iShares' AI ETF Returned 91% Last Year. Can It Do It Again in 2026?

BlackRock’s iShares Future AI & Technology ETF (ARTY) has delivered a 91.22% 1-year return with $2.81 billion in net assets, while rival Global X AIQ has $8.56 billion AUM and a 51.53% 1-year return. The article argues that AI-related capex, data-center buildout, and commercialization trends are broadening the opportunity set across semiconductors, power, commodities, and ancillary industries. Overall tone is constructive on AI as a durable multi-sector investment theme rather than a short-lived spike.

Analysis

The key signal is not that AI demand is strong; it is that the market is now treating AI infrastructure as a multi-year capex supercycle with unusually durable visibility. That shifts alpha away from “AI software story stocks” and toward the picks-and-shovels stack where pricing power is still underappreciated: memory, networking, foundry capacity, and power-related equipment. In this regime, the biggest winners are firms with constrained supply or strategic bottlenecks, while the losers are the generic “AI attached” names that trade on narrative but lack direct throughput leverage.

Second-order effects matter more than the headline ETF flows. If hyperscalers keep spending, the real margin expansion accrues to suppliers with tight capacity and long lead times, but that also raises the odds of a later digestion phase once orders normalize and channel inventory rebuilds. Semis with AI exposure could see multiple expansion for another few quarters, but the risk is that consensus is extrapolating current growth rates into 2026 without fully discounting cyclical payback, especially in memory and networking where capex can overshoot demand.

The contrarian read is that BlackRock- and Global X-style ETF inflows may be creating crowded ownership in the same handful of constituents, which increases fragility if any one of the megacap AI spenders signals budget discipline. A short-duration drawdown could come from earnings guide-downs on AI monetization timing, not from AI demand fading. The market is likely underpricing how quickly beneficiaries beyond the obvious mega-cap software layer can re-rate, but overpricing the durability of current multiples if growth slows even modestly.

For positioning, the cleanest expression is long the AI supply chain with balance-sheet support and earnings leverage, while fading the most crowded parts of the AI basket on any strength. The next 3-6 months should favor names with order backlog and capacity scarcity; the next 12 months could be more selective if spending remains strong but valuation dispersion widens. The main risk is a capex pause, regulatory pressure, or a pullback in risk appetite that forces de-grossing of crowded AI factor exposure.

Key Decisions for Investors

• Go long MRVL / MU / AVGO as the highest-conviction AI infrastructure beneficiaries over a 3-6 month horizon; prefer pullbacks of 5-8% for entry, with a 15-20% upside target if AI capex remains intact.
• Use any post-earnings strength in AI basket proxies to fade crowded beta: short a proxy basket of BLK-linked AI ETF exposure versus long a narrower semiconductor supply-chain basket; this captures crowding without betting against AI outright.
• Initiate a tactical long AMD vs short CSCO pair over the next 1-2 quarters; AMD has more direct AI revenue torque, while CSCO is more vulnerable to delayed enterprise networking spend and valuation compression.
• Buy 6-9 month call spreads on MSFT only on pullbacks; the risk/reward is favorable if AI monetization remains gradual, but avoid outright chasing after strong guidance because the market is already paying for long-duration optionality.
• For a lower-beta expression, overweight BLK on a medium-term basis only as a flow beneficiary, but hedge with short-duration semis if AI ETF inflows remain strong; that captures asset-gathering upside while limiting factor drawdown risk.