Hyosung Heavy Industries Corporation (KRX:298040) is bought for transformers. Its share price rose 10.08% on August 27 to ₩3,079,000, valuing the company at ₩28.71tn on the strength of a global shortage of high-voltage grid equipment that has taken its operating margin from 1.48% in FY2020 to 15.67% in the second quarter of 2026.
On July 31 its management committee resolved something unrelated to any of that. The company would assume ₩346,300,000,000 of debt owed by a project finance vehicle called Jamwon-dong Complex Development PFV, because a completion deadline had arrived and the building was not finished.
The filing scores the amount at 13.91% of consolidated equity.
US readers will recognise completion guarantees, but the Korean version is harsher and it is worth understanding because it is the standard structure behind most Korean property development lending.
A developer sets up a project financing vehicle, borrows against the future building, and hires a contractor. The lenders do not want to underwrite a thinly capitalised developer, so they require the contractor, which is a large listed company with a real balance sheet, to give a 책임준공 undertaking: a promise that the building will be completed by a specified date.
If the contractor misses that date, it does not pay a penalty or make the lenders whole for damages. It assumes the loan. The entire outstanding principal and interest moves from the project vehicle onto the contractor's balance sheet.
That is what happened here. Item three of the filing gives the reason in plain terms: the responsible completion deadline of July 31, 2026 arrived and the obligation was not fulfilled. Item two describes the ₩346.3bn as the unpaid principal and interest that the borrower owed to the lenders as of the assumption date.
The filing gives no explanation for why the project was not completed.
An important distinction before this gets overstated. Assuming the debt is not the same as writing off ₩346.3bn.
Hyosung Heavy now owes the lenders, but it also stands behind an unfinished development in Jamwon-dong, which is in Seocho-gu, one of the more expensive districts in Seoul. The land and the partly built structure have value. In the normal course, the contractor finishes the building, sells or leases it, and recovers some or all of what it assumed.
Whether it recovers all of it depends on the sale prices achievable in a Korean property market that has been difficult for three years, on how much construction cost remains, and on how long the carry runs. None of that is in the filing.
What is certain is the liability. ₩346.3bn moved onto the balance sheet on August 1. What is uncertain is the loss, and the loss is what will eventually show up in the income statement.
The assumption date is August 1. The half-year report is dated June 30. So this obligation is not in the balance sheet that most investors will look at.
That June 30 balance sheet is worth examining anyway, because it shows a company with less room than the earnings suggest. Cash and equivalents were ₩172.6bn, on total assets of ₩8,555.0bn, which is 2.0%. Current liabilities were ₩4,919.0bn against current assets of ₩4,644.1bn, so working capital was already negative by ₩274.9bn. Total liabilities of ₩5,826.2bn sat against ₩2,728.8bn of equity, a ratio of 2.13 times.
Add ₩346.3bn of assumed debt to that and the picture tightens further. The assumed amount is roughly twice the cash balance.
The comparison with the pure-play alternative is unflattering. HD Hyundai Electric, which sells the same transformers into the same shortage, held ₩988.1bn of cash at June 30, had current assets exceeding current liabilities, and carried only ₩155.8bn of non-current liabilities in total. It also has no construction division.
Item five of the filing records the decision date as July 31 and notes that this refers to a management committee decision. The fields for outside director attendance are blank.
Korean companies routinely delegate defined categories of decision to a management committee, and there is nothing irregular about it. But a commitment equal to 13.91% of shareholders' equity, arising from a guarantee that has now been called, is the kind of item most investors would expect to see a board vote on, or at least an outside director present for.
The absence is not evidence of anything improper. It is a data point about how this company escalates decisions, and worth noting alongside the fact that the underlying guarantee was given years ago by whoever signed it.
Three points in mitigation, and they are real.
First, scale. ₩346.3bn is large against equity but small against a ₩28.71tn market capitalisation. Even a total loss would be roughly 1.2% of the market value.
Second, the asset. A prime Seoul development site is not a stranded factory. Korean contractors have absorbed completion guarantees before and recovered most of the money by finishing and selling.
Third, the direction of travel elsewhere. Operating income in the first half of 2026 was ₩416.6bn against ₩266.6bn a year earlier, up 56.1%, and cash from operations was ₩453.7bn against ₩162.0bn. The grid business is generating enough to absorb a problem of this size.
The counter is that a responsible completion guarantee that gets called is rarely the only one outstanding. Korean contractors typically carry several such undertakings at once, and the half-year report's contingent liabilities note is where the remainder sit. That note is not in the summary financials, and anyone holding this stock should read it before assuming ₩346.3bn is the whole exposure.
The contingent liabilities note in the half-year report, and specifically the schedule of outstanding responsible completion undertakings with their deadlines. One guarantee has now been called. The relevant question is how many more there are and when they fall due.
The second marker is the third-quarter report in November, which will be the first balance sheet to include the assumed debt. Watch whether it appears as borrowings with a matching asset, which would mean the company took control of the development, or whether a provision is taken against it, which would mean management does not expect to recover the full amount.
The third is simply whether another debt assumption filing appears. Korean rules require one each time, so the disclosure is reliable. A second within a year would say the construction division's problems are structural rather than about one project, and at that point the transformer story stops being the whole investment case.
kstock reads DART every morning and writes up what moved — the contract, the buyback, the number that does not add up. The daily post and a Saturday roundup, by email.