Fees & rates

Fee tiers and rate curves are estimated day-zero values, subject to change before launch. Published source, including example rate curves: Trading Fees & Rates.

Trading fees

Spot trading is free. Margin trades pay volume-tiered fees in USD, measured on 14-day rolling USD volume:

Tier	Volume	Taker	Maker
0	—	0.025%	0.015%
1	> $50m	0.02%	0.005%
2	> $100m	0.015%	0.00% + rebate
3	> $250m	0.01%	0.00% + rebate

Tier 2 and 3 makers receive a rebate of 25% and 50% respectively of the taker fee paid on their filled volume. Trading fee rates are 1,000,000-scaled integers (150 means 0.015%). Taker and maker fees round up; maker rebates round down from the taker fee.

taker_fee = ceil(notional * taker_rate / 1,000,000)
maker_fee = ceil(notional * maker_rate / 1,000,000)
maker_rebate = floor(taker_fee * rebate_share / 1,000,000)
maker_net_fee = maker_fee - maker_rebate

Example: a 10,000 USD margin fill at Tier 3 has taker_rate = 100 and maker_rate = 0. The taker pays ceil(10,000 USD * 100 / 1,000,000) = 1.00 USD. With a 50% rebate share, the maker receives floor(1.00 USD * 500,000 / 1,000,000) = 0.50 USD, so maker_net_fee = -0.50 USD. Every order carries a max fee rate field (ppm); the order will not match at a higher fee. See Orders.

Funding rates

The overweight leverage side of a market pays the underweight side, bounding skew. Rates are computed and applied once per hour from time-weighted averages over the preceding hour. Positive rate: longs pay shorts. Negative: shorts pay longs.

skew           = time-weighted (net long − net short) OI ÷ max OI
excess         = max(0, |skew| − skew_cliff)          # deadzone absorbs small skew
skew_component = sign(skew) × skew_factor × excess² ÷ 1,000,000
premium_component = avg_premium + clamp(base_funding_rate - avg_premium, ±premium_clamp)
funding_rate   = clamp(skew_component + premium_component, ±max_funding_rate)

Skew component: skew (time-weighted net long/short open interest ÷ max OI) inside the per-asset deadzone (skew_cliff) contributes nothing; beyond it, the rate grows with the square of the excess, scaled by a per-asset factor.
Premium component: tracks the deviation between the book’s impact prices (the volume-weighted price to fill a reference notional on each side) and the oracle price, pulled toward a per-asset base rate, with a per-asset clamp on the adjustment. This is mainly a stress/deposit-scarcity term; skew is the normal funding driver.

Funding rates use the same 1,000,000 scaling. Example:

average_skew = 600
skew_cliff = 500
skew_factor = 9,000
excess = 600 - 500 = 100
skew_component = floor(9,000 * 100^2 / 1,000,000) = 90

average_premium_index = 1,200
base_funding_rate = 500
premium_clamp = 500
premium_component = 1,200 + clamp(500 - 1,200, ±500) = 700

funding_rate = 90 + 700 = 790 = 0.079%

A 50,000 USD long pays 50,000 USD * 790 / 1,000,000 = 39.50 USD for that funding interval.

Funding rate vs skew: zero inside the deadzone, growing with the square of the skew beyond the cliff, clamped at the maximum funding rate

Funding rate vs skew (illustrative). Zero inside the ±skew_cliff deadzone; beyond it, it grows with the square of the excess and is clamped at ±max_funding_rate.

Imbalanced books

A market is almost always skewed, so funding is settled through the protocol rather than transferred one-to-one between traders:

The paying side is charged in full: position notional × |rate|.
The receiving side is paid its pro-rata share of what was collected, capped per account at |rate| × position notional.
Because the paying side is usually the larger side, more is collected than is owed to receivers. The surplus goes to the protocol fee account.

This is how a one-sided book still funds: the overweight side always pays the rate, the underweight side is paid up to the same rate on its smaller notional, and the protocol absorbs the difference. A single rate is charged to both sides; per-side rates to rebalance the surplus are a planned refinement.

Funding fee tiers

Accounts carrying large open interest pay discounted funding. The tier is set by the account’s open interest (OI) averaged over a 14-day rolling window, in notional USD, and reduces funding charges only — not borrow rates or trading fees.

14-day average OI	Funding discount
—	0%
> $50m	5%
> $100m	10%

The highest tier whose threshold the account meets applies. Discounts are 1,000,000-scaled registry values (50000 means 5%). The discount does not come out of the receiving side. The headline rate grosses up the raw skew-and-premium rate by the paying side’s average discount, so the same total is collected and receivers are paid in full — full-rate accounts carry the cost, not receivers or the protocol.

avg_discount   = Σ(discount × OI) ÷ Σ OI         # paying side, OI-weighted
headline_rate  = raw_rate ÷ (1 − avg_discount)
account_charge = headline_rate × (1 − discount) × OI

Σ account_charge = raw_rate × Σ OI               # the gross-up offsets the discounts

The gross-up uses the OI-weighted average discount, so total funding collected stays at raw_rate × Σ OI no matter how the discounts are distributed — what the receiving side is owed never changes. Example: the raw skew-and-premium rate is 0.095%, and the paying side holds 100m USD of open interest — 50m at full rate and 50m in the 10% tier, a 5% OI-weighted average discount. The headline rate is 0.095% ÷ (1 − 0.05) = 0.10%.

