Tokenomics
The token earns its supply.
OROG is the settlement token of the Orogen network, and its design answers the question a crypto-native reader asks first: where does new supply come from. Issuance is minted only against verified inference work. Customers pay in stablecoins, the gateway burns OROG to settle each completed job, and new OROG enters circulation in proportion to the revenue that real demand creates. There is no fixed inflation schedule, no halving calendar, and no discretionary mint the foundation can reach for. The chain computes the cap and enforces it.
Operators are paid per completed, verified job at the oracle spot price, denominated against the dollar, with most of each payout auto-swapped to a stablecoin at settlement. The token is the rail underneath the settlement, not a position operators are forced to carry, and because issuance is minted only against verified work, honest operators are not diluted by emissions paid to anyone who did not do the job.
Utility
Four uses, all productive.
Gateway burn at top-up
Customer pays USDC/USD/ETH/BTC; the gateway buys OROG from the protocol-managed AMM at oracle spot and burns it. The customer receives non-transferable USD-pegged CUC credits (30-day expiry). This is the only legitimate path into the network. No skip-the-token backdoor.
Operator stake
Operators bond OROG to participate. Tier-dependent: 500 OROG dc-premium → 25 OROG prosumer. Slashable for sampling mismatches, fraudulent attestation, or coldkey duplication.
Validator stake
Validators bond 10× the operator-min stake for the tier they sample. Verification compensation is proportional to (replays performed) × (consensus alignment).
Governance
Parameter changes (elasticity factor, sampling rate, ceiling/floor, slashing schedule) are governed by stake-weighted vote with a 14-day timelock. No one-time discretionary mint is possible, because the chain rejects mints above the pallet-computed cap.
Per-job settlement split
75 / 15 / 5 / 5.
Every per-job emission, sized by the pallet rule below, splits four ways, and every slice maps to a working role. Governance stakers take their 5 percent only while governance holds a real security function such as slashing arbitration or parameter review, and that slice folds back into verification otherwise.
Per-job emission · pallet-bme
75 / 15 / 5 / 5 split
sum = 100.0 %
Operator
Verification
Treasury
Governance
- Operator 75%
- Provider compute, the product
- Verification 15%
- Validators · opML watchers · zkML provers
- Treasury 5%
- Audits, security response, integrations
- Governance 5%
- Slashing arbitration, parameter review
Split is an at-launch parameter, constitutionally bounded so the operator share cannot fall below 60% without supermajority governance and a 60-day timelock. Every recipient maps to a working role; nobody collects rent without doing the work.
Emission envelope
Bounded above and below, pallet-enforced.
The chain rejects any mint above
MIN(year_N_cap, rolling_180d_burn × elasticity, 5% supply/yr)
and clamps below by a
0.5% supply/yr floor.
Synthetic envelope below illustrates how the rule binds across a 36-month run from TGE.
Mint envelope · pallet-bme
Demand-elastic, with floor and ceiling
synthetic envelope · illustrative
- mint rate (actual)
- burn × elasticity
- 5% ceiling
- 0.5% floor
Per epoch (~72 minutes), mint =
MAX(floor, MIN(ceiling, year_N_cap, rolling_180d_burn × elasticity))
.
Bootstrap cap declines from 8%/yr at TGE to 4%/yr by month 24; after that
the rolling-burn envelope governs, clamped by the 5% ceiling above and the
0.5% floor below. The synthetic curve is shown for illustration; the rule
itself is pallet-enforced, and the chain rejects any mint above the computed bound.
Burn-and-mint engine
The loop, with $-flows.
Burn-and-mint engine · the loop, with $-flows
USD in → OROG burned → execute → OROG minted → USD out
oracle-anchored · pallet-enforced
01 · USD in
Customer
Pays USDC / USD / ETH / BTC into the gateway.
02
Gateway (LiteLLM)
Buys OROG from protocol-managed AMM at oracle spot.
03 · burn
pallet-bme · burn
OROG retired. CUC minted to the customer account.
-
04
Customer CUC
Non-transferable, USD-pegged, 30-day expiry.
-
05
Operator (vLLM / SGLang / TRT-LLM)
Deterministic kernel in TEE; signed receipt returned.
-
06
Validators replay 1–5 % of receipts
Bit-identical re-execution on attested hardware.
-
07 · finalise
on_job_finalised
CUC consumed; the mint pallet computes the bounded mint.
08 · mint
pallet-bme · mint = bounded()
mint = MIN(cap_year_N, rolling_burn_90d × elasticity, ceiling_5 %_supply); clamped above floor_0.5 %_supply.
09 · split · 75 / 15 / 5 / 5
- Operator 75%
- default 70 % auto-swap → USDC at protocol AMM
- Validators 15%
- share of replay work × consensus alignment
- Treasury 5%
- audits, security response, integrations
- Governance 5%
- slashing arbitration, parameter review
10 · USD out
Operator wallet · USDC
Configurable auto-swap on operator share (default 70 %) routes through the protocol-managed AMM, so operators are paid in USD-equivalent units mechanically and are not exposed to OROG price between job and payout.
