At the end of September 2025, SK Innovation Co., Ltd. (KRX:096770) carried ₩56,348,325,040,000 of property, plant and equipment. At the end of December, it carried ₩47,195,026,079,000.
Nine trillion, one hundred and fifty-three billion won of plant left the balance sheet in three months.
For scale, that is more than the entire market capitalisation of most companies in the KOSPI 50. It is roughly the capital cost of building the Shaheen petrochemical complex at Onsan, twice over. And it happened in a company that, on its operating line, made money that year.
FY2025 in three numbers: operating income of ₩448.7bn, pretax loss of ₩5,868.8bn, net loss of ₩5,436.4bn.
The company sold ₩80,296.1bn of product and made a positive operating margin on it. It then lost more than ₩5tn.
Almost all of that happened in the fourth quarter. On the derived quarterly figures, Q4 2025 produced a pretax loss of ₩4,344.6bn and a net loss of ₩4,184.4bn, against positive operating income of ₩337.4bn.
Retained earnings fell from ₩8,231.3bn at the end of FY2024 to ₩4,304.3bn a year later. Total equity fell from ₩39,648.9bn to ₩36,391.5bn, cushioned by other comprehensive income.
The ₩9.15tn drop in plant is not all a writedown, and the distinction matters more than most investors will bother with.
Reporting around the results attributed roughly ₩4.2tn to asset impairments taken in the fourth quarter, largely at SK On and largely connected to the restructuring of BlueOval SK, the battery joint venture with Ford.
That leaves about ₩5tn unaccounted for by impairment. The most plausible explanation is deconsolidation: as the BlueOval SK arrangement was unwound, the joint venture's plant stopped being consolidated into SK Innovation's balance sheet. Assets leaving because a structure changed is a completely different event from assets being written down because they are worth less.
The first destroys value and hits the income statement. The second is a presentation change and does not.
I cannot split the two from the summary statements. The FY2025 annual report note on property, plant and equipment carries the reconciliation — additions, disposals, impairment, changes in scope of consolidation — and it is worth twenty minutes of anyone's time who owns this stock. Reading a ₩9.15tn decline as a ₩9.15tn writedown would be badly wrong, and reading it as harmless would be worse.
The context for both is the scale of what was built.
Purchases of property, plant and equipment: ₩6,776.6bn in FY2022, ₩11,238.1bn in FY2023, ₩10,027.9bn in FY2024, ₩5,367.2bn in FY2025. That is ₩33,409.8bn across four years, and it does not include FY2021's ₩3,175.2bn.
SK Innovation borrowed to do it. Financing activities were inflows of ₩10,507.2bn, ₩9,490.4bn, ₩7,327.1bn and ₩2,345.7bn across the same four years — roughly ₩29.7tn raised. Total liabilities went from ₩29,924.2bn at the end of FY2021 to ₩70,881.2bn at the end of FY2024.
So the sequence is the same one that played out across the global battery industry, just at unusual scale: build capacity for demand forecasts made in 2021, fund it with debt, discover in 2024 and 2025 that Western electric vehicle demand grew at perhaps half the assumed rate, and mark the assets accordingly.
BlueOval SK was the vehicle for much of the American build. Reporting put the cost of the breakup at around $2.6bn. Unwinding it is expected to save roughly ₩500bn a year in depreciation and interest from here — which is another way of saying the plant it removed was generating that much fixed cost against production that was not there.
Two readings are possible and both are partly right. The generous one is that SK On protected itself by exiting a structure whose economics had stopped working, took the hit in a single quarter, and reduced its recurring cost base permanently. The unkind one is that a company committed $2.6bn to escape an arrangement it had entered three years earlier, and that the ₩4.2tn of impairment is the measure of how wrong the original decision was.
The second-quarter 2026 result is evidence for the generous reading. SK On produced ₩821.8bn of operating profit against years of losses, and the group produced record operating income of ₩3,487.3bn. A business that could not have earned that with BlueOval SK's fixed costs attached is a business that was right to detach them.
There is a good argument that this was the healthiest thing SK Innovation has done in years.
Impairments are a judgment. A management team that wanted to protect its reported earnings could have taken ₩1tn a year for four years, kept the loss out of the headlines, and left investors carrying assets on the books at values nobody believed. Taking ₩4.2tn in one quarter, alongside a structural exit, clears the deck completely and makes every subsequent quarter honest.
The evidence that the deck is genuinely cleared: capital spending has fallen to ₩1,814.4bn across the first half of 2026, financing activities have swung to an outflow of ₩3,842.0bn as debt is repaid, and property, plant and equipment has been roughly stable since — ₩47,195.0bn at year-end, ₩48,954.1bn in March, ₩47,978.0bn in June. Assets that are still being written down do not sit flat for two quarters.
And there is no sign of a second wave. Q1 2026 and Q2 2026 both produced positive net income, ₩896.1bn and ₩73.5bn respectively, without further large impairment charges.
The property, plant and equipment note in the FY2025 annual report, split between impairment and change in consolidation scope. Everything in this piece is provisional until someone reads it.
Second, whether the remaining ₩47.2tn of plant holds its value through 2026. The best test is the FY2026 impairment review, which will be the first conducted after a year in which SK On was profitable. An impairment test passed on the strength of actual earnings rather than forecast earnings is worth considerably more than one passed on a business plan.
Third, depreciation expense. If the ₩9.2tn reduction in the asset base is real, depreciation should fall visibly in 2026, and that reduction alone would add meaningfully to operating income for years. If it does not fall, then most of what left the balance sheet was never generating depreciation in SK Innovation's accounts in the first place — and the fourth quarter of 2025 was more presentation than pain.
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.