Same Position, Four Different Bills: How Funding Rates Differ Across Perp DEXs in 2026

Funding is not a trading fee — it is a recurring payment between longs and shorts that keeps a perpetual's price tethered to spot. The rate you face is not set by the exchange's price list; it is produced by that venue's own order book and its specific formula, cadence and caps. So the identical $10,000 long can sit in a mildly positive funding regime on one perp DEX and a near-zero or negative one on another at the very same moment. Over a month, that difference is the gap between a rounding error and a hole in your PnL. This is the cost most calculators ignore and most bot operators discover only after the fact.

TL;DR

• Funding is paid on your notional position size, not your margin — so leverage multiplies its bite against your actual capital.
• Three of the four venues here — Hyperliquid, dYdX v4 and EVEDEX — settle funding hourly; Aster defaults to an 8-hour interval, with some pairs moved to hourly.
• Hourly settlement is not inherently cheaper or pricier: the rate is computed the same way, just sliced finer. What it changes is when you can exit without paying.
• The real driver of "different bills" is each venue's premium component — how far its perp is trading from the index, which depends on that book's own crowding and liquidity — plus different interest baselines and caps.
• A position held for 30 days at a sustained +0.01% per 8h costs roughly 0.9% of notional; at +0.03% per 8h, roughly 2.7%. For a long-lived DCA or COMBO bot, that is often larger than your trading fees.

What funding actually is (and what it is not)

A perpetual future has no expiry date, so there is no settlement moment at which its price is forced to meet spot. Funding exists precisely because of this: the mechanism continuously charges the side that is pushing the contract price away from spot, which incentivizes arbitrageurs to take the other side and pull it back. When the perp trades above the index, longs pay shorts; when it trades below, shorts pay longs.

The critical point for cost accounting is who collects it. On a venue like Hyperliquid, the exchange does not collect funding at all — it debits one side of the book and credits the other. The same is true on Aster, where funding is a peer-to-peer transfer that the platform neither charges nor receives. That distinction matters: funding is not a fee you can shop away by finding a cheaper venue's commission schedule. It is a market price for being on the crowded side of a trade, and it is as volatile as sentiment itself.

Two structural facts flow from this. First, funding is computed on notional value, not on the margin you posted — so at 10x leverage, a funding cost that looks like 0.9% of position size is effectively 9% of the capital actually at risk. Second, because the rate is driven by each venue's own premium, two honest exchanges can quote materially different funding for the same market at the same time without either being "wrong."

How each venue computes the bill

The four venues share the same skeleton — a premium component that measures perp-versus-index drift, plus a small interest component — but differ in cadence, formula detail and caps. These are the structural facts that survive from one day to the next, regardless of where rates happen to sit.

Hyperliquid. The rate is computed over an 8-hour window using a CEX-style formula, but Hyperliquid then settles one-eighth of that rate every hour, producing continuous hourly cash flows instead of three daily snapshots. The formula is F = Average Premium Index + clamp(interest rate − premium, −0.05%, +0.05%), where the premium index is built from impact prices sampled across the window. The interest component is fixed at 0.01% per 8 hours — about 0.00125% per hour, annualizing to roughly 11.6% APR structurally paid to shorts. Funding is capped at 4% per hour — far less aggressive than most centralized venues. A position opened and closed inside the same hour pays no funding at all, which makes intraday scalping cheap relative to venues that snapshot every eight hours.

