Eight reactors at about $15 billion apiece is arithmetic, and the arithmetic belongs to the source: $120 billion, or 60 percent of the $200 billion pledge, consumed at once [4][5]. What the two sides actually settled late in the negotiations is narrower, two APR-1400 units first and expansion later if possible, with broad language to that effect in the document [16]. Two units at $15 billion is about $30 billion, roughly 15 percent of the pledge [1]. The remaining six are about $90 billion of direction, 45 percent of the total [2], resting on sites, schedules, unit costs and profit-sharing that all remain to be discussed [9].
The more interesting version of the question is where Seoul's return actually sits. An investment special purpose vehicle will oversee funding for the projects [8], but the return described most concretely in the Seoul Economic Daily account is not the vehicle's: Korean firms would inevitably supply the APR-1400's core equipment, so they earn during construction even where the Korean government provides the money [17], and they can also manufacture main components for the AP-1000 that Washington wanted the funds pointed at, which analysts cited by the outlet say would lift their share of the supply chain [18][14]. Equipment margin is paid against construction progress. Equity return is paid against power prices at a site not yet chosen, in electricity markets that differ by region and where AI data centre building clusters, so profitability varies widely by location [19]. Encinal is the live test of that: the cost was raised late in the talks at Washington's request, and designing a structure that secures adequate returns is treated as the open problem [13].
If nuclear does take $120 billion, about $80 billion is left for the Encinal combined-cycle plant, the Alaska natural gas pipeline development and everything else the package names [3][6][10]. That is the allocation worth tracking, because capital committed to US energy infrastructure under a tariff-driven pledge [12] is capital not deployed against Korean industrial policy at home, and the eight-unit headline books 60 percent of it before a single site exists [5][9].
The reading this desk will defend is that the agreement is export financing with an equity wrapper, and that the eight-unit figure is a ceiling drafted to be quotable rather than a capital plan; the two units are the deal, the six are an option Korea holds without paying for it. Two other readings are available. The signed text could name sites, schedules and a profit formula for more than two units, in which case $120 billion of genuine capex is in motion and the returns question becomes urgent rather than theoretical. Or the six never get sites, the $90 billion drifts toward LNG and pipeline work, and Korea ends up with a lower-return, faster-cycling book than the nuclear framing implies. Ruling party and government officials briefed Democratic Party lawmakers behind closed doors on the 7th [11]; the text expected as early as the 18th [1] is what settles it. Sites, schedules and a profit-sharing formula for more than two units in that document would show this read to be wrong.
