Update (2026-04-10): This post is a historical snapshot from 2026-03-14 covering the switch from 50 UDAG/round to 1 UDAG/round. Since then the distribution model was overhauled again:
- 7-bucket model, not 4-bucket: the per-round split is now 44% validators / 10% council / 16% treasury / 5% founder / 8% ecosystem / 5% reserve = 88% of the nominal reward. The remaining 12% is never minted — it’s offset by a 2,520,000 UDAG IDO genesis pre-mine that bootstraps the private round and Uniswap liquidity. Total supply cap is unchanged at 21M.
- Only one pre-mine: the IDO distributor address. All other buckets (validators, council, treasury, founder, ecosystem, reserve) still start at zero and earn through per-round emission.
- Per-validator emission table:
MAX_ACTIVE_VALIDATORSis 100, and the table below has been recomputed with the 44% validator share of the nominal reward.- “Year 5.7 supply cap”: incorrect. The geometric series converges over tens of halvings until the per-round reward integer-shifts to zero (around halving 27, ~45 years at 5 s/round). The
block_reward()function returns 0 forhalvings >= 64.See tokenomics/supply for the current canonical numbers.
UltraDAG inherited Bitcoin’s halving constants: 50 coins per block, halving every 210,000 blocks, 21 million max supply. Bitcoin produces blocks every 10 minutes. UltraDAG produces rounds every 5 seconds.
That difference — a factor of 120 — meant the entire emission schedule would complete in three months instead of a hundred years. We caught it before mainnet launch and redesigned the schedule from first principles.
The Constraint
Bitcoin’s emission has an elegant identity that most people don’t think about:
initial_reward x halving_interval x 2 = max_supply
50 BTC x 210,000 blocks x 2 = 21,000,000 BTCThe factor of 2 comes from the geometric series 1 + 1/2 + 1/4 + ... = 2. This means the three constants — initial reward, halving interval, and max supply — are mathematically coupled. You can’t change one without adjusting another.
UltraDAG kept 21M max supply (the right choice for brand clarity). It kept 210,000 as the halving interval (wrong — this is a block count, not a time duration). And it kept 50 as the initial reward (wrong — this was sized for 10-minute blocks).
Two Bugs, Not One
Bug 1: Per-Validator Instead of Per-Round
Pre-Staking Emission Bug: Before staking is active, the pre-staking emission model should split block_reward equally among all validators. Instead, each validator received the full reward. With 5 validators: 250 UDAG/round instead of 50.
Root cause: the validator count used for splitting was derived from DAG vertex counts in the previous round, which was 0 at startup. max(0, 1) = 1, so each validator got the undivided reward.
Fix: Added configured_validator_count field to StateEngine, set from the --validators N CLI flag. Both engine and validator loop now use this deterministic value instead of counting vertices in finality batches.
Bug 2: The Schedule Itself
Even after fixing the per-validator split, the emission was still far too fast. At 50 UDAG/round with 5-second rounds:
- 17,280 rounds/day x 50 UDAG = 864,000 UDAG/day
- First halving at 210,000 rounds = 12.15 days
- 60% of all mining rewards emitted in under two weeks
- Supply effectively exhausted within 3 months
For comparison, Bitcoin’s first halving took 4 years. Most serious L1s measure halvings in years, not days.
The Redesign
We fixed max_supply at 21M and solved for reward and interval combinations that produce credible timelines:
| Reward/Round | Halving Interval | First Halving |
|---|---|---|
| 50 UDAG | 210,000 | 12 days |
| 5 UDAG | 2,100,000 | 4 months |
| 2 UDAG | 5,250,000 | 10 months |
| 1 UDAG | 10,500,000 | ~1.66 years |
We chose 1 UDAG per round with a 10,500,000-round halving interval. The identity holds:
1 UDAG x 10,500,000 rounds x 2 = 21,000,000 UDAGThe New Schedule
Genesis allocates 2,050,000 UDAG (faucet reserve + developer allocation), leaving 18,950,000 UDAG for mining. The supply cap in StateEngine enforces the 21M ceiling — once cumulative mining + genesis reaches 21M, block rewards drop to zero regardless of the halving schedule.
| Period | Reward/Round | Mined | Cumulative | % of 21M |
|---|---|---|---|---|
| Year 0-1.7 | 1.0000 UDAG | 10,500,000 | 12,550,000 | 59.8% |
| Year 1.7-3.3 | 0.5000 UDAG | 5,250,000 | 17,800,000 | 84.8% |
| Year 3.3-5.0 | 0.2500 UDAG | 2,625,000 | 20,425,000 | 97.3% |
| Year 5.0-5.7 | 0.1250 UDAG | 575,000 | 21,000,000 | 100% |
Supply cap reached at approximately year 5.7. After that, validators earn only from transaction fees.
Validator Economics
Daily nominal reward is 17,280 UDAG (one per round × 17,280 rounds per day). The validator pool is 44% of that — 7,603.2 UDAG/day. Split among validators:
| Validators | UDAG/Day Each | UDAG/Year Each |
|---|---|---|
| 5 (current testnet) | 1,520 | 555,000 |
| 10 | 760 | 277,500 |
| 25 | 304 | 111,000 |
| 100 (max active) | 76 | 27,750 |
Math: 86,400 s/day ÷ 5 s/round = 17,280 rounds/day. Validator pool per day = 17,280 × 0.44 = 7,603.2 UDAG. Divide equally by N validators for per-node income, before any delegation effects or commission. At the full 100-validator cap each node earns ~76 UDAG/day in the first halving period. The remaining 44% of the nominal reward is split 10% council + 16% treasury + 5% founder + 8% ecosystem + 5% reserve; the final 12% is uncreated, matching the IDO genesis pre-mine.
Historical note: this table was updated on 2026-04-10 when the distribution model changed from the 75% validator / 10% council / 10% treasury / 5% founder split to the current 7-bucket model. Earlier versions of this post showed higher per-validator daily numbers based on the 75% validator share.
Why Not Exactly Bitcoin’s Timeline?
Bitcoin’s 4-year halvings at 5-second rounds would require an initial reward of ~0.033 UDAG per round. That’s too small to meaningfully reward early validators, especially when split across up to 100 active nodes. The daily emission per validator would be a fraction of a UDAG — not enough to matter.
1.66-year halvings hit the sweet spot: aggressive enough to reward early participants, conservative enough that the schedule extends over 5+ years of meaningful emission. The supply cap provides a hard ceiling regardless.
What Changed
// Before
INITIAL_REWARD_SATS = 50 * COIN // 50 UDAG per round
HALVING_INTERVAL = 210,000 // ~12 days at 5s rounds
// After
INITIAL_REWARD_SATS = 1 * COIN // 1 UDAG per round
HALVING_INTERVAL = 10,500,000 // ~1.66 years at 5s roundsTwo constants changed. The rest of the system — supply cap enforcement, stake-proportional distribution, observer penalties, halving logic — was already correct. The fix was purely parametric.
The emission schedule is enforced at three levels:
block_reward(height)computes the halved reward,apply_vertex_with_validators()caps coinbase atMAX_SUPPLY - total_supply, and the supply invariant (liquid + staked == total_supply) is checked unconditionally in release builds.