# The V4 emission curve: a linear base plus a halving bonus

> Canonical: https://blazephoenix.xyz/learn/the-v4-emission-curve/
> License: CC BY-NC-ND 4.0 (attribution + link) · © 2026 BlazePhoenix
> Updated: 2026-11-16

Each staking deployment schedules 90,000,000 BZPX over eight years: 60,000,000 BZPX at a constant rate plus a 30,000,000 BZPX bonus whose rate halves every two years, and the closed form closes at exactly 90,000,000 with a remainder of 0 wei. It is a calendar of the most a deployment may emit, not a payment.

Short answer: 60,000,000 BZPX of linear base plus 30,000,000 BZPX of bonus, and 60,000,000 + 30,000,000 = 90,000,000, per deployment. The base is paid at a constant rate for all eight years. The bonus runs at a rate that halves every two years. This is a calendar: it states how much the contract is allowed to emit in each window, not how much is paid. A window in which no position holds weight emits nothing, and nobody inherits it.

The staking engine is pre-launch (the site gives Q2 2027), so everything below is the schedule written in the repository, not a record of anything paid. The simpler version of this schedule is in the article on the halving schedule; this one derives it.

## The constants

Five constants fix the whole curve. Their names are in the staking source and the paper (section 25.3) uses the same names.

## The year-by-year table

The base is 60,000,000 / 8 = 7,500,000 BZPX in every year. The bonus is 8,000,000 in years one and two, 4,000,000 in years three and four, 2,000,000 in years five and six and 1,000,000 in years seven and eight. You can check each row with a calculator.

## Why the series closes exactly

The bonus series starts at 32,000,000 and the contract emits 32,000,000 minus 32,000,000 / 16 = 30,000,000 of it, because the schedule stops after the fourth halving and the last step leaves 2,000,000 over. The closed form reads, after the four periods, base + series - (series >> 4), where the shift is an exact division by 16. That is 60,000,000 + 32,000,000 - 2,000,000 = 90,000,000, remainder 0 wei.

The shift is exact because the series, 3.2 x 10^25 wei, is divisible by 2^4. The paper states it as a precondition of the closed form, so a change to the constants that broke it would be a change to the arithmetic, and not an invisible one.

## What integer arithmetic does inside a period

Per-second rates are whole numbers of wei, so each rate is rounded down by less than one wei per second. Inside a period the cumulative curve grows linearly and sits just under the ideal curve. At each of the first three turns the bonus part steps up to its exact value, by 2,176,000, 1,088,000 and 544,000 wei; the base keeps its small shortfall until the close, where the whole curve steps up to exactly 90,000,000 BZPX, by 229,184,000 wei (197,376,000 wei of base and 31,808,000 wei of bonus). The largest of these steps is a few hundred million wei, set against a total of 90,000,000 BZPX. The curve is monotone non-decreasing and it never exceeds the cap.

## From the calendar to a payment

At every entry point the contract pays out the window since the last update, and only that. Written in words: the reward of a window is the scheduled emission of the window, scaled down if the total weight is below a minimum weight, and then capped by what the reward reserve holds. A window with no weight emits nothing, and nothing is carried forward for later stakers.

So the table is an upper bound per deployment, never a promise to a person. What any one position receives depends on its weight against all the others in the same deployment.

## What this does not claim

It does not say that 90,000,000 BZPX will be paid out: it is the most a deployment can schedule. It gives no rate of return, because a share of a pot that depends on everyone else's weight is not a rate you can read from the contract in advance. And it describes the code as written for a launch that has not happened.

**Verify it yourself:** Redo the table by hand from the five constants in the staking repository at commit c0366cf (TOTAL_REWARDS, BASE_REWARDS, HALVING_SERIES, HALVING_PERIOD, EMISSION_LENGTH). The closing sum is pinned by the test test_Dim13_EmissionCurveHalvingAnchors in test/DimensionsExtended.t.sol, and the same close is anchored by tests in test/MetamorphicRelations.t.sol and test/assurance/EmissionEightYear.t.sol. Reading: section 25.3 of the 2.4 whitepaper.

Related: https://blazephoenix.xyz/learn/the-halving-schedule/ · https://blazephoenix.xyz/learn/stochastic-invariants/ · https://blazephoenix.xyz/learn/one-clock-per-chain/
