The appeal of after-hours trading is easy to understand. In the past six months, we’ve seen some of the most influential and consequential announcements and decisions happen after market close, ranging from offhand comments from the US President to Nvidia earnings.
You might have an opinion about what semiconductor stocks will do when trading resumes, and you’d rather back it immediately than wait until morning.
There’s a complication, though. When you trade a stock or an index of stocks whose primary market is closed, you’re trading an estimate of what those stocks are worth. That estimate isn’t necessarily the price you’d get if you tried to buy or sell the underlying shares.
Usually, the difference is manageable, and most traders don’t notice it. But sometimes, especially with leverage, it can become the entire trade.
The bets aren’t closed
MarketVector has licensed its US semiconductor index, which is tracked by VanEck’s SMH exchange-traded fund, to Paragon for a perpetual futures contract on Hyperliquid. The product uses an extended-hours index calculated with Pyth price data, letting traders speculate on semiconductor stocks outside the regular US trading session.
Paragon says it has launched 29 markets and handled nearly $500 million in trading volume since April 2026, though those figures don’t show how much activity the semiconductor contract itself has attracted.
The contract works a lot like the perpetual futures crypto traders know and love.
You can bet on an asset’s price without buying it, and unlike ordinary futures, the contract doesn’t expire. You can hold the position as long as you have enough collateral, paying or receiving periodic funding payments that help keep its price connected to the reference market.
Bitcoin is perfect for this arrangement because it trades everywhere, all the time. Someone buying Bitcoin perps at 2 a.m. can compare the contract against actual Bitcoin prices on exchanges around the world.
Traders who notice a large discrepancy can buy one and sell the other, profit from the difference, and help bring prices back together.
Semiconductor stocks, on the other hand, aren’t quite so accommodating.
Nvidia, Broadcom, AMD, and the other companies represented in semiconductor benchmarks trade on exchanges with established operating hours. While some shares are available through premarket, after-hours, or overnight services, that doesn’t mean every constituent trades continuously with the depth of the regular session.
The companies don’t stop being valuable when the exchange closes, of course. Their earnings prospects can improve or deteriorate overnight, and investors will adjust their expectations accordingly.
What disappears is the most liquid place to see what everyone else is willing to pay.
An extended-hours index tries to bridge that gap using prices available outside the regular session. The exact inputs, stale-price rules, and fallback procedures depend on its methodology. The specific MarketVector-Pyth calculation hasn’t been independently established, so we can’t assume precisely how it handles missing or thinly traded constituents.
But even with the best data, the index has to contend with a market that can look very different at midnight than at noon.
Everyone’s trading the value of stocks
Imagine Nvidia closes at $200, and an hour later, the company announces earnings so good that traders immediately start pricing the shares at $215.
Semiconductor stocks can move sharply on earnings, especially when AI spending expectations make a quarterly report feel like a referendum on the entire technology industry.
Now imagine the announcement comes after conventional after-hours trading ends. Some overnight exchanges may still be operating, but the liquidity available to hedge a position in the actual shares is much thinner than during the regular session.
Someone who believes Nvidia is worth $220 might buy semiconductor-index exposure through a perpetual, while someone who thinks the enthusiasm has gone too far might sell it. Their trades establish a market price, even though neither participant needs to own a single share.
There’s nothing inherently wrong with that, as futures markets have helped investors price expectations for decades. The problem is knowing how far the derivative can wander from the assets it’s supposed to represent.
During regular trading hours, a professional trader who notices that a semiconductor-index derivative has become too expensive can sell it and buy the underlying stocks or a related ETF. If the prices converge, the trader profits from the difference.
But overnight, they may not be able to buy all the stocks they want in sufficient quantities. An ETF might provide a partial hedge if it’s trading, while Nasdaq futures could offset broader market exposure, but neither necessarily replicates the index.
The trader has to decide whether the discrepancy is worth holding an imperfect hedge until the underlying market becomes more liquid.
Suppose the perpetual trades 4% above its reference index. Under ordinary conditions, that premium might attract sellers. Overnight, it could persist because the people best equipped to exploit it can’t confidently lock in the other side. It might even get larger.
Crypto traders are accustomed to checking Bitcoin prices across several exchanges. Equity-index perpetuals introduce a situation where the derivative may be one of the few actively traded expressions of a particular market view at that hour.
Its price can reflect expectations about tomorrow’s stock market and tonight’s shortage of people willing to take the other side.
