Divide 316.8 billion won by 1.53 and the year-earlier first half lands near 207 billion [1][1], so roughly 110 billion won of fresh operating profit arrived on a revenue line that went slightly backwards [1][2], which works out to about 9.0% at the operating line this half against something under 5.9% on a larger base before [3][4]. That gain came from price, not from volume.
LS Securities said the housing division held margins above 20% for a second consecutive quarter and that one-off items did little to the plant result [3], while Samsung Securities read the same quarter as a significant housing improvement with civil engineering and plant steady from the prior quarter [2]. Those are two separate reads of the same quarter.
The balance sheet behind it carries about 1.2 trillion won of net cash against a debt-to-equity ratio of 86.4% [4], so real borrowings coexist with more cash than borrowings, and July's trust agreement to repurchase shares came to 55.5 billion won, or 4.6% of the net cash [6][7]. That is modest against the pile.
The order book is where the intention gets a number. First-half intake of 5.24 trillion won was 110.7% above a year earlier [9], which back-solves to about 2.49 trillion in the comparable half [12], so part of that growth rate is simply a small base, and the resulting 28.9 trillion won backlog is 4.1 times annualised first-half revenue [9][11]. Against that, the second-half plan of roughly 2.5 trillion won in plant work and about 2 trillion in data centres [10] is a combined 4.5 trillion [8], equal to 86% of everything booked in the first six months [9], with the plant leg implying a 25% conversion of the roughly 10 trillion won preliminary pipeline [11][10].
This is probably wrong, but the more interesting version of the story is that the data centre attempt costs DL E&C very little to make: the targets are bids, the named geographies are greater Seoul and Chungcheong [12], the X-energy agreement is a standard design contract for fourth-generation SMRs rather than a booked order [13], and the funding is a housing margin nobody has to believe in past next quarter. The thesis fails if housing reverts to single digits before plants and data centres are large enough to carry revenue, because then the cash is covering a hole rather than opening a door; it also fails, less visibly, if 2 trillion won of data centre awards land at margins that dilute a 20% housing book, and the filings say nothing about what data centre work pays. On whether any of this is cheap, the material gives earnings and cash and no share price, so that stays open.
