VGT vs PSI: What's the Better Buy?

The piece compares Vanguard Information Technology ETF (VGT) and Invesco Semiconductors ETF (PSI), noting VGT holds 322 tech stocks across multiple subsectors (semiconductors, software, hardware, infrastructure) while PSI is a concentrated 30-stock semiconductor fund. Over the past 10 years VGT averaged a 22.18% annual return versus PSI's 24.98%, with PSI described as having higher earning potential but materially greater volatility and concentration risk. The article advises choice depends on investor risk tolerance and portfolio diversification needs and notes the Motley Fool analyst team did not include PSI in its top-10 stock picks; the author discloses a position in VGT.

Analysis

Market structure: The clear winners are concentrated semiconductor designers and foundries (NVDA, AMD, TSM, ASML, AVGO) and ETFs that concentrate that exposure (PSI) as AI-driven chip demand keeps gross margins elevated; diversified tech (VGT) benefits but sees dilution from slower software/hardware names. Pricing power will skew to leading node/AI-accelerator suppliers; less-advanced IDMs and legacy consumer chip vendors face margin pressure and potential market-share loss within 6–18 months.

Risk assessment: Tail risks include export-control escalation (U.S./China), a rapid inventory correction from hyperscalers, or ASML supply shocks — each could trigger 30–50% drawdowns in semiconductor-centric names within 3 months. Near-term (days–weeks) expect volatility around earnings; medium-term (3–12 months) fundamentals hinge on order-backlog vs. capex pace; long-term (1–3 years) secular AI demand supports higher structural TAM but concentration risk remains high.

Trade implications: Direct plays favor concentrated exposure (NVDA long, PSI tactically) with portfolio sizing limits: treat PSI as a 1–2% tactical overweight, VGT as a 3–5% core holding to dampen volatility. Options: prefer 3–6 month call debit spreads on NVDA to express upside with defined risk, and sell short-dated (30–60 day) covered calls on VGT to harvest yield if no high-conviction catalyst.

Contrarian angles: Consensus underestimates supply-side choke points (ASML, EUV lead times) and overestimates persistence of recent outperformance — historical parallel: 2017 GPU cycle then 2019 inventory bust. A realistic plan: prepare to add on semiconductor drawdowns of 20–30% (buy window) and beware capex-led overcapacity that can flip winners to losers within 12–24 months.

Key Decisions for Investors

• Establish a 2% portfolio long in NVDA (ticker NVDA) sized to risk 1–1.5% of portfolio; target +30% upside over 6–12 months, set a hard stop-loss at -18% from entry or hedge with a 3–6 month 10% OTM put if volatility cheapens.
• Take a tactical 1–2% long position in PSI (Invesco Semiconductors ETF) to capture concentrated semiconductor beta; trim or exit if PSI outperforms VGT by >300 bps over any 3-month period or if PSI implied volatility >50%
• Use VGT as a 3–5% core holding for lower-volatility AI exposure; write 30–60 day covered calls 5–8% OTM to generate 4–8% annualized yield on position while retaining upside, roll monthly unless market dislocates.
• Implement a dollar-neutral pair: long PSI vs short VGT at 1:1 notional for 3–6 months to express pure semiconductor vs broad tech spread; close if spread widens >5% or compresses to <1%, rebalance weekly.
• Monitor three binary catalysts in the next 30–90 days before increasing size: (1) U.S./China export-control announcements (material change = new restrictions), (2) ASML shipment/lead-time updates (any delay >30 days), and (3) hyperscaler guidance on inventory/capex (downgrade >10% signals inventory risk).