Best 3 AI ETF Picks for the Second Half of 2026

The article argues the AI trade remains favored going into 2H 2026, with semiconductors leading and the S&P 500’s best-performing sector in 2026 being tech. It highlights three AI-related ETFs—SMH, AIQ, and DTCR—while noting SMH is concentrated with Nvidia and TSM accounting for ~30% of assets and trades at ~24x next 12-month earnings. Overall, it frames valuation support and continued earnings growth as the key tailwind, despite some signs of industry “exhaustion” in segments.

Analysis

The edge here is not in owning the obvious AI beta; it’s in the breadth trade underneath it. The first leg of the cycle was a narrow multiple expansion in hyperscaler-exposed names, but the next 1-3 months should be driven by who can monetize the capex wave without paying the highest narrative premium. That favors second-order beneficiaries like MU and INTC on a relative basis, while the most crowded parts of the basket are increasingly hostage to any disappointment in order growth or lead-times.

SMH still works as a macro barometer, but its concentration makes it fragile if the market starts pricing normalization rather than acceleration. The more interesting structural pocket is DTCR, because data-center scarcity is really a power/interconnect and lease-duration story, not just a GPU story; if that bottleneck persists, REIT cash flows can de-risk the AI theme even if semiconductor multiples compress. The key falsifier is a step-down in hyperscaler capex guidance or evidence that memory pricing and foundry utilization are peaking before end-market revenue catches up.

Contrarian view: consensus is still treating "AI infrastructure" as one trade, but the winners will likely diverge sharply over the next 6-18 months. If NVDA/TSM regain earnings revision momentum, the rotation thesis is premature; if not, the market will keep migrating down the stack toward lower-concentration, more cash-flow-visible exposures. That makes this a breadth, not a chase, setup.

Key Decisions for Investors

• Prefer AIQ over SMH for new AI exposure over the next 1-3 months; it captures the theme with less single-name concentration risk and should outperform if the market keeps rotating beyond megacaps.
• Pair trade: long MU and INTC vs short SMH on any strength, sized beta-neutral, targeting 8-12% relative outperformance over 1-3 months if second-order beneficiaries keep getting bid; exit if NVDA earnings revisions re-accelerate.
• Start a tactical long in DTCR on pullbacks for a 6-12 month horizon; the risk/reward improves if power/interconnect scarcity keeps data-center lease rates elevated, but trim if vacancy rises or utility queues clear faster than expected.
• Use NVDA and TSM as hedge legs rather than core longs until the next earnings/capex readout; if either company raises guidance materially, the breadth-rotation thesis should be reduced quickly.