Best 3 Tech ETF Picks for the 2nd Half of 2026

Tech rebounded in 2Q 2026 with a resurgence tied to strong corporate earnings and easing geopolitical tensions, restoring its position as one of the top S&P 500 sectors for H1. The article argues investors should continue riding the rally as AI-driven capex is translating into AI-related revenues/earnings, while valuations are described as relatively reasonable. It also flags near-term risk in semis, noting chip stocks saw a ~13% drawdown in just three weeks, and software remains out of favor (IGV down ~17% YoY vs VGT up ~39%).

Analysis

The cleanest expression of this tape is still infrastructure over applications: SMH and NVDA have the most direct earnings torque from continued hyperscaler capex, while VGT is a lower-beta wrapper that dilutes the AI upside with mature mega-cap software/hardware names. The second-order beneficiary set is broader than the article implies: advanced packaging, networking, and power-management vendors should keep taking share before software budgets catch up, so the near-term profit pool stays with chip suppliers rather than end-user productivity software.

The main risk is that semiconductor demand is now a consensus long with crowded positioning, so any hint that cloud capex is normalizing, inventory is building, or export controls tighten could trigger a fast multiple reset. That matters on a 1-3 month horizon; semis can de-rate before fundamentals roll over. By contrast, IGV is more of a 6-18 month digestion trade: if AI starts converting into measurable seat expansion, workflow automation, or lower churn, the software group can work even if revenue acceleration lags the hardware cycle.

The contrarian mistake is assuming software is dead rather than merely early-cycle impaired. A lot of the current underperformance is sentiment-driven, and if earnings hold up through the next two quarters, IGV could outperform simply through mean reversion. But if AI tools keep compressing pricing power or reducing hiring needs, the margin pressure shows up first in high-multiple SaaS and then in the index-level software ETFs, not in the chip names until later.

Key Decisions for Investors

• Long SMH vs short IGV for 1-3 months: own the part of the AI stack with the most immediate revenue translation and hedge broad tech beta; this works unless software revisions inflect upward faster than expected.
• Add NVDA on post-rally pullbacks only, preferably into the next earnings reset window; upside remains tied to capex durability, but the trade should be sized for a sharp valuation air pocket if hyperscaler guidance softens.
• Use VGT as a lower-volatility core, but not as the primary AI expression; it should lag SMH if the market keeps rewarding infrastructure spend over broad tech breadth.
• Watch for a 5-10% drawdown in SMH on any capex pause commentary from hyperscalers; that would be the thesis-falsifier and the point to trim or hedge semis aggressively.
• If IGV stops underperforming on improving software guideposts over the next two earnings seasons, rotate into a tactical long; otherwise treat the sector as a slow-burn mean-reversion story, not a near-term momentum trade.