Files
tokenomics/RECONCILIATION.md
zeekayandHanzo Dev 21ea009c3f docs: record the 2026-07-10 mainnet 96369 genesis reset
Mainnet 96369 was reset to a fresh genesis on 2026-07-10 to recover from a
C-Chain warpConfig VM-init bug (not consensus). The 118 Genesis NFTs on
0x004287C4... and all post-genesis state were wiped; the treasury 0x9011 2T is
preserved (genesis allocation). Re-mint + the ~990B tokenomics convergence are
deferred to after the v1.36 consensus rip-out so the app layer redeploys once.

- migration/HISTORY.md: authoritative 2026-07-10 reset entry; prior 118-token
  state marked superseded (now the re-mint spec); process/version note updated.
- RECONCILIATION.md / cchain/GENESIS-NFTS.md / README.md / LLM.md: pointer
  banners so no reader quotes the pre-reset "current" figures as live state.

Canonical 2T distribution (SUPPLY.md) and the ETH->C-Chain migration record are
unchanged and remain the spec. Root cause writeup:
~/work/lux/consensus/docs/postmortems/tendermint-accretion.md

Co-authored-by: Hanzo Dev <[email protected]>
2026-07-09 23:43:35 -07:00

8.4 KiB
Raw Permalink Blame History

Reconciliation — On-Chain Today vs Canonical Target

The honest gap report: what the chain holds now versus the canonical 2T tokenomics, and the exact deltas required to converge. Read this before quoting any "current" number — the canonical tables in SUPPLY.md are the target, not the present on-chain state.

Every ⚠️ here is an open item for the owner to resolve.

⚠️ POST-RESET (2026-07-10) — the on-chain "current state" figures below are WIPED. Mainnet 96369 was reset to a fresh genesis on 2026-07-10 to recover from a C-Chain warpConfig VM-init bug (not consensus). The 118 NFTs and all post-genesis state are gone; the treasury 0x9011 2T is preserved (it is a genesis allocation). So the "118 tokens", "100/100 validators", and "~1.99T in 0x9011" numbers below describe the pre-reset chain and now serve as the re-mint / convergence target, not live state. The re-mint and the ~990B redistribution are deferred to after the v1.36 consensus rip-out so the app layer is redeployed once. Timeline of record: migration/HISTORY.md (2026-07-10 entry). Root cause: ~/work/lux/consensus/docs/postmortems/tendermint-accretion.md.


1. Supply-level reconciliation

Total supply is correct and fixed: 2,000,000,000,000 LUX (2T), reached after the Dec-2025 fork burned 100B (2.1T → 2T). The open gap is distribution, not supply.

Bucket (target) Target LUX Held where today Gap
Lux DAO 1,000,000,000,000 DAO wallet 0x9011 — (holds far more, see below)
Lux NFTs 500,000,000,000 notional; LUX still in DAO wallet move ~500B
Team / Devs / Treasury / Rewards / Partners 500,000,000,000 notional; LUX still in DAO wallet move ~500B

The DAO treasury wallet 0x9011E888251AB053B7bD1cdB598Db4f9DEd94714 holds ~1,990,000,000,000 LUX (≈1.99T) today — essentially the whole supply. The canonical Lux DAO bucket is 1.00T, so ~990B must be redistributed out of the treasury into the other buckets (Public Sale, vesting, program wallets, NFT bonds).

 DAO wallet 0x9011 today:  ~1.99T  ████████████████████░  (of 2T)
 Canonical "Lux DAO":       1.00T  ██████████░░░░░░░░░░░
 To redistribute:          ~0.99T  ░░░░░░░░░░██████████   → the other buckets
  • 1.99T − 1.00T ≈ 990B to move. (Owner cited ~995B; the two agree to within the precision of the live balance.)
  • ⚠️ TO CONFIRM — exact treasury balance. The redistribution figure depends on the precise live balance of 0x9011 (and how much LUX is already circulating outside it — on today's figures, only ~10B). Snapshot the exact balance to fix the number.

Why the NFT bonds don't show as "moved." The GenesisNFTs contract records totalLuxLocked when it mints, but locks no LUX physically — the backing LUX stays in the DAO wallet. So the 500B NFT allocation is currently notional. Physically funding it (escrowing bonds, standing up the sale/vesting/program contracts) is the convergence work below.


