# Non-Crypto Perps on Hyperliquid: Every Market, Its Hours, and Its Funding Cost

> Every equity, commodity, index and FX perp listed on Hyperliquid, when each one trades against its underlying market, and why funding pins to half the crypto baseline overnight.

*Source: https://hyperliquidguide.com/guides/trading/non-crypto-perps-on-hyperliquid*

A perpetual on Nvidia is not a stock, and it is not the Nasdaq contract on the same name either. It is a contract that never expires, quotes around the clock, and stays tethered to its underlying by a payment that changes hands every hour. For most of its life, the market it is tracking is shut.

That last point is the one that gets underestimated. A US equity trades 6.5 hours a day, five days a week: 32.5 hours out of the 168 in a week. The perp on it quotes for all 168. So roughly 80.7% of the time a US equity perp is trading, there is no live cash reference for it to price against.

This page is the reference for what that means in practice. Which non-crypto markets are listed, what hours each underlying actually keeps, how funding behaves once the reference market closes, and what the gap costs to hold. Every figure comes from Hyperliquid's own public API or a venue's published schedule, with the endpoint and the date beside it.

> **Key takeaway:** When the underlying market closes, funding on a non-crypto perp stops tracking a premium and settles onto a baseline of **0.00000625 per hour**. That is exactly half the crypto baseline. At the moment this data was read, 37 of 103 live non-crypto markets sat on that number and **none** of the 233 native crypto perps did.

---

## What a non-crypto perp is, and how it differs from the futures contract

The instrument is a cash-settled perpetual swap. You post USDC as margin, take a long or short position sized by leverage, and settle profit and loss in USDC. You never hold the share, so no dividend reaches you and there is nothing in custody with your name on it. The mechanics are the same ones the [native crypto perps](/guides/trading/hyperliquid-xyz-explained) run on, pointed at a different oracle.

Against a listed futures contract on the same underlying, the differences that actually change how you trade it are these.

**There is no expiry, so there is no roll.** A CME contract has a delivery month, converges toward settlement, and has to be rolled if you want to keep the exposure. A perp has neither. Nothing forces convergence on a date, which is why it needs funding instead.

**Funding replaces convergence.** Rather than a term structure pulling price toward a settlement, an hourly payment moves between longs and shorts to keep the perp near its oracle. When the perp trades above the reference, longs pay shorts. When it trades below, shorts pay longs. Our [funding rates explainer](/guides/trading/funding-rates-explained) covers the mechanism in full.

**The access model is different.** There is no brokerage account to open and no expiry calendar to manage. Deposit USDC, connect a wallet, trade. That is the trade being made: you give up ownership of the underlying and you take on the funding cost described below.

One nuance cuts against the usual framing here. The futures side is not standing still. CME Group has moved its 1-ounce gold and 10-barrel crude contracts to [24/7 trading](https://www.cmegroup.com/globex/files/1-ounce-gold-and-10-barrel-crude-futures-expansion-to-24-7-trading.pdf), so the round-the-clock advantage is narrowing at the small-contract end. What has not narrowed is the absence of an expiry and the size of the listed roster.

**Trade Stocks and Commodities Without a Broker** — Equity, commodity and index perps settle in USDC on Hyperliquid. Use code Concept211 for a 4% lifetime fee discount. [Start Trading](https://app.hyperliquid.xyz/join/Concept211)

## Every non-crypto market currently listed

A `meta` call against the trade.xyz DEX on **1 September 2026** returned **117 markets**, of which **103 were live** and **14 carried an `isDelisted` flag**. The retired 14 were URANIUM, ALUMINIUM, DXY, VIX, CORN, WHEAT, TTF, VOL, KRW, H100, NIFTY, IBOV, IBIDEN and KSTR, which is worth knowing if you find one referenced in older material.

Leverage caps on the live set clustered tightly. 73 markets capped at 10x, 23 at 20x, two at 25x, one at 30x, and four at 50x. Every live market returned a `deployerFeeScale` of exactly **1.0**, which puts the whole roster on the standard HIP-3 schedule of 0.09% taker and 0.03% maker rather than on a bespoke rate. The [fee guide](/guides/fees/fees-explained) covers how that stacks with the referral discount and staking tiers.

