The Global X Artificial Intelligence & Technology ETF turned $10,000 into $13,400 in six months, powered by a semiconductor concentration that now faces its first real test.
The AIQ ETF returned 34% year to date through June 4, nearly tripling the S&P 500's 11% gain, as a 35% semiconductor weight in Nvidia, Broadcom, Samsung and SK hynix captured the AI infrastructure buildout.
"The fund's Korean memory exposure gave it differentiation that a Nasdaq-100 tracker structurally cannot replicate," said Austin Smith, a financial publisher covering AI ETFs. Samsung and SK hynix together account for 9.1% of the portfolio.
Nvidia's Q1 FY27 revenue surged 85% to $81.6 billion, with data center alone at $75.3 billion and gross margins holding at 75%. Broadcom's AI semiconductor revenue jumped 143% to $10.8 billion, though the stock fell 13% on June 4 as the market re-rated the print. Alphabet guided 2026 capital expenditure to as much as $185 billion, anchoring the demand side of the chip thesis.
The fund's $6.97 billion in net assets reflects price appreciation more than fresh inflows — assets have fallen from $7.80 billion since February even as the ETF gained. With the Nasdaq plunging 4.2% on June 5 and the 30-year Treasury yield topping 5%, the same concentration that drove AIQ's outperformance now exposes it to a regime where strong economic data delays rate relief and compresses tech multiples.
Three signals determine whether the setup holds
The first is Nvidia's data center gross margin. As long as the 75% non-GAAP figure holds and the quarterly run rate keeps stepping up — Q2 guidance calls for $91 billion in revenue — the chip leg of the thesis remains intact. The second is hyperscaler capital expenditure. Alphabet's $175 billion to $185 billion plan for 2026 is the anchor; any negative revision from Alphabet, Microsoft, Meta or Amazon would land hard on AIQ's top ten holdings. The third is Broadcom's AI semiconductor cadence, where the $16 billion Q3 guide implies an annualized run rate above $60 billion. Thursday's 13% drop after the print is the early warning that meeting expectations no longer pays the same way beating them used to.
The valuation question changes the math from here
AIQ's 65% trailing 12-month return against the S&P 500's 27% was earned fairly — the fund owns the memory duopoly that supplies high-bandwidth DRAM for every Nvidia accelerator, plus the hyperscaler names funding the buildout. But the entry point has shifted. Composite sentiment on AIQ reads 66.42, bullish with medium confidence, while Broadcom sentiment cratered from the mid-60s to 25.48 in roughly a week. Alphabet sentiment is climbing, Nvidia is neutral. That dispersion is what late-cycle AI exposure looks like when valuations are no longer cheap and earnings have to do the work alone.
For investors holding AIQ, the number to watch is Alphabet's next capital expenditure update. That single data point is the demand signal against which everything else in the top ten is sized. If hyperscaler spending holds, the fund's semiconductor weight continues to compound. If it contracts, the same concentration that made AIQ work in 2025 and the first half of 2026 will work in reverse.
This article is for informational purposes only and does not constitute investment advice.