- Stock perpetual futures are no-expiry futures on a single US stock, held near the share price by funding payments between longs and shorts; the CFTC deemed them approved as security futures on 3 November 2026, and as of 4 November only Kalshi lists them.
- Funding prices the gap between contract and stock, so when the stock market shuts the venue's rulebook decides the weekend: Kalshi takes no readings and pays nothing, Bitnomial's filed rule repeats Friday's rate at seven funding times, and Coinbase's may publish no rate at all.
- Before holding a stock perp past Friday's close, find the venue's closed-market funding clause and its margin floor (15% to 15.5%, which caps leverage near 6.5x); the SEC can still abrogate the rules and a lawsuit seeks to block the approvals.
A perpetual future needs a live price to chase, and from Friday evening until Sunday evening a US stock has none to give. Until this week that was a designer's puzzle. Then on 3 November the CFTC said that stock perpetual futures "can begin trading for the first time" on US exchanges, and Kalshi listed the first of them. Hold one into a Friday close and it feels parked, with the chart flat and Monday the next moment anything can touch you.
Something does touch it, because the contract still owes funding. Funding is what keeps a perpetual close to its stock. When the contract sits above the share, longs pay shorts; below it, shorts pay longs. Over a weekend there is no stock price to compare against, so the venue's rulebook decides what the comparison is worth. On Kalshi, the weekend costs nothing in funding. Under one rival's filed rule, it costs Friday's rate seven times.
Eunha studies the beliefs that make risk feel smaller. A closing bell brings an old one: a market at rest costs nothing. Her way through stock perpetuals passes over the headline rate. It goes to the weekend clause in four US rulebooks, and then to what one Friday-to-Monday hold would cost.
The clauses are easy to find once you know the word. Eunha types "closed" into each filing's search box and keeps only the sentences that contain it.
What are stock perpetual futures, and where can you trade them?
Strip away the novelty and a stock perpetual future is a futures contract on one company's shares with no expiry date. A quarterly future meets the stock because it must settle. A perpetual never settles, so funding does the pulling instead. On Kalshi one contract covers a hundred shares, tradable in fractions. The contracts are security futures, so the CFTC and the SEC share them, and only an exchange registered with both may list them. Each also clears through a registered clearinghouse, which is more than an offshore perpetual has when its exchange shuts.
Permission arrived without a vote. Kalshi filed its rulebook in September and the CFTC's forty-five-day review ran out, so the contracts were "deemed approved", in Chairman Michael Selig's words. As of 4 November, Kalshi is the only venue listing them. Its product update, as reported, named contracts on Amgen, Costco, McDonald's and Qualcomm. Customers apply separately, and Kalshi warns they may lose everything they commit and may have to post more margin. Coinbase Derivatives, Bitnomial and Crypto.com's Nadex have filed rulebooks too, and all three are still waiting on the CFTC. Bitnomial's would admit only giants: at least $100 billion of market value and $450 million traded on an average day over six months.
Those filings sit in the Federal Register, so the weekend can be priced before three of the four products trade. None of them trades on a Saturday. Kalshi runs from six on Sunday evening to five on Friday afternoon, New York time, with an hour's pause each day. Bitnomial and Coinbase propose nearly the same week, and Nadex's notice describes continuous trading five days a week.
Crypto's round-the-clock contract has come onshore and picked up the stock market's calendar.
Funding needs two prices, and Friday evening takes one away
What, Eunha asks, is the rate a measurement of? The answer is a gap. Every perpetual futures funding rate compares the contract's price with the price of whatever it tracks, then charges the side pushing the two apart. While the contract runs above the stock, holding a long costs money, and buyers ease off until the prices meet. Funding isn't a forecast of anything, it's a tether, and a tether needs both ends tied to something.
By Friday night one end has come loose. The regular session ends at four in New York, after-hours quotes stop at eight, and no US venue quotes the stock again until Sunday evening. The contract may find buyers and sellers in that stretch, or none, but there's no stock price to hold them against. Kalshi already copes with a milder form of this on its index perpetual, where the contract's hours outrun the index it follows.
On a single stock, the stretch is a whole weekend.
"So the rate needs a source other than the market. Either it stops, or someone writes down what it should be," Eunha says.
A closed market can't price its own idle hours. Their cost becomes policy, set by whoever wrote the rulebook and accepted by everyone who opened a position under it.
How each venue handles funding while the stock is shut
Typed into all four filings, Eunha's word turns up four answers, no two alike.
Kalshi's is the plainest. It takes a reading each minute of the regular session, 390 on a normal day, and each one compares the contract with the stock's recent trades. The average becomes the day's rate, paid once at the four o'clock close "on each day on which the cash market is open for regular trading". No reading happens while the market is shut. So Saturday and Sunday carry no payment, and the hours the contract trades on Sunday night feed no reading. A small band nudges the rate toward zero, and a cap holds it to 2% a day.
Kalshi makes no separate dividend payment either. An ordinary dividend reaches holders only through those readings, which is one answer to how a dividend reaches a perpetual.
Bitnomial's filed rule does the opposite. It sets three funding times every calendar day, "including weekends and holidays when trading is closed". In New York they fall at four in the morning, noon and eight at night, though the filing gives them in Central time. Bitnomial's notice builds Friday's eight o'clock rate from readings taken between noon and six. That rate and Friday's closing price "carry forward through subsequent closed-market Funding Times", up to the one on Sunday evening when trading resumes. Count from Friday night to Sunday night and there are seven.
Coinbase Derivatives keeps its options open. Its funding schedule sits in an appendix the notice leaves out. Its rules say that when a contract is closed for a whole funding interval, the exchange "may decline to publish a Funding Rate for that interval". Nadex lists three funding times a day, set in UTC, for a contract that trades five days a week.
