Investor Stephanie Link swaps Chevron for technology stocks

Stephanie Link sold her Chevron position after realizing a 35% gain and rotated proceeds into unspecified technology names, while retaining a large position in oilfield services SLB. She highlighted data-center and power-infrastructure beneficiaries including Quanta Services (shares up 37% YTD) and named Vertiv (+74% in 2026) and Eaton (+21% in 2026) as attractive 'pick-and-shovel' plays supporting AI buildout. Implication: tactical rotation from an energy provider gain into tech and power-infrastructure exposure tied to data-center and AI demand; unlikely to be market-moving but relevant for sector positioning.

Analysis

Capital rotation from commodity cyclicals into AI and data-center enablement has an outsized supply-chain footprint: incremental data-center power needs (UPS, transformers, switchgear, medium-voltage construction) translate into multi-year demand for copper, power-conversion semiconductors, and specialty contractors. That favors companies exposed to long lead-time engineered equipment and installation services where backlog visibility and order-book conversion give 12–36 month revenue predictability, while semiconductor leaders capture volatile, front-loaded GPU demand that is more output- and cadence-driven.

A meaningful second-order is project execution risk — permitting, labor and commodity inflation can push multi-MW projects out by seasons not quarters, creating asymmetric revenue phasing for equipment suppliers vs sticky margins for chip makers. Macro and policy shocks (higher rates, export controls, a China capex slowdown) compress long-duration tech multiples quickly; conversely, a surprising acceleration in energy prices would reverse opportunity-costs for capital redeployed away from hydrocarbons within 3–9 months.

Tactically, the trade-off is dispersion: pick-and-shovel names offer lower beta to GPU spot cycles but higher project/capex sensitivity; pure-play GPU/ASIC names offer convex upside on AI adoption but are exposed to inventory cycles and sequential demand misses. Position sizing should reflect that dichotomy: size equipment/contractor exposure to capture multi-year secular upgrades, and use option structures on semiconductor leaders to express asymmetric upside while limiting delta exposure to near-term macro noise.

Key Decisions for Investors

• Long PWR/ETN (equal-weighted allocation, 12–36 month horizon): accumulate on pullbacks of 8–15%. Target +35–50% upside if multi-year grid/data-center buildouts hold; set 20% stop to protect against project deferral or commodity-driven margin squeeze.
• Long NVDA via 9–15 month call spreads (buy 1–sell 1 at ~30–40% higher strike) to capture AI demand convexity while financing cost. Risk limited to premium; reward 2–4x if GPUs remain volume-constrained and ASPs hold.
• Pair trade — long select power-infrastructure contractor (PWR or similar) / short integrated energy major (CVX) sized 1:1 by dollar exposure for 6–12 months: captures secular grid upgrade vs cyclical oil-price exposure. Exit if oil price rally >20% in 60 days or contractor backlog-to-revenue reversion >30%.
• Long SLB (oilfield services exposure) as a hedge to an oil-price reacceleration, size smaller (~25% of tech exposure) with 6–12 month horizon: expected low correlation to GPU cycles, protects portfolio if energy tightens; cut if rig count falls >10% sequentially.
• Event-driven options: buy 3–6 month straddles on major semiconductor equipment names around earnings windows only if implied vol falls into 30–40% range — favorable skew when market underprices positive guidance from AI OEMs. Limit allocation to 2–3% of portfolio to avoid vega decay.