Aster. Aster's funding is a periodic peer-to-peer payment built on a premium index, with positive rates meaning longs pay shorts and negative rates the reverse. The default interval is 8 hours, but Aster reserves the right to adjust the funding floor, cap and interval during extreme volatility, which may differ from that 8-hour default. In practice it has already done so for individual markets — the PTBUSDT perpetual was moved from a 4-hour to a 1-hour funding interval as of late March 2026, making settlement smoother and more responsive. The practical takeaway: on Aster, you cannot assume a single fixed cadence across every pair — check the interval on the specific market you trade.

dYdX v4. Funding payments on dYdX are paid or received every hour, with the purpose of keeping each perpetual trading close to its oracle price. At the end of each hour, the one-hour premium is calculated as a TWAP of the sixty per-minute premium samples gathered over that hour, on top of which sits a fixed interest-rate component. The cap is unusual because it is tied to margin parameters and adjustable by governance: the 8-hour rate is capped at 600% × (Initial Margin Fraction − Maintenance Margin Fraction) — for a large-cap market like BTC-USD that works out to 600% × (5% − 3%) = 12%. Governance can adjust the funding clamp factor and related parameters, so the rules themselves are not static.

EVEDEX. The funding rate is computed once every 8 hours per trading pair, but settlements occur hourly, with each hour charging or crediting a pro-rated FR ÷ 8 between longs and shorts. The rate is determined using an Average Premium Index, with positive funding sending payment from longs to shorts and negative funding reversing it. Structurally this is closest to Hyperliquid's approach — an 8-hour rate sliced into hourly settlements — which means EVEDEX shares the same "close within the hour to skip the next payment" property.

What 30 days actually costs

Here is the part calculators skip. Funding looks trivial per interval and compounds into something real over the holding period of a bot. The arithmetic below is illustrative — it shows what a sustained rate would cost, not a prediction of any specific venue's rate, which moves hourly. Funding is paid as notional × rate each interval, so for a flat $10,000 position:

Two things to read from this. First, even the benign baseline — which is roughly where funding gravitates in a balanced market — is on the order of trading-fee magnitude over a month, so for any position a bot holds for weeks it cannot be ignored. An 8-hour-equivalent rate above 0.01% (~11% APR) is generally considered an attractive level to "farm" by taking the paid side, which tells you that this baseline is the neutral hinge, not an extreme. Second, the leverage interaction is where it stops being academic: that same $90 over 30 days, on a position you opened with $1,000 of margin at 10x, is 9% of your capital gone to carry alone, before a single adverse price tick.

This is why the same position produces four different bills. The numbers above assume one fixed rate, but in reality each venue's premium is set by its own book's lean. A heavily long-crowded perp DEX can be running +0.05% per 8h on BTC while a more balanced venue sits near zero — same market, same direction, same notional, radically different monthly cost. The live rate per venue must be pulled before you commit, not assumed; CoinGlass aggregates real-time and historical funding per market and per exchange and is the place to check the current number for each of these venues before opening a multi-week position. (Editorial note for internal use: do not insert specific per-venue funding figures here — they are stale within the hour. Link the CoinGlass per-venue pages instead.)

What this means for grid, DCA and COMBO bots

Automated strategies are exactly the profile that funding hurts most, because they hold exposure across many funding intervals by design. A futures DCA or COMBO bot that builds and holds a directional position for days or weeks is paying (or receiving) funding the entire time, and that flow lands on the same side of the ledger as your edge — it either pads a winning thesis or quietly drains a marginal one. The structural implications are worth being concrete about:

• Funding can flip your trade's sign. A strategy that is barely profitable on price alone can be net-negative after a month of adverse funding. Conversely, being on the paid side (shorts in a crowded-long market) is a real, if unreliable, tailwind.
• Cadence changes your exit math. On the three hourly-settling venues here, closing a position before the top of the hour forfeits at most a partial hour of funding; on a strict 8-hour venue, mistiming an exit can mean eating a full interval. For a bot rebalancing frequently, that adds up.
• The premium is a sentiment signal, not just a cost. Persistently high positive funding means the long side is crowded and paying to stay — which is information your strategy can use, not merely a line item to minimize.

The honest framing is that funding belongs inside the strategy: modeled in backtests, watched in live monitoring, and priced into position sizing — not treated as a surprise at month-end.

The risks funding introduces that traders underestimate

Funding can reverse against you. The rate is not fixed; it tracks sentiment. A position that was receiving funding can start paying it the moment the book's lean flips, and over a long hold that swing can be larger than your expected price gain. Any plan that assumes today's funding regime persists for weeks is making an unhedged bet on sentiment.