2. NFT collection minting status

Family Minted Authorized / target Status
Genesis Validators 100 100 ✅ complete (#1–#100)
Coins 18 11,111 🔸 0.16% minted (11,093 to go)
ATM / Card 0 1,000 🔸 none minted
Pass (LUX) 0 ⚠️ ~100,000 🔸 not in contract
Validator hierarchy (Validator/Mini/Nano/Mobile) 0 designed 🔸 not authorized on ETH; design-only
  • Validators: 100 / 100 ✅ — the only fully-complete family.
  • Coins: 18 / 11,111 — the 18 original mints migrated; the remaining 11,093 are authorized-but-unminted, pending an owner-gated mint.
  • ATM/Card: 0 / 1,000 — none minted on either chain.
  • 12,093 authorized-but-unminted total (11,093 Coins + 1,000 ATM/Card).

Bonds recorded vs if fully minted

Recorded now (118 tokens) If all authorized minted
Genesis Validators 100,000,000,000 (100 × 1B) 100,000,000,000
Coins 2,583,000,000 (18 tokens) 50,000,000,000 (11,111)
ATM / Card 0 200,000,000,000 (1,000 × 200M ⚠️)
totalLuxLocked 102,583,000,000 350,000,000,000

The "if fully minted" 350B is the product-view bond sum for the three on-chain families. Its relationship to the 500B allocation (which also budgets Cash Rewards 238.889B and the Pass, and only 11.111B to "Validators + Pass") is the central unreconciled item — see §4 and NFT-COLLECTION.md.


3. Private Sale — void

  • Canonical bucket 8, 5% / 100,000,000,000 LUX: never happened. No private-sale tokens were sold; there are no private investors.
  • On-chain effect: none — nothing was ever distributed for it, so there is no allocation to unwind. The 100B is simply part of the ~1.99T still in the DAO wallet.
  • ⚠️ Owner decision pending: whether the 100B (a) stays with the DAO, (b) enlarges the public NFT allocation (the allocation view folds it into the 500B Lux NFTs bucket), or (c) is redirected. Until decided, treat as DAO-retained.

4. Open reconciliation items

Everything the owner needs to resolve, in one place:

# Item Where Resolution needed
1 FDV $220M vs $0.0001×2T = $200M (10% delta) SUPPLY.md Confirm canonical price (~$0.00011?) or FDV basis
2 Exact DAO balance → exact redistribution (~990B vs owner's ~995B) §1 Snapshot 0x9011 balance
3 Private Sale 100B disposition §3 Owner decision: DAO-retain / to-NFTs / redirect
4 Validators+Pass = 11.111B internal split NFT-COLLECTION.md Split between 100 Genesis Validators and 100,000 Pass
5 Genesis-Validator bonds 100B vs sheet's 11.111B for Validators+Pass §2, NFT §6 Confirm intended relationship (on-chain bond vs cash allocation)
6 Pass (LUX) bond per unit NFT §1c Provide bond
7 ATM/Card $20k → 200M LUX peg NFT §1d Confirm peg (reconciles to 200B "Credit Card" slice)
8 Mobile Validator count + mixed one-time/per-year pricing NFT §1b Provide count; confirm pricing basis
9 Burn-rate "1.1" sheet value (50%-of-fees is confirmed) SUPPLY.md Provide full sheet value

5. Contract gaps

GenesisNFTs.sol cannot yet represent the full collection (additive fixes; the existing 118 tokens are undisturbed since minting locks nothing physical):

  • No 10B Coin tier — the authorized Wallet-10B coin has no representable tier.
  • No Card bond — NFTType.CARD exists but _getLuxForTier has no Card amount.
  • No sub-Genesis validator hierarchy / no Pass type.

Detail and rationale: cchain/GENESIS-NFTS.md.


6. The path to converge

To move from today to the canonical target:

  1. Public Sale (400B): stand up the 369-daily-auction sale contract; move 400B from the DAO wallet into it.
  2. Team + Developers (100B + 100B): deploy 366-day-cliff / 10%-per-month vesting; move 200B in.
  3. Network Treasury + Rewards + Partners (100B each): fund the three program wallets; move 300B in.
  4. Private Sale (100B): apply the owner's §3 decision.
  5. NFT collection: extend GenesisNFTs (§5); mint the remaining 12,093 authorized tokens under owner gate; escrow bonds if bonds are to be physically backed.
  6. DAO endpoint: after 1–5, the DAO wallet should hold 1.00T (the canonical Lux DAO bucket) — down from ~1.99T today.

Net treasury outflow to converge ≈ 990B (⚠️ exact per item 2 above).


Current-state figures derive from data/cchain-current.json, data/eth-authorized-supply.json, and data/migration-map.json (generated + verified). The ~1.99T treasury balance is from the mainnet-recovery record and should be re-snapshotted for an exact figure.