An equity injection does not pay a fuel bill. It moves gearing, which moves borrowing headroom, and headroom is what the government is actually buying: the stated expectation is that more capital improves KEPCO's capacity to borrow [11]. The size has not been settled [12], which tells you the decision to do it came first and the number second.
Look at what the borrowing has to carry. Interest alone is said to run at 11.5 billion won a day [5], about 4.2 trillion won a year [1], equal to roughly 85 percent of the 4.9127 trillion won of consolidated operating profit KEPCO booked in the first half [2]. That profit was down 16.6 percent year on year as fuel purchase costs rose 8.8 percent after the Middle East conflict [6]. Debt grew by more than 5 trillion won in six months [4], a pace above 10 trillion won a year if it holds [6]. A company in that position does not fund a national grid programme out of retained earnings.
The programme is 70 transmission lines covering 3,855 kilometres by 2038 at an estimated 72.8 trillion won, plus 10.2 trillion won for distribution, so more than 80 trillion won over the next decade or more [7][8]. That is about 38 percent of the debt already on the books [3], and works out to roughly 18.9 billion won per kilometre of transmission, near 13.6 million dollars [4].
For scale on how unusual the instrument is: the last time the government took KEPCO equity was May 2011, about 10.6 billion won of new shares issued in exchange for taking over power lines to front-line military units [10]. That was an in-kind swap worth less than a single day of today's interest [5]. KEPCO's own account is that apart from 3 trillion won of paid-in capital at the 1989 listing, the government has never made a cash equity investment in the company [9].
The engineering being chosen alongside the money points the same way. The Ministry of Climate, Energy and Environment plans to lean on existing corridors and expand two-circuit lines to four circuits, which it says can cut the need for new high-voltage towers by nearly 40 percent [13], with underground routing prioritised through dense residential areas [14]. Fewer towers is cheaper, but it is mostly faster, because towers are where siting fights happen. Meanwhile 26 projects tied to the Manufacturing AI Transformation push and the three mega projects were cleared of preliminary feasibility studies [15], and the trade ministry intends 6 trillion won of contribution projects from 2027 to 2030 for clusters in Yongin and the southwestern and Chungcheong regions [16]. The cost-benefit gate comes off in the same week the state takes the equity risk.
None of this expands the quantity of chips Korea can make. A senior semiconductor industry official put the problem as fabs and data centres that get built but are hard to run without power and water, and said speed matters more than volume [3]. What the injection finances is a calendar.
