The "AI Magnificent 10" Is Not the Magnificent Seven Plus Three

Defiance rang in a new ticker on September 10, 2026 for AIMG, the "AI Magnificent 10" ETF. The name invites a comfortable picture — the Magnificent Seven, two or three upgrades, a bigger tray of the same blue chips. The picture is wrong, and the gap between the label and the contents is the actual story. This fund is an equal-weight bet on the machines underneath the AI buildout, and only two of the Magnificent Seven even qualify. Before anyone treats it as "the Mag Seven, now with more room," the card needs looking at.

The matchup. Same bell, same clock, one question: where does the AI profit pool live — in the platforms that own customers and cash flow, or in the hardware that makes the buildout physically happen?

• Contestant A: AIMG, a basket of ten AI supply-chain companies, equal-weighted, tracking the BITA AI Magnificent 10 Select Index. Expense ratio 0.61%.
• Contestant B: a Magnificent Seven basket, market-cap weighted, sold for about 0.30%.
• Start: today's close. Formula: total return, twelve months, both normalized to 100 paper points.

That is the investor decision wearing a matchup costume. AIMG doesn't blend both philosophies — it picks the hardware side.

The name disconnects. Of the seven "Mag" stocks, only two of the Magnificent Seven even qualify: Nvidia and Alphabet. Microsoft, Apple, Amazon, Meta, and Tesla are not. Around them the index stacks Broadcom, TSMC, Samsung, SK hynix, Micron, Marvell, Lumentum, and Coherent — foundries, memory makers, optics houses, and custom-silicon shops. This is a semiconductor-and-networking portfolio wearing the language of the platform giants.

Equal weight is the mechanism.Each of the ten names carries roughly a tenth of the fund. That means a mid-sized optics maker like Lumentum or Coherent holds the same paper weight as Nvidia. A cap-weighted Magnificent Seven is the opposite trade: Nvidia's roughly $5.3 trillion sits beside Microsoft's $3.6 trillion with Alphabet far behind, so the megacaps dominate. AIMG deliberately flattens that ordering to give the small, high-growth suppliers of the buildout the same vote as the largest. The thesis under the equal weight is simple: historically the market treated the components as separate trades, and Defiance wants one ticker for the entire AI supply chain.

Why the building, and what it costs. The launch narrative is a spending surge: the five largest hyperscalers are expected to lift capital spending to about $1.3 trillion next year from roughly $800 billion in 2026, with Nvidia reporting record data-center revenue, Broadcom's AI semiconductor line up more than 200% year over year, and high-bandwidth-memory spot prices reported at four to five times long-term contract levels. If the pool truly sits at the silicon layer, AIMG is positioned where the money physically lands.

The risk side of that bet is priced in. Growth comes at rich multiples — Marvell trades near 76 times trailing earnings and Broadcom near 45, against Microsoft at about 27 and Google at 16. The fund is non-diversified and concentrates in the semiconductor industry, so a memory downturn would hit a freshly spiked HBM cycle and most of the book at once. And there is the fee: 0.61% is roughly double the Magnificent Seven wrapper and triple an S&P 500 ETF, a headwind a paper race must carry from the first checkpoint.

The close. AIMG is not a safer, more diversified version of the Meg Seven trade. It is a higher-variance wager on the cyclical machinery below it — equal-weighted, single-industry, costlier, and priced for continuation. Investors who want their AI exposure inside mature platforms have a cheaper, more diversified wrapper. Investors who specifically believe the profit pool lives in the custom silicon, memory, and optics of the buildout are the audience this fund was built for. The label says ten Magnificent. The holdings say: pick your layer, and know what you are paying to play it.