By 24-hour notional volume at the time of the read, the busiest twelve were:

| Market | Max leverage | 24h notional volume |
|---|---|---|
| SKHX | 10x | $240,935,977 |
| [SNDK](/markets/xyz/sndk) | 10x | $204,767,229 |
| [CL](/markets/xyz/cl) | 20x | $184,768,419 |
| XYZ100 | 30x | $170,111,062 |
| SP500 | 50x | $160,178,688 |
| [SPCX](/markets/xyz/spcx) | 20x | $108,809,233 |
| [MU](/markets/xyz/mu) | 10x | $104,787,325 |
| [NVDA](/markets/xyz/nvda) | 20x | $88,202,456 |
| SILVER | 25x | $83,034,122 |
| [DRAM](/markets/xyz/dram) | 20x | $81,632,752 |
| [BRENTOIL](/markets/xyz/brentoil) | 20x | $68,540,928 |
| GOLD | 25x | $67,806,863 |

The complete list below is read from the same API call when this page is built, so a market listed last week appears without anyone editing the article.

---

## Trading hours: the perp against its underlying

Every market above quotes 24 hours a day, seven days a week. None of the underlying markets do. The gap is where the interesting behavior lives.

| Underlying | Regular session | Source |
|---|---|---|
| US equities and ETFs | 09:30 to 16:00 ET, Mon to Fri | NYSE / Nasdaq |
| Korean equities | 09:00 to 15:30 KST, Mon to Fri | [Korea Exchange](https://global.krx.co.kr/main/main.jsp) |
| Japanese equities | 09:00 to 11:30 and 12:30 to 15:30 JST | [Japan Exchange Group](https://www.jpx.co.jp/english/equities/trading/domestic/01.html) |
| Gold and silver | Sun 17:00 to Fri 16:00 CT, 60-minute daily break | [CME Group](https://www.cmegroup.com/trading-hours.html) |
| WTI crude | Sun 17:00 to Fri 16:00 CT, 45-minute daily break | [CME Group](https://www.cmegroup.com/trading-hours.html) |

Read that table against a 168-hour week and the asymmetry is stark. A US equity perp has a live cash reference for 32.5 hours and no live reference for the other 135.5. The commodity perps are far better covered, since the CME session runs most of the week, which is part of why [Brent](/markets/xyz/brentoil) and [WTI](/markets/xyz/cl) behave more like their futures than the equity perps behave like their stocks.

What fills the gap is the oracle plus whatever order flow shows up. News that lands after the US close reprices the perp immediately, and the cash market spends the next session catching up or fading it. Our [after-hours trading guide](/guides/trading/after-hours-trading-guide) covers how to trade that window; this page is about what it costs to sit through it.

## How funding behaves when the reference market is closed

This is the part that surprised me, and it is the reason the page exists.

Both readings below were taken from `POST https://api.hyperliquid.xyz/info` on **1 September 2026 at 08:04 UTC**, a Tuesday, roughly five and a half hours before the US cash open. The US equity market was shut.

The non-crypto set (`{"type":"metaAndAssetCtxs","dex":"xyz"}`), across 103 live markets:

| Funding state | Markets | Share |
|---|---|---|
| Exactly 0.00000625 per hour | 37 | 36% |
| Negative | 35 | 34% |
| Positive, above the baseline | 22 | 21% |
| Positive, below the baseline | 7 | 7% |

The native crypto set (`{"type":"metaAndAssetCtxs"}`), taken at the same moment across 233 markets: **zero** sat at 0.00000625. BTC, ETH and DOGE all returned exactly 0.0000125. SOL returned -0.0000007418 and HYPE -0.0000026926.

Two separate things follow from that, and they get conflated easily.

**The venue applies a different baseline.** The value 37 non-crypto markets pinned to, 0.00000625, is precisely half the 0.0000125 that the crypto majors were paying. That is a property of the venue, not of the moment.

**Only the non-crypto markets pin to it.** No crypto perp sat on its baseline, because crypto spot never closes, so there is always a premium to track and funding always has something to respond to. Take the reference market away and there is nothing to price against, so a market with no independent order flow settles onto the floor and stays there. Roughly a third of the roster was doing exactly that.

> **Tip:** The 36% figure is a snapshot with the US market closed, not a constant. Read the same endpoint mid-session and the share collapses as live prices give funding something to track again. The point is not the number itself, it is that this is the state an equity perp spends most of its week in.