Key invariant. Steps "burn" and "mint" happen at different oracle prices. If OROG appreciates between top-up and consumption, net burn occurs and supply is permanently reduced (the Akash post-BME pattern). If OROG depreciates, the protocol mints enough to honour the customer's USD-quoted payment, still bounded by the emission cap below.
Emission policy
Demand-elastic, with floor and ceiling.
| Phase | Rule | Cap / bound |
|---|---|---|
| Year 0 to Y1 TGE | Centralised stack; no token live; bootstrap revenue. | n/a |
| Year 1 (TGE → +12m) | Hard mint cap 8% of supply / year. Subsidising operator onboarding. | 8% / yr |
| Year 2 | Hard cap declines linearly to 4% of supply / year. | 4% / yr by Y2-end |
| Year 3+ | Pallet rule: mint ≤ rolling-180d burn × elasticity (governance ∈ [0.8, 1.5]). | Hard ceiling 5% / yr; floor 0.5% / yr |
Per epoch (~72 minutes), the mint is the minimum of (a) the Year-N bootstrap cap, (b) the rolling-180-day burn times the elasticity factor, and (c) the 5%-of-supply hard ceiling, with a 0.5% floor for trough resilience. The rule is pallet-enforced. The chain rejects mints above the rule's output, including any attempted foundation override.
The design has no halving schedule, no foundation buyback program, no subnet alpha tokens, and no pool AMMs. One token, one network.
Allocation at TGE
100% of genesis supply.
| Bucket | Share | |
|---|---|---|
| Provider emissions | 40% | |
| Locked backers | 12% | |
| Team | 12% | |
| Treasury | 10% | |
| Ecosystem grants | 8% | |
| Provider bootstrap | 8% | |
| Community distribution | 5% | |
| Reserve | 5% |
Vesting summary. Team and locked-backer tranches vest over 36 months with a 12-month cliff. Ecosystem grants and provider bootstrap unlock per a milestone-based schedule managed by foundation governance with on-chain attestations. Public-sale tranche unlocks at TGE. Reserve is a long-tail buffer for adverse events; unlocking requires a supermajority governance vote with a 30-day timelock.
Subsidy targets
From subsidised to self-funded.
The subsidy ratio is (USD-equivalent OROG minted) ÷ (USD revenue burned). A ratio > 1 means the network is paying operators above what customers contribute; a ratio < 1 means net burn (deflationary). The roadmap commits to these year-by-year targets:
| Phase | Target | Rationale |
|---|---|---|
| Year 1 (post-TGE) | < 2× subsidy ratio (mint $ / real revenue $) | Bootstrap onboarding budget; operators paid above pure revenue share. |
| Year 2 | < 1.2× subsidy ratio | Network approaches self-funded equilibrium. |
| Year 3+ | < 1.0× subsidy ratio | Pallet enforces mint ≤ rolling-180d burn × elasticity. Akash post-BME pattern. |
Pricing tiers (informational)
Per-tier pricing posture.
Tier pricing as multiples of base ($0.40 / 1M output tokens for Llama-3.1-70B FP8 at dc-standard). The base price is itself a smoothed function of operator cost-of-service + verification overhead and is updated by the gateway on a slow cadence.
| Tier | Hardware floor | Verification | Typical models | Pricing | Min stake | Use case |
|---|---|---|---|---|---|---|
| dc-premium | 8× B200 / 8× H200 / NVL72 | L1+L2+L3+L4 (1–5% sampled) | DeepSeek-V3 671B, Llama-4-MoE, frontier MoE | 1.0× base | 500 OROG | Enterprise frontier-tier, premium latency SLA |
| dc-standard | 8× H100 SXM | L1+L2+L3+L4 (1–5% sampled) | 30–70B dense, Mixtral, large MoE | 0.6× base | 100 OROG | Mainstream production inference |
| cloud-rented | 1–2× H100 PCIe / H200 | L1+L2+L3+L4 (5–10% sampled) | 7–30B dense | 0.4× base | 50 OROG | Spot capacity, secondary regions |
| prosumer | 1–2× RTX 5090 / PRO 6000 | L1 stake; L3 best-effort; L4 (10%+ sampled) | 7–14B quantized | 0.25× base | 25 OROG | Lower-tier, hobbyist-friendly |
| edge | Mac Studio Ultra, dual 3090 | Stake-only | ≤ 32B single-user | 0.15× base | 0 OROG (deposit only) | Private single-tenant inference |
| embed-only | CPU AVX-512, Apple M-series | Stake-only, optional L6 zkML | Embeddings, re-ranking, classification | 0.10× base | 0 OROG (deposit only) | Cheap embedding + classification flows |
Stake floors are at-launch parameters; governance can adjust ±20% per epoch with timelock. Verification layers are described in How it works.