SIP-3: Emission Schedule

• Status: Accepted (Rob sign-off 2026-09-22 — numbers locked as drafted; changeable only via SIP-1 burn-weight signaling, per “Changing this schedule”)
• Implementation: crates/consensus/src/schedule.rs (normative once wired)
• Author: Sova (orchestrated draft)
• Depends on: SIP-1 (burn recognition), SIP-2 (epochs, ranking, settlement derivation)

Summary

SOVA is minted only by epoch rewards to burners (SIP-2). This SIP locks the reward schedule: a Zcash-style slow start, then a flat per-epoch reward that halves every 1,680,000 epochs — Zcash’s own halving interval, which Sova inherits naturally because one epoch is one Zcash block. Burn-less epochs mint nothing and their rewards are never made up. Long-term security funding transitions to fees: the EIP-1559 basefee is burned, priority fees and the sealer tip pay sealers.

Constants

Constant	Value	Meaning
BASE_EPOCH_REWARD_GWEI	6,250,000,000,000 gwei (6,250 SOVA)	full per-epoch reward, era 0
ERA_EPOCHS	1,680,000	epochs per halving era (~4 years at 75 s; = Zcash’s halving interval)
SLOW_START_EPOCHS	20,000	linear ramp length (~17.4 days at 75 s)
SLOW_START_STEP_GWEI	312,500,000 gwei (0.3125 SOVA)	exact ramp step (BASE / SLOW_START, divides exactly)
sealer tip	1/10 of the epoch reward	SIP-2’s derivation; the fraction is locked here
MIN_BURN_ZAT	1,000	SIP-1’s dust floor, restated for the economics section

All reward arithmetic is in the gwei domain (SIP-2): every value above is integer-exact, no rounding anywhere in the ramp.

The schedule

For 0-based epoch index E (epochs since network genesis; the epoch at Sova height H has index H − 1):

reward(E) = SLOW_START_STEP_GWEI × (E + 1)          if E < 20,000
          = BASE_EPOCH_REWARD_GWEI >> (E / 1,680,000)  otherwise

• Slow start (epochs 0–19,999): the reward ramps linearly from 0.3125 SOVA to the full 6,250 SOVA in exact 0.3125-SOVA steps.
• Eras: from epoch 20,000 to 1,679,999 the reward is flat 6,250 SOVA; each subsequent era halves it (integer gwei floor). Era 42’s reward is 1 gwei; era 43’s is 0 — emission ends (~176 years).
• Asymptotic supply: exactly 20,937,503,124.97144 SOVA — “just under 21 billion” the same way Bitcoin’s cap is just under 21 million. The gap below 21B is the slow-start shortfall (62,496,875 SOVA the ramp deliberately never mints) plus sub-gwei halving dust (~0.03 SOVA across the eras). We do not shift era boundaries to claw it back; the schedule stays a pure function of the epoch index.

The exact asymptote is pinned by a supply-audit test in schedule.rs that sums the whole schedule.

No mint without burns

A burn-less epoch mints zero and the scheduled reward for it is gone — there is no carry-over, no accumulation, no retroactive claim. Consequences, all intended:

• Realized emission is ≤ the schedule and automatically tracks real participation: no demand, no dilution.
• There is no “jackpot epoch” to time: carrying unminted rewards forward would turn the first burn after a quiet stretch into a lottery and make conservation checking stateful. Rejected.
• The supply asymptote above is therefore an upper bound; the honest statement is “at most ~20.94B, minted only against demonstrated demand.”

Why this shape (rationale)

Fixed reward + pro-rata split is the anti-farming mechanism. The epoch reward does not grow with burn volume; more burners slice the same 6,250 SOVA thinner. Farming intensity adjusts the price of a share (in destroyed ZEC), never the supply. Early over-farming self-corrects: it permanently destroys ZEC for a thinner slice.

The slow start closes the worthless-token window. In the first days SOVA has no market and dust burns would otherwise capture full 6,250- SOVA rewards. Ramping the reward makes week-one capture proportionally small while everyone learns the tooling — and it is a direct homage to Zcash’s own launch (Zcash ramped its block subsidy over its first 20,000 blocks for exactly this reason). Sova reuses the number.

Halving at 1,680,000 epochs is inherited, not invented. Sova epochs are Zcash blocks; adopting Zcash’s halving interval means Sova’s eras tick in lockstep with the host chain’s own emission cadence.

The far future is fees. As the subsidy halves away, sealers are paid by (a) the 1/10 sealer tip while subsidies last, and (b) EIP-1559 priority fees. The basefee is burned (standard EIP-1559), so SOVA is deflationary under sustained usage. This is Bitcoin’s fee-transition story with Ethereum’s fee mechanics, on top of a chain whose block cadence never depends on the reward (cadence blocks are produced rewardlessly per SIP-2 — liveness needs nodes, not burns).

Cost floor note (from SIP-1’s economics). At the dust floor, the dominant per-epoch cost of mining is the Zcash ZIP-317 fee (~20,000–25,000 zat for the burn tx shape), not the burn itself (1,000 zat minimum). Keeping a miner alive every epoch costs pennies per day and is dominated by fees paid to Zcash miners — a spam floor for Sova and a small ongoing gift to the host chain’s security budget.

Changing this schedule

Only by soft fork via SIP-1 burn-weight signaling (the version-bits tally machinery, workstream C7). No admin keys, no discretionary issuance, no council.

Implementation status

crates/consensus/src/schedule.rs implements epoch_reward_gwei(E) with boundary and supply-audit tests, and C8 wired it through the node: the Schedule type feeds the sealer, the expectations follower, and the validator from one construction site (SOVA_EMISSION_SCHEDULE=sip3 selects this schedule; the default stays flat for regtest and the box scenarios’ deterministic exact-mint assertions). The sip3 mode’s live debut is mainnet genesis: the public testnet mints a flat 6,250 SOVA per epoch (Rob, 2026-09-23), so testers hold usable balances from the first epoch.