A full-rate 50,000 USD long pays 50,000 × 0.10% = 50.00 USD.
A 50,000 USD long in the 10% tier pays 50,000 × 0.10% × 0.90 = 45.00 USD.

Funding collected across the paying side is 50m × 0.10% + 50m × 0.09% = 95,000 USD — exactly 0.095% × 100m, what receivers are owed at the raw rate. The gross-up and the discounts cancel.

Borrow rates

Borrowers pay depositors a utilization rate, quoted as an annual percentage. Two rates per asset:

Base borrow rate: borrowing the asset itself (a short borrows ETH). Utilization is measured against that asset’s deposits.
Quote borrow rate: borrowing USD against that market (a leveraged long). Utilization is measured against the protocol-wide USD deposit pool.

Utilization = borrowed ÷ (lendable + borrowed), with lendable capped at the asset’s spot OI cap. Rates follow a piecewise-linear kink curve over per-asset utilization breakpoints (registry parameters): flat at low utilization, steep near 100%. An account borrows on at most one side per asset at a time (base borrows take priority over quote).

lendable = min(time_weighted_deposits, spot_OI_cap)    # if the cap is enabled
utilization = floor(borrowed * 1,000,000 / (lendable + borrowed))
borrow_rate = rate_below
            + floor((utilization - kink_below)
                    * (rate_above - rate_below)
                    / (kink_above - kink_below))

Example: if time-weighted borrows are 5m USD, deposits are 15m USD, and the spot OI cap is 5m USD, then lendable = 5m USD and utilization = 5m USD / (5m USD + 5m USD) = 50%. With kink points (40%, 2%) and (70%, 5%), the interpolated borrow rate is:

borrow_rate = 2% + ((50% - 40%) * (5% - 2%) / (70% - 40%)) = 3%

Borrow rate vs utilization: a piecewise-linear kink curve, flat at low utilization and steep near 100%

Borrow rate vs utilization (illustrative). Flat at low utilization, steepening through each kink toward 100%; the kink points are per-asset registry parameters. Charges accrue hourly in USD — each hour applies roughly the annual rate divided by the hours in a year (~8,760). Funding is instead quoted per hour, so on a leveraged position the hourly borrow charge is a small fraction of the funding payment. Collected fees are distributed to lenders pro-rata to their time-weighted average spot deposits of the lent asset, credited as USD, capped per user at deposit notional × max rate; excess goes to the protocol fee account.

lender_credit = collected_fees * lender_TWA_deposit / total_TWA_deposits
lender_credit_capped = min(lender_credit, deposit_notional * max_rate)

Example: if 3,000 USD of borrow fees are collected and a lender supplied 200,000 USD of the 1,000,000 USD time-weighted deposits, that lender receives 3,000 USD * 200,000 USD / 1,000,000 USD = 600 USD, subject to the per-user cap.

Rate data

GET /funding-rate/{asset_id}: current indicative rate.
GET /funding-rate/history: hourly applied rates (the authoritative record), newest first in the standard pagination envelope: items, page, limit, total_pages, cursor_size, has_next_page, and next_page.
Indicative funding and borrow rates also stream over the public interest-engine /rates WebSocket; see WebSocket streams.
Per-asset rate parameters (max_funding_rate, skew_cliff, base rates, kink breakpoints): GET /registry/risk-parameters.

Charge cap

Each hourly funding or interest charge is capped so the account’s equity cannot fall below the midpoint between its liquidation threshold and its bad-debt threshold. A charge can make an account liquidatable but never insolvent; an account already at bad-debt level is charged nothing. Conversion rounding favors the user.

charge_floor = (liquidation_threshold + bad_debt_threshold) / 2
max_charge = max(0, equity - charge_floor)
applied_charge = min(raw_charge, max_charge)

Example: if account equity is 1,000 USD, the liquidation threshold is 700 USD, and the bad-debt threshold is 400 USD, then charge_floor = (700 USD + 400 USD) / 2 = 550 USD. The account can be charged at most 1,000 USD - 550 USD = 450 USD that hour.

Other fees

Rebalancing: deposits, withdrawals, and settlements incur a fee or earn a reward of up to 2.5% to keep per-chain vault liquidity balanced.

Overview

Trading

Funds

Security

Trading fees

Funding rates

Imbalanced books

Funding fee tiers

Borrow rates

Rate data

Charge cap

Other fees

​Trading fees

​Funding rates

​Imbalanced books

​Funding fee tiers

​Borrow rates

​Rate data

​Charge cap

​Other fees

Trading fees

Funding rates

Imbalanced books

Funding fee tiers

Borrow rates

Rate data

Charge cap

Other fees