You predicted the rally and lost your money
Consider a trader who believes semiconductor stocks will rally when Nasdaq opens.
The trader deposits $2,000 and opens a $10,000 long position in a semiconductor-index perpetual, using 5x leverage. Every 1% move in the contract produces roughly a 5% gain or loss relative to the initial collateral, before fees and funding.
The trader expects the underlying stocks to open 3% higher because of a major overnight announcement.
The prediction might be perfectly reasonable, but while the stock market is closed, the perpetual sells off 15% as leveraged traders unwind positions, market makers pull back, and buyers become unwilling to pay the previous price.
The position has now lost $1,500 on paper, leaving only $500 of the original collateral before costs. Whether it gets liquidated depends on maintenance margin, the mark price used for risk calculations, and the exchange’s liquidation rules.
A 15% move doesn’t automatically liquidate every 5x position, but this particular trader has very little room left. If the position is liquidated, the trader can’t wait for Nasdaq to open and prove the prediction correct.
Suppose the underlying semiconductor index subsequently opens 3% higher, exactly as expected. Someone holding the actual shares through the event might have made money. The perpetual trader could have lost most of the collateral because the derivative’s overnight price went down.
This is a good example of what happens when leveraged positions are continuously marked against a market whose underlying assets aren’t equally accessible.
Hyperliquid explains that mark prices are used for margin calculations and liquidations, while oracle prices provide external reference information. Its HIP-3 framework lets independent deployers define contracts, supply oracle prices, and set leverage limits.
The price at which traders transact, the external index value, and the mark price used to determine whether a position has enough collateral aren’t necessarily identical. An extreme transaction price won’t automatically trigger liquidation, but a sufficiently large move in the mark price can.
Funding payments also encourage perpetual prices to stay connected to their references. When a perpetual trades above its reference, longs generally pay shorts. When it trades below, payments can run the other way.
Yet someone collecting funding on a short position still has to survive any further price increase, and the promise of a payment offers little consolation if the trade gets liquidated first.
Everybody gets a reality check when trading starts
At 9:30 a.m. New York time, the regular equity session begins, bringing a much larger group of buyers and sellers into the market.
Overnight expectations can now be tested against transactions in the underlying stocks. Sometimes the overnight market gets the direction right. Other times, prices move sharply when regular trading begins because the overnight market overestimated the news or lacked enough liquidity to absorb large orders.
The opening price forms in a deeper market, with more participants able to transact in the underlying securities.
By then, the perpetual market may have already had its own trading session. Positions could have been liquidated, funding payments exchanged, and traders forced to reduce exposure. Even if the derivative and underlying index converge once regular trading resumes, those overnight losses won’t be reversed.
The index itself makes this even more complicated. An overnight headline might be excellent for Nvidia but terrible for another chipmaker. Some constituents may trade actively outside regular hours while others barely trade at all, leaving the index provider to combine different kinds of price information into a single number.
Traders can end up comparing the published index, the ETF, the perpetual, and whatever overnight prices are available for the constituent shares. None is automatically wrong because it disagrees with another, but they don’t necessarily represent equally executable prices.
Hyperliquid’s HIP-3 framework doesn’t establish that every index perpetual has identical safeguards or that its reference feed will always be available.
Without Paragon’s contract-specific methodology, liquidation parameters, and historical trading data, we can’t assess how often its semiconductor perpetual diverges from the reference index or how its safeguards performed during volatile periods.
Those details become more important as crypto venues offer products built around assets that don’t trade on the same hours as the derivatives referencing them.
Wall Street has spent years extending equity trading beyond the traditional session. Crypto is taking that idea further by letting people trade synthetic exposure without waiting for the shares themselves to become available.
There’s nothing irrational about wanting to react to news immediately. Waiting until morning doesn’t make uncertainty disappear, and overnight derivatives give investors somewhere to transfer risk while the primary market is closed.
But being able to trade something at every hour doesn’t mean you can value or hedge it equally well at every hour.
Someone buying a semiconductor-index perpetual at midnight is betting on the companies in the index, but also on the quality of the reference prices, the liquidity of the derivative, and whether other traders will hold risk until the underlying market opens.
That’s considerably more to get right than deciding whether Nvidia had a good earnings report.
The stock exchange can be closed, and the underlying shares can be unavailable at the prices everyone is discussing, but the perpetual will still accept your order.
And if you’ve borrowed enough money to make that order interesting, it may also close your position long before Wall Street gets around to opening.