The corollary matters for anyone sizing a carry position. A market pinned to the baseline is not a market with no opinion. It is a market with no input. The moment the cash session opens, the premium reappears and funding can move hard in either direction, which is why the [90-day funding study](/ecosystem/what-hyperliquid-perps-cost) shows the equity and memory names swinging much wider than their overnight baseline suggests.

**Every Market, One USDC Balance** — Equities, commodities and crypto perps share the same margin pool on Hyperliquid. Start with a 4% lifetime fee discount. [Get the Discount](https://app.hyperliquid.xyz/join/Concept211)

## What the gap costs to hold

Work it through on the weekend, since that is the longest stretch a US equity perp goes without a reference.

The cash market closes Friday at 16:00 ET and reopens Monday at 09:30 ET. That is **65.5 hours**. At the baseline rate of 0.00000625 per hour:

```
65.5 hours x 0.00000625 = 0.00040938
 = 0.0409% of position notional
```

On a $10,000 position, **$4.09** to hold a long across the weekend. Against the HIP-3 taker fee of 0.09% confirmed above, that whole weekend costs less than half of a single taker fill, and under a quarter of the 0.18% you pay to enter and exit at taker.

Annualized, the two baselines compare like this:

| | Per hour | Per day | Per year |
|---|---|---|---|
| Non-crypto baseline | 0.00000625 | 0.015% | 5.475% |
| Crypto baseline | 0.0000125 | 0.03% | 10.95% |

That arithmetic is cleaner than reality, in ways that matter.

The calculation assumes funding sits at the baseline for the entire 65.5 hours. It often does on a quiet weekend and it does not when something happens. A market with real weekend flow prices away from the floor, and the 34% of markets running negative at the time of the read were doing precisely that.

It also assumes a long. A short across the same window at the same rate **receives** 0.0409% instead of paying it, which is the part most people skip past when they assume holding overnight is simply a cost.

And it ignores what the position does to your margin. Holding through a closed session means holding through the gap risk of the reopen, and the maintenance-margin math does not pause because the cash market did. [What happens when an equity perp position is liquidated](/guides/trading/liquidation-explained) covers that side.

---

## Methodology and sources

Market roster, leverage caps, deployer fee scale and funding rates come from `POST https://api.hyperliquid.xyz/info`, read on **1 September 2026 at 08:04 UTC**. The non-crypto set uses `{"type":"metaAndAssetCtxs","dex":"xyz"}` and the crypto control set uses `{"type":"metaAndAssetCtxs"}` with no `dex` parameter, both read within the same minute so the comparison is like for like. The endpoint is public and unauthenticated, so every count and rate above is reproducible.

The complete market list is not a static table. It is rendered at build time from the same `meta` call that decides which market pages exist on this site, so it reflects the roster as of the most recent build rather than the date in the prose.

Trading hours are the venues' own published schedules, linked in the hours table. The fee figures are the published HIP-3 base-tier schedule, applicable here because every live market returned a `deployerFeeScale` of 1.0 on the date read. Per-year and per-weekend figures are arithmetic on the hourly rates, shown in full above so the working can be checked.

What this page does not claim: that any market will still be listed, still capped at the same leverage, or still funding at the baseline when you read it. Those are point-in-time readings from one call on one date. The live figures on each market page are the current ones.

You may republish these figures with attribution and a link to https://hyperliquidguide.com/guides/trading/non-crypto-perps-on-hyperliquid.

Hourly funding baseline by venue, read 2026-09-01 08:04 UTC. Source: Hyperliquid metaAndAssetCtxs API, via Hyperliquid Guide (https://hyperliquidguide.com/guides/trading/non-crypto-perps-on-hyperliquid)SetMarkets readBaseline funding per hourAt baselineAnnualized baselineNon-crypto (trade.xyz HIP-3)103 live0.0000062537 (36%)5.475%Native crypto perps2330.00001250 (0%)10.95%`}
/>

Related reading on this site: [how the trade.xyz venue works](/guides/trading/hyperliquid-xyz-explained), the [equity perps guide](/guides/trading/equity-perps-guide) for the stock side in detail, the [commodities guide](/guides/trading/commodities-trading-guide) for metals and energy, and [HIP-3 builder codes](/ecosystem/hip-3-builder-codes) for how a venue like this gets deployed in the first place.