Alaska LNG was supposed to be a footnote in a much bigger document. Instead it is one of the reasons that document is late. South Korea and the United States had aimed to sign a memorandum of understanding on September 18 covering Korea’s investment commitments in the US. That signing slipped. Korea’s industry ministry pulled a closed-door briefing to lawmakers off the calendar and rescheduled it for September 22, and the foreign minister said only that procedural matters were still being sorted out.
Here is the shape of the deal. Last year Korea agreed to invest $350 billion in the United States in exchange for having a threatened 25 percent tariff cut to 15. Of that, $150 billion is earmarked for shipbuilding. The other $200 billion goes through an investment vehicle that picks specific projects. So the number was agreed a year ago. What is being argued about now is which projects it buys, and on what terms.
The Alaska LNG project is the clearest example of why that distinction matters. Washington wants Korea in it. Seoul has agreed to keep it on the table without committing to it — which sounds like a small difference in wording and is actually the entire negotiation. The project would move gas about 1,300 kilometers from Alaska’s North Slope to an ice-free port in the south, then liquefy it for export. It is enormous, it is slow, and Korean officials have openly doubted whether it pays for itself. An exploratory line in a document costs nothing. A commitment moves the risk of a project that has not reached a final investment decision onto whoever signed.
The rest of the package is mostly things I have written about here one at a time, and it is worth seeing them together.
The first project is a 6.3-gigawatt gas plant at Encinal, Texas, worth roughly $22 billion, built to feed AI data centers — I went through how its economics depend on the Texas grid when it first surfaced. The nuclear leg is eight reactors, six using Westinghouse’s AP1000 and two using Korea’s APR1400, which runs straight into the rules on who is allowed to build what and where, the same rules I looked at through the Saudi nuclear agreement. The whole thing exists because of a tariff, which is the other half of a story I covered when chip tariffs pushed Korean fabs toward American ground.
And then there is the piece that best explains the delay. Korea has pushed for a stake of 15 percent or more in Westinghouse plus a seat on the board. The US side has floated 5 to 10 percent. Both numbers buy the same kind of thing on paper. Only one of them buys a vote. That gap is not about price, and it is the oldest question in taking a government equity stake: are you supplying capital, or are you joining the decision?
There is a cash constraint underneath all of it. The MOU as described caps remittances at $20 billion a year. That cap is not a formality. Moving tens of billions of dollars out of Korea in a compressed window would show up in the won, which is why the timing argument is also a currency argument.
Whatever gets announced next week, the mechanism is already visible. A headline number was agreed under tariff pressure, and everything that actually determines the return — which projects, what ownership, whose board, how fast the money leaves — was left to be settled afterward, under the same pressure. Alaska LNG is simply where that shows most clearly, because it is the one item Korea has been unwilling to write down as a promise.
My take: The number was the easy part. Everything that decides whether it was a good number came after the signature.
Not advice. Just how I see it.