How to Trade MAGS (Roundhill Magnificent Seven ETF) on Hyperliquid

What MAGS Actually Holds

MAGS is the Roundhill Magnificent Seven ETF, and the Hyperliquid perp references one share of it. The fund launched on April 11, 2023 as the first ETF built to track the Magnificent Seven as a group, and it now lists on Cboe BZX after moving off Nasdaq in February 2025.

The pitch is simple: seven mega-caps, roughly equal weight, one ticker. As of the fund's June 30, 2026 factsheet, weights were Apple 14.32%, Microsoft 14.31%, NVIDIA 14.30%, Meta 14.29%, Alphabet 14.24%, Amazon 14.23% and Tesla 14.13% — about as flat as a seven-name basket gets. Roundhill rebalances back to equal weight quarterly, which mechanically trims whichever name just ran and tops up whichever one lagged. Sector exposure sits at 42.93% information technology, 28.53% communication services and 28.36% consumer discretionary. The fund held roughly $3.70 billion in assets against 57.5 million shares outstanding, at a 0.30% gross expense ratio.

One detail is worth knowing before you trade it: the fund does not simply own seven stocks. That same factsheet lists 23 holdings, because MAGS combines physical shares with total return swaps — written with Goldman Sachs and other counterparties and collateralized with Treasury bills — to stay inside the RIC diversification limits that US funds must live under. Seven flat 14% positions would otherwise break those tests. The practical effect: you are trading a swap-wrapped basket, not a pile of share certificates.

Performance has been strong over the full life of the fund. Through June 30, 2026 the factsheet shows 17.70% over one year, 29.76% annualized over three years, and 35.61% annualized since inception. That is the record the 2026 tape has been arguing with.

Why the Mag 7 Trade Got Hard in 2026

For three years the Magnificent Seven were the market. In 2026 they stopped being it. Through midyear the group was down 3.7% while the other 493 S&P names were up 12.9% — a spread of nearly 17 points and a near-perfect inversion of the prior regime. Defiance's ex-Mag 7 fund XMAG has outrun both MAGS and the S&P 500 this year, which is a strange sentence to write and exactly the point.

June was the break. MAGS fell about 9% on the month, among its worst since launch, while the CNBC Magnificent 7 index dropped 10% and the group shed roughly $2.3 trillion in market value. The fund closed at $63.14 on June 23 in correction territory, and more than $1 billion left it during the drawdown. What made the move unusual was its breadth: all seven fell together rather than one or two dragging, with Microsoft the worst of the group.

The cause was not broken businesses. It was one question — when does the AI build-out pay? The group is on track to spend more than $700 billion on AI capex in 2026, up from roughly $400 billion in 2025, and Apollo chief economist Torsten Slok has flagged that free cash flow at Meta, Alphabet, Microsoft and Amazon fell sharply from its 2024 peak. Companies bought for cash returns turned capital-intensive, and the market repriced them accordingly.

The other side of that trade is the de-rating itself. In early August the forward P/E on MAGS fell below 24, its lowest since the fund was created, against consensus for 23% EPS growth in 2026 — roughly 12 points ahead of the S&P 500 ex-Mag 7. Cheap versus its own history while still lagging the index is the kind of disagreement that produces genuine two-sided flow, which is what makes a perp market worth having.

The HIP-3 Perpetual

MAGS trades on Hyperliquid as xyz:MAGS, a perpetual future deployed under HIP-3, the framework that lets a builder stake 500,000 HYPE and operate its own perp DEX on HyperCore with its own markets, margining and oracle configuration. The xyz namespace is trade.xyz, the first team to deploy under HIP-3 and the dominant venue for equity perps on the network.

The contract references one share of MAGS, quoted in USD, with up to 10x leverage. You are not buying the ETF. You hold no shares, receive no distributions, and never touch the creation-redemption mechanism. Price tracks an oracle feed, and a funding rate paid between longs and shorts holds the contract near the underlying instead of settling at an expiry — so a crowded directional book costs you carry whether or not the trade works.

The structural feature that matters most is hours. The ETF trades on Cboe BZX during US market hours; the perp runs continuously. Mag 7 risk becomes expressible on a Sunday night, through Asian hours, and across the gap between the US close and the next open — windows where earnings, capex guidance and AI headlines have historically landed with no listed way to react. That is the real argument for this market over simply buying the fund. It also means the perp must price an underlying that is not trading, which is where both the opportunity and the hazard sit.

Key Trading Considerations

Liquidity sets your size. The perp turned over about $580,000 over the last 24 hours, with price near $67.80 inside a 52-week range of $55.09 to $71.16. That is a thin book by any measure. On a market this size, slippage and the cost of exiting a leveraged position into a fast tape matter more than being right on direction. Size to the book, not to the thesis.

Gap risk is structural, not occasional. Because the perp runs 24/7 and the ETF does not, a position carried through a closed cash session can gap when US equities reopen and a weekend's worth of repricing arrives at once. Leverage that feels reasonable during regular hours is a different instrument on a Sunday.

You are two wrappers deep. The perp references an ETF that itself sources much of its exposure through total return swaps against Treasury collateral. In normal conditions this is invisible. Under stress, counterparty and collateral mechanics inside the fund are a layer of risk that a single-name perp does not carry. Hyperliquid lists Mag 7 names individually if you want the exposure without the wrapper.

Equal weight is a position, not a footnote. Quarterly rebalancing means MAGS systematically sells strength and buys weakness inside the basket. When one name carries the group, that drags badly versus a cap-weighted vehicle. When the group moves together — as it did in June 2026 — it barely registers. Decide which regime you are trading before you assume equal weight is neutral.

It is a single-theme trade. Seven names, one narrative, high internal correlation. June 2026 showed that plainly: all seven fell together and the basket had nothing to offset with. MAGS is an efficient way to express a view on the AI capital cycle in one position. It is a poor way to diversify, and treating it as broad tech exposure is the most common way traders misuse it.