Its notice never uses the word weekend.
| Venue | Status on 4 November | Trading week (New York time) | Funding while the stock is shut | Margin floor | Leverage cap (computed) |
|---|---|---|---|---|---|
| Kalshi | Live | Sunday 6 p.m. to Friday 5 p.m. | None: no readings outside the regular session, payments only on trading days | 15.50% | about 6.45x |
| Bitnomial | Filed, awaiting the CFTC | Sunday 8 p.m. to Friday 6 p.m. | Friday's rate and closing price repeat at each funding time through Sunday evening: seven payments | 15.25% | about 6.56x |
| Coinbase Derivatives | Filed, awaiting the CFTC | Sunday 8 p.m. to Friday 5 p.m. | May publish no rate for a fully closed interval | 15% (proposed, awaiting the SEC) | about 6.67x |
| Nadex (Crypto.com) | Filed, awaiting the CFTC | 24 hours, five days a week | Not stated; three funding times a day in UTC | 15.05% | about 6.64x |
So a Saturday is free on one venue, billed on another and unwritten on the other two. The difference is not in the formulas but in the sentence each venue wrote for when there is nothing to feed them. None of it shows on an order screen, where the ticket offers a rate and never the clause behind it.
What one weekend costs when Friday's rate repeats
Bitnomial isn't trading yet, so Eunha invents one contract to test its text. Take a stock at $150. One contract covers a hundred shares, or $15,000 of exposure, and the 15.25% floor puts $2,287.50 of margin behind it. Say the contract trades a hair above the stock on Friday afternoon and the rate comes out at 0.01%. Seven payments then cost $10.50. That is 0.07% of the position, about 0.46% of the margin.
It looks like nothing, and on calm Fridays it nearly is. At 0.05% the weekend costs $52.50, about 2.3% of the margin. A positive rate bills the longs and pays the shorts, and a negative one reverses that. Either way, Saturday's payer is chosen on Friday afternoon.
How far can that Friday reading stray? The filing gives it no ceiling, since its small adjustment "does not cap the total Funding Rate". It has a shallow base instead. Each reading looks no more than $50,000 deep into each side of the order book, a depth a few resting orders can fill. The window also runs until six, two hours past the stock's close, when the index behind the contract is built from after-hours quotes.
On this rule, one Friday afternoon prices the whole weekend.
The clock then closes the exits. Trading stops at six on Friday evening and resumes at the moment the Sunday payment falls due. A position still open at Friday's close owes all seven. Holidays stretch the bill, since each weekday the stock market takes off adds three more funding times at the carried rate. On the filed wording, a long weekend around Christmas Day, a Friday this year, would mean ten payments.
Kalshi's weekend, by contrast, costs nothing in funding. What you give up is not money but the tether. From Friday's close until half past nine on Monday, Sunday night's trading included, nothing in its funding pulls the contract toward the stock. Any weekend gap is settled by Monday's price.
"Kalshi charges nothing for the weekend and holds nothing to the stock during it. Bitnomial's filing bills the weekend at Friday's rate, then goes back to measuring on Sunday night," Eunha says.
Why stock perpetual futures stop near 6.5x leverage
The margin behind the contract Eunha invented looks heavy to anyone arriving from crypto, and it is heavy on purpose. US rules set the minimum margin on security futures at 15% of the position's value, down from 20% six years ago. Each venue sits on that floor or within half a point of it, with Kalshi highest at 15.50%. Divide one by the margin and each floor stops leverage at roughly six and a half times. Crypto venues offer many times that on bitcoin.
A floor that high makes sense once you remember that stocks gap. A share that opens 10% lower on Monday takes about two-thirds of a minimum Bitnomial margin with it, before a cent of funding is counted. Beside that, seven payments at a quiet Friday's rate barely register. Eunha expects attention to go to the funding line anyway, as the only weekend cost you can price on Friday. The Monday gap decides whether the margin survives.
Can the rules change while you hold?
Every launch brings the second belief on Eunha's list: that a product you can buy will still be there to sell. The paperwork puts dates on that. Kalshi's rulebook took effect on 2 November, and within sixty days the SEC may "summarily abrogate" it and make Kalshi file again. That window closes around 1 January. The pending three carry the same clause, counted from their CFTC approval. Each notice gives it a single sentence, and none says what it would mean for an open position.
The challenge has already reached a court. The day the CFTC's statement came out, the Capital Markets Alliance sued in New York to overturn the approval. Its members include Cboe, Citadel Securities and Healthy Markets. It argues the contracts are security-based swaps under the SEC's jurisdiction and that the CFTC broke the Administrative Procedure Act. It also asks that the approvals not take effect while the case runs. SIFMA had written to both agencies the day before. It asked them to block the products and send them through the SEC's standard review, with public comment. CME's June suit over the CFTC's approval of crypto perpetuals is still pending, and the CFTC has moved to dismiss it for lack of standing.
All of it comes back to whether the contract is a future or a swap, and the answer decides which agency has the final say over the rulebook. What can change here is not a weekend's price but the product itself. An approval that arrived because a review period ran out can leave through a sixty-day order or a judge's ruling.
If the three pending rulebooks are approved, choosing a venue will also mean choosing a weekend. Eunha's advice is to make that choice on a weekday, with the filing open.
By Friday's close, the choice has already been made for you.
Carry a 2x perpetual short on a tokenized stock from Friday's close to Monday, and check its Funding Accrued line on Saturday. It keeps moving, because the simulator works funding out every two minutes from its own balance of longs and shorts: a fifth weekend rule that copies no venue. Price those two days under Kalshi's clause, which charges nothing, and under Bitnomial's filing, which bills Friday's rate seven times.
