Factbox-Texas power plant, nuclear projects under discussion in South Korea’s US investment package

South Korea and the United States are finalizing a pledged $350 billion investment package linked to a trade deal that reduced U.S. tariffs on South Korean goods to 15%, including $150 billion earmarked for shipbuilding. Proposed projects include a $22.3 billion, 6.3GW Texas gas power project for AI data centers and semiconductor plants; up to eight U.S. nuclear reactors at roughly $15 billion each; and possible South Korean participation in the $44 billion Alaska LNG project. Seoul is also discussing a potential 5%-15% Westinghouse stake ahead of its planned IPO, potentially valuing a 15% holding at $2.25 billion-$3 billion, while retaining caution on high-risk Alaska LNG exposure.

Analysis

The most investable read-through is not Korean capital deployment itself but the conversion of proposed AI-load power capacity into equipment orders. A staged gas build would favor GE Vernova (GEV), Quanta Services (PWR), and potentially Caterpillar (CAT) before it benefits merchant generators: turbines, interconnection, and transmission are contracted earlier than generation cash flows. Over 1-3 months, data-center power scarcity should support a premium for Texas-exposed dispatchable generation (VST, NRG), but a 6.3GW headline should not be capitalized until EPC awards, grid interconnection status, and signed offtake contracts are disclosed.

Brookfield (BN) and Cameco (CCJ) have a credible valuation catalyst through any Westinghouse listing or strategic minority sale, but reactor construction is a poor near-term earnings catalyst: U.S. licensing, financing, and labor constraints make first concrete unlikely to translate into material cash flow for several years. The more important second-order effect is that a strategic Korean shareholder could improve Westinghouse's bid credibility and procurement access, raising IPO optionality for BN/CCJ; it does not eliminate fixed-price construction risk or the risk that reactor commitments remain diplomatic targets rather than bankable projects.

Consensus may overvalue the aggregate commitment and underweight execution conditionality. Alaska LNG and CCUS are particularly vulnerable to commercial returns, permitting, and customer take-or-pay requirements; they should be treated as negotiating currency, not incremental capex. The thesis fails if Texas power projects lack contracted hyperscaler/semiconductor offtake, if ERCOT interconnection timing slips, or if a Westinghouse transaction occurs at a valuation that leaves little upside versus BN/CCJ's already embedded stakes.

Key Decisions for Investors

• Initiate a 3-6 month long GEV / short XLU pair: turbine and grid-equipment order momentum should be monetized earlier than regulated-utility rate-base growth. Target 10-15% relative upside; exit if no named EPC or turbine award emerges within 90 days.
• Accumulate BN and CCJ selectively ahead of a Westinghouse IPO or minority-stake announcement, sized as event optionality rather than a reactor-build thesis. Use a 6-12 month horizon; reduce if implied Westinghouse valuation materially exceeds comparable nuclear-equipment and fuel-cycle multiples without contracted backlog disclosure.
• Maintain a watchlist, not a position, in VST and NRG for Texas capacity/offtake announcements. Enter only after disclosed long-term AI or semiconductor customer contracts; absent contracts, incremental gas capacity can dilute scarcity pricing rather than enhance it.
• Avoid treating WAB, APP, NVDA, and SMCI as direct beneficiaries: no disclosed revenue linkage exists. For NVDA/SMCI, the relevant confirmation would be customer data-center build schedules and utility power reservations, not diplomatic investment headlines.