How the treasury health split works

Every evaluation fee a firm collects gets split on-chain, in one transaction, between the firm's treasury and everyone else the protocol owes a share: the operator, DecentralProp, stakers, token buybacks, and the shared safety pool. The treasury health split changes those proportions as the firm's treasury fills up. A firm that has already banked more payout capital than it plausibly needs stops hoarding new revenue in a vault, and the marginal fee flows instead to the people who made the firm successful. Nobody turns this on, nobody tunes it, and no operator can opt out. It's computed live from the firm's on-chain treasury balance on every single purchase.

The problem it solves

Without it, the firm treasury takes the same residual share of every fee, roughly half, whether the firm holds 50 SOL or 10,000. Past a point, that's dead capital: the treasury exists to pay traders, and a treasury that already covers its obligations several times over doesn't protect anyone better by growing further. Meanwhile the operator's cut stays flat, staker yield stays flat, and the Universal Treasury PoolUniversal Treasury PoolA cross-firm shared liquidity pool, funded by a slice of every firm's fees, used only as the last layer of the payout waterfall. that backstops every firm on the protocol grows no faster. The split was solvency-first by design; this mechanism makes it solvency-first and success-rewarding, in that order.

The five zones

A firm's zone is a pure function of two things: its deployment tier and the SOL balance of its on-chain treasury. No stored state, no oracle, no admin instruction. The program recomputes it inside every fee-paying transaction, the same way it already prices the liquidity leg off live curve depth.

The SOL thresholds that separate the zones scale with deployment tier:

Starter, Growth, and Pro share one ladder on purpose. What a firm owes traders tracks the account sizes it sells, not the launch fee its operator paid, so treasury health is judged the same way across those tiers. Scale and Enterprise firms sell into larger obligations and their ladders sit higher.

Note what the measurement is: the firm's own raw SOL treasury, read directly from chain state. It is deliberately not the combined, USD-denominated figure that gates which account sizes a firm may sell (Operating a firm covers that ladder). Two different questions, two different mechanisms: the account-size ladder asks "can this firm safely sell a $500k account," the health split asks "does this firm's vault need the next fee more than its stakeholders do."

What gets redirected, and to whom

Each tier has a maximum redirect, reached at Saturated: 44% of the fee minus the operator's baseline share. That works out to 38% of each fee for Starter firms, 36.5% for Growth, 35% for Pro, 33% for Scale, and 30% for Enterprise. Each zone applies its fraction of that maximum, and the redirected amount splits by fixed weights:

Everything else in the fee is untouched. The liquidity leg, the backstop premium, the affiliate share, the insurance fund, and loss-back accrual all serve functions unrelated to how full the treasury is, and the redirect never draws from them. It comes out of the treasury remainder alone, and the on-chain math guarantees the split still sums to exactly the fee, to the lamport.

A worked example

A Pro-tier firm's operator starts at a 9% share of every fee. Here's the full climb:

In the protocol's own end-to-end test of this exact ladder, run against the real compiled program with real purchase transactions, the treasury's share of each fee fell from 49% in Growing to 17.5% at Saturated while every other destination grew monotonically. The exact percentages depend on the firm's other dynamic legs (curve depth, backstop, referrals), but the shape is the mechanism: the treasury's share only ever falls as zones climb, and it falls to roughly a fifth of the fee, never to zero.

What it means for each stakeholder

Firm operators. Your effective fee share is no longer capped at your tier's baseline. Run a firm well enough to fill its treasury and your share climbs by up to 7 points of the fee at the top tier: Starter saturates at 13.6%, Growth at 14.8%, Pro at 16%, Scale at 17.6%, Enterprise at 20%. There is no application, no toggle, and no way to game it other than the intended one, selling evaluations and honoring payouts until the treasury crosses each line. The operator console's liquidity page shows your current zone, your effective rate, and how much SOL stands between you and the next zone.

Traders. The price of an evaluation doesn't change by a lamport; the split of that price is internal. Your payout protection doesn't weaken either, and that's worth spelling out: the redirect only begins once the treasury already holds 500 SOL or more of payout capital, it never touches the insurance fund or backstop legs, and the treasury keeps receiving the largest single share of every fee even at Saturated. What you gain is indirect but real, since 10% of every redirect flows into the Universal Treasury Pool, the last-resort tier that stands behind every funded payout on the protocol (Funded accounts & payouts).

Stakers. Both staking pools earn more from fees at healthy firms. The no-risk pool's yield leg grows by 15% of whatever the zone redirects, and $DPROP staking yield grows by the same. A staker choosing between firms now has one more reason to read the solvency page: a Thriving firm literally pays its stakers a larger slice of each sale than a Growing one (Staking economics).

Token holders. A fifth of every redirect buys tokens, split evenly between $DPROP and $FIRMA buybacks. Successful firms translate directly into standing buy pressure, scaled by how successful they are.

DecentralProp. The protocol takes 20% of the redirect, on the reasoning that the platform carrying the shared infrastructure and the shared risk should participate in firm success the same way operators do.

Why it's locked

An operator-configurable version of this would be pointless: the whole mechanism is an incentive to capitalize the treasury before enriching anyone, and letting an operator redraw the ladder would let them collect the reward without the capitalization. So the thresholds and weights are platform constants, the same for every firm in a tier, changeable only by protocol governance. This follows the same pattern as the backstop premium, the affiliate rate, and the loss-back stake floor, all of which are fixed platform-wide for the same reason.

The design is also deliberately oracle-free. Zones are measured in raw SOL, not USD, because the fee-paying instruction runs on every purchase and shouldn't depend on a live price feed to work. The tradeoff is disclosed rather than hidden: a large move in SOL's dollar price shifts what a 500 SOL threshold means in USD terms, and the thresholds don't chase it.

Where it stands today

The mechanism is live on devnet and platform-locked on. Its arithmetic is covered by a 60,000-seed conservation fuzz (the split must sum to exactly the fee, every time, including the corner case where an over-configured firm's redirect exceeds what the treasury leg has left, which scales every destination down proportionally rather than breaking the sum). The full five-zone ladder was proven against the real compiled program on a local validator, and the live devnet deployment was proven with a real purchase transaction against a real firm. As of this page's last-verified date, every live firm still sits in Growing, below its first threshold, so the split you'll observe on chain today is the baseline one. The mechanism is waiting for the first firm to earn its way up.

You can check any firm's zone yourself: its public solvency endpoint reports the current zone, the operator's effective rate, and the SOL remaining to the next threshold, alongside the treasury figures that drive it. Verify it yourself covers reading protocol state directly.