neuron-tai/docs/adr/0002-dual-token-payment-model.md

TAI token: revenue-backed rewards for nodes, USDT/SOL payments for clients

Status: Accepted (settlement mechanics superseded by ADR-0015)

Settlement update (2026-07-04): Alpha and near-term production use USDT-direct custodial settlement with pending-balance forfeiture penalties (ADR-0015). TAI emission, backing-price buyback, and stake deposits described below remain the roadmap tokenomics — not the live payment path. When TAI ships, it will mint from accumulated protocol cut; this ADR's incentive structure still governs long-term design.

Core principle

Nodes must get paid reliably for inference work. This is the primary growth engine. All tokenomics decisions serve this goal first. The model is intentionally simple and sound — parameters can be tweaked later, but the structure must have no fatal flaws from day one.

Token

Name: TAI
Total supply: 1,000,000,000 (1 billion) — fixed forever, no minting after initial issuance.
Inflation: Zero. BTC-style hard cap. No additional tokens are ever created.
Emission schedule: Team distributes TAI to nodes from team holdings on a halving-style decay curve — early nodes earn more TAI per dollar of compute than later nodes. This rewards bootstrapping risk. Exact halving intervals are TBD but the principle is locked.

Clients never hold or see TAI. They pay in USDT or SOL. Nodes earn TAI. This separation keeps the client experience frictionless while giving node operators asymmetric early-adopter upside.

Revenue-backed price model

TAI has no open market during the bootstrap phase. The only way TAI enters circulation is as a node reward for completed inference work. Price is therefore backed by real compute revenue, not speculation.

TAI backing price = USDT in protocol treasury / TAI in circulation

Price starts near zero (team sets a symbolic starting price, e.g. $0.0001) and only rises as inference payments accumulate. There is no AMM, no liquidity pool, no external price discovery during bootstrap. The team is the sole issuer and sole buyer.

Payment and reward flow

Client pays $100 USDT for inference
    └─ Protocol treasury receives $100 USDT

Node completes compute work
    └─ Node earns TAI worth $95 at current backing price
       (team transfers TAI from team holdings to node wallet)
    └─ Protocol keeps $5 as operating spread

Node wants to exit:
    └─ Team buys back at 90% of current backing price
    └─ Node receives ~$85.50 USDT
    └─ Team recovers TAI + keeps ~$14.50 total on round-trip

Node holds TAI:
    └─ Team keeps $95 USDT in treasury → backing price rises → TAI appreciates
    └─ Holding is always strictly better than selling immediately

Dynamic spread

The protocol spread (~5% on issuance + ~10% effective on full round-trips) is the minimum viable amount to sustain operations and maintain the wind-down reserve. As inference volume grows, the spread decreases. The target is the lowest spread consistent with:

1. Covering infrastructure and operating costs
2. Maintaining the wind-down reserve (see below)
3. Maintaining enough reserve to resist hostile market actions post-listing

There is no transfer tax, no reflection mechanic, no fee on wallet-to-wallet TAI transfers. Democratising AI inference means keeping friction as low as possible. Protocol revenue comes only from the inference spread.

Buyback floor and wind-down guarantee

The protocol maintains a 90% buyback floor on the current organic backing price at all times.

Floor = 90% × (USDT treasury / TAI in circulation)

This floor ratchets upward — it never falls — because the backing price only rises as inference volume accumulates USDT into treasury. TAI therefore cannot go to zero as long as the protocol operates.

Wind-down guarantee: If the project is shut down, the protocol buys back all circulating TAI at 90% of the backing price at time of shutdown. Every node that earned TAI through legitimate compute work can exit at this floor. The floor is always funded because the protocol collected at least $90.25 for every $85.50 it owes on a full round-trip — the reserve builds automatically from the spread.

Post-listing risk (acknowledged): Once TAI is listed on an open market, speculative demand may push the market price above the organic backing. Nodes who earned TAI during a speculative spike and hold through a correction back to backing may redeem below their earn price. This is an accepted market risk — the protocol guarantees the backing floor, not speculative peaks. The incentive to hold is that backing grows with every inference payment; selling at the wrong moment after a speculative spike is the node's risk to manage.

Team distribution schedule

Team begins holding ~100% of the 1 billion TAI supply. Over approximately 5 years of inference rewards to nodes, the team reaches a long-term target of 36% holdings (~640 million TAI distributed to node operators).

Distribution is not linear — the halving decay means early node operators receive a disproportionate share, rewarding those who take the bootstrapping risk.

Open-market listing (provisional)

Open-market listing is announced only when both of the following thresholds are met:

• Volume: $50,000 cumulative USDT inference volume paid through the network
• Nodes: 25+ active nodes across 15+ unique wallet operators

Both gates must pass simultaneously. Neither alone is sufficient. These specific numbers are provisional and will be reviewed as the network grows — what matters is that listing happens only after real usage and real decentralisation are established, not before.

At listing, only a portion of team holdings is made available to the open market. The specific percentage will be decided at listing time based on market conditions. The team does not dump its full 36% at listing.

Node staking bootstrap

New nodes have no TAI and TAI is not on any open market. Bootstrapping is solved through the existing probationary period (ADR-0003):

• During probation (first N jobs): zero TAI stake required, zero TAI earnings, zero slashing risk. The node proves good faith through free compute work.
• After probation: node has accumulated TAI rewards from which it locks a minimum stake. Full earning and slashing enforcement begins.

No loan, no grant, no special mechanism. The probationary period is the stake substitute.

Prototype contract boundary

The packages/contracts settlement layer tracks:

• caller_balance[api_key] — USDT/SOL prepaid by clients
• node_tai_balance[wallet] — TAI earned, unredeemed
• node_tai_stake[wallet] — TAI locked as fraud collateral
• protocol_usdt_reserve — accumulated spread, earmarked as wind-down reserve

Epoch settlement debits caller USDT, credits node TAI weighted by layers served × node speed benchmark. No live Solana settlement in prototype — contract boundary is a local deterministic adapter (ADR-0007).

Considered options

• SOL only: no early-adopter upside for nodes, no viral growth mechanic — rejected
• Own token only: clients must acquire TAI — high friction, kills adoption — rejected
• AMM / open market from day one: exposes price to speculation before real volume establishes backing floor; bank-run risk before treasury is deep — rejected
• Transfer tax / reflection mechanic: breaks DEX compatibility, hostile to users, contradicts democratisation goal — rejected
• Revenue-backed TAI, no open market during bootstrap: price anchored to real compute work, wind-down guarantee is always funded by the spread, speculation cannot crash price before network effect is established — chosen

Open parameters (to be resolved before token launch)

• Exact halving schedule: at what cumulative TAI distributed does each halving trigger?
• Dynamic spread curve: at what volume levels does the spread decrease, and what is the minimum floor?
• Percentage of team holdings offered at open-market listing
• Rate limit on buybacks (max USDT/day per wallet) to prevent coordinated drain attacks
• Legal structure for token distribution in target jurisdictions (securities classification review required before any public listing)