"Cash-and-carry" is not risk-free. The textbook funding-arbitrage trade — long spot, short perp to collect positive funding — is often described as market-neutral, but it carries real exposure: funding can turn negative and force the carry to bleed; the spot and perp legs can diverge under stress; and on a perp DEX you layer on smart-contract, oracle and venue-solvency risk that does not exist in a simple spot holding. The "free yield" framing hides these.

Leverage turns a small rate into a large drag on capital. As the 30-day table shows, funding measured against notional looks modest; measured against the margin you actually posted at high leverage, it can rival or exceed your liquidation buffer's worth of slow bleed. Position sizing should account for cumulative funding over the expected hold, not the per-interval figure.

Caps and parameters can change. dYdX funding parameters are subject to adjustment by governance, and Aster reserves the right to alter the funding floor, cap and interval during extreme volatility. The rules you modeled against are not guaranteed to be the rules in force during the next stress event.

Final takeaway

The slogan fits the math: the same position really does generate four different bills, because funding is priced by each venue's own order book rather than by a shared rate card. The structural differences — Hyperliquid's hourly slicing and 4% cap, Aster's adjustable 8-hour default, dYdX's governance-tuned margin-linked cap, EVEDEX's FR ÷ 8 hourly settlement — set the frame, but the bill itself is written by where each venue's perp is trading relative to spot at the moment you hold. For a long-lived automated strategy, that is not a footnote; over thirty days it can be larger than every commission you pay. Model it, monitor it, and check the live rate per venue before you carry a position for weeks.

The point of comparing funding across venues is to put the cheaper carry to work in a real strategy. Bitsgap runs GRID, DCA and COMBO bots across connected perp venues from a single dashboard, with demo trading and backtesting so you can model funding drag before committing capital, and smart orders to time exits around settlement. Compare carrying costs where it matters — in your own bots.

FAQ

What is a funding rate on a perpetual DEX? It is a recurring payment exchanged directly between long and short traders to keep the perpetual's price aligned with the spot/index price. When the perp trades above the index, longs pay shorts; when below, shorts pay longs. It is a peer-to-peer transfer, not a fee collected by the exchange.

Is funding paid on my margin or my position size? On your notional position size, not your margin. That is why leverage matters so much: a funding cost equal to 0.9% of a $10,000 position is 9% of the $1,000 of margin you posted at 10x.

Do all perp DEXs charge funding every 8 hours? No. Hyperliquid, dYdX v4 and EVEDEX settle funding hourly; EVEDEX and Hyperliquid compute an 8-hour rate and pay one-eighth each hour, while dYdX averages per-minute samples into an hourly figure. Aster defaults to an 8-hour interval but has moved some pairs to hourly, so the cadence can vary by market.

Why is the funding rate different on different exchanges for the same coin? Because each venue's rate is driven by its own order book's premium — how far that perp is trading from the index, which reflects that book's specific crowding and liquidity — plus venue-specific interest baselines and caps. The same BTC long can face very different funding on two venues at the same instant.

How much can funding cost over a month? At a sustained +0.01% per 8h (roughly the neutral baseline) a $10,000 position costs about $90 over 30 days, or 0.9% of notional. At +0.03% it is about $270, and in a hot, crowded market at +0.10% per 8h it can reach ~$900 — figures that often exceed total trading commissions for a position held that long.

Can funding turn a profitable trade into a loss? Yes. For a strategy held across many intervals, adverse funding can accumulate into a cost larger than the price gain you were targeting, flipping a marginally profitable trade negative. It can also reverse mid-hold as sentiment shifts.

Is funding-rate arbitrage (cash-and-carry) risk-free? No. It is exposed to funding turning negative, divergence between the spot and perp legs under stress, and — on a DEX — smart-contract, oracle and venue-solvency risk. The "market-neutral" label understates these.