Run a standard free cash flow screen on Hyundai Motor Co. (KRX:005380) for FY2025 and you get roughly negative ₩14.4tn. Cash from operating activities was negative ₩5.99tn. Purchases of property, plant and equipment were ₩8.37tn. Subtract the second from the first and the company appears to have destroyed an amount of cash close to a tenth of its market capitalisation in a single year, while reporting ₩10.36tn of net income.
It did not. The screen is measuring the wrong thing, and understanding why is the single most useful piece of knowledge for anyone looking at this company from outside Korea.
Start with the pattern, because a one-year anomaly and a structural feature look different.
Cash from operating activities, by fiscal year: FY2020 negative ₩0.41tn, FY2021 negative ₩1.18tn, FY2022 positive ₩10.63tn, FY2023 negative ₩2.52tn, FY2024 negative ₩5.66tn, FY2025 negative