Former Intel CEO warns US chip comeback still has long way to go

Former Intel CEO Pat Gelsinger cautioned that restoring advanced semiconductor manufacturing to the U.S. will take significant time because wafer production migrated to Asia over decades, and the key metric is how many wafers are built domestically. His comments come as the U.S. government took a nearly 10% stake in Intel and is pushing onshore chip fabrication for national-security and AI-competitiveness reasons, but Gelsinger said major fabless designers such as Nvidia and AMD must commit to U.S. foundries before reshoring materially alters supply dynamics.

Analysis

Market structure: The near-term winners are U.S. fabs (INTC) and capital-equipment suppliers (Applied Materials AMAT, Lam Research LRCX) that capture incremental wafer-capex; losers are Asian foundries (TSMC) and vendors exposed to Asia-only fabs if political pressure forces reshoring. Pricing power shifts to equipment/materials vendors as single-source tools (EUV from ASML) and skilled labor constrain throughput; expect wafer-start growth in the U.S. to rise from ~<20% today toward 25–30% over 3–7 years, not months.

Risk assessment: Tail risks include execution failure at Intel (multi-billion-dollar write-offs), US-China escalation disrupting Taiwan shipments, or ASML export bottlenecks; these are low probability but high impact and could occur within 0–36 months. Short-term (days–months) news will drive volatility around subsidy announcements and capex contracts; long-term (3–7 years) outcomes depend on cumulative capex, talent availability, and EUV tool cadence.

Trade implications: Direct plays favor suppliers and politically-backed Intel: establish modest exposure to INTC (2–3% portfolio) and larger exposure to AMAT/LRCX via 6–12 month call spreads to capture cyclical capex; implement a relative-value pair (long AMAT, short AMD) sized small (1%/1%) to play equipment demand vs fab concentration risk. Options: sell short-dated covered calls on NVDA to monetize IV if long, or buy 9–12 month AMAT/LRCX calls (OTM 15–25%) on capex re-acceleration; rotate out of EM semiconductor exposure into U.S. industrials over 1–6 months.

Contrarian angles: Consensus underestimates time and cost — expect multi-year dilution of gross margins for new U.S. fabs as wages and supplier scarcity push capex +20–40% vs baseline estimates. Government ownership of INTC raises political support but increases governance and exit-risk; mispricings will appear if markets price instant wafer-share gains (sell rallies >25% in Intel or equipment names absent confirmed wafer-start guidance).

Key Decisions for Investors

• Establish a 2–3% long position in INTC over the next 2–6 weeks, target hold 6–18 months; set a 20% downside stop and add on confirmed quarterly guidance showing >$5bn incremental U.S. capex awards or fabs Y/Y wafer-start growth.
• Allocate 1–3% to AMAT and/or LRCX via 9–12 month call spreads (buy 12-month OTM 15–25% calls, sell nearer-term calls to fund) to capture expected equipment demand; trim if shares rally >30% absent concrete tool-delivery schedules.
• Implement a small pair trade: long AMAT 1% vs short AMD 1% (equal $ notional) for 6–12 months to play equipment demand vs foundry concentration risk; unwind if AMD reports multi-quarter guidance beats or TSMC announces U.S. fab commitments.
• If long NVDA, sell 30–60 day covered calls to harvest elevated IV or buy 3–6 month protective puts if NVDA exposure >3% of portfolio; target to reduce net delta during headline-driven rallies tied to reshoring commentary.
• Reduce EM semiconductor/fabless exposure by 1–3% and rotate into U.S. industrials/capital goods over 1–3 months; re-evaluate when U.S. wafer-starts reported quarterly increase by >5 percentage points.