Perpetual contracts have no expiry. Funding payments anchor the perp price to the underlying index by transferring value between longs and shorts every hour.
The mechanism
When the perp trades above the index, longs pay shorts. When it trades below, shorts pay longs. The payment is proportional to position notional and is settled directly in USDC.
funding_payment = position_notional × funding_rateA positive rate means longs pay; a negative rate means shorts pay. The rate is applied to the position at the funding timestamp, not to the entry price.
Rate calculation
The rate has two components:
premium = (mark_price - index_price) / index_price
funding_rate = clamp(premium / 8 + interest_component, -cap, +cap)- The interest component is 0.01% per hour on mainnet.
- The cap is ±0.01% per hour for majors and ±0.04% per hour for long-tail markets.
- The rate is computed from the mark price and the index at each hourly boundary.
Funding is not a fee
Funding is a transfer between traders, not revenue for Nexwest. The protocol does not take a cut. The payment appears in your ledger as funding and is included in portfolio PnL.
Example
A trader is long 2 BTC at a mark price of 100,000 USDC. The funding rate is 0.005% for the hour.
notional = 200,000 USDC
funding rate = 0.005%
payment = 200,000 × 0.00005 = 10 USDCThe long pays 10 USDC to shorts. If the rate were negative, the long would receive 10 USDC.
When funding is charged
| Network | Interval | Timestamp |
|---|---|---|
| Mainnet | 1 hour | On the hour, UTC |
| Testnet | 5 minutes | Every 5 minutes |
Positions opened after the timestamp do not pay for that interval. Positions closed before the timestamp do not receive the payment. The exchange snapshots positions at the boundary.
Estimating carry
Use the fundingHistory endpoint to fetch the last 500 rates for a market, or subscribe to activeAssetCtx for the current predicted rate. A simple annualized estimate is:
annualized_rate = hourly_rate × 24 × 365At 0.005% per hour, the annualized cost is about 43.8%. In practice rates mean-revert; use the historical distribution rather than a single snapshot.