derived constants; they are the two numbers most likely to move before mainnet if the seed budget in §3 or the ops cost model changes. Weight math and split percentages are frozen and tested before mint per §11 — not patched after, precisely because this is the parameter everyone will want to lobby on once real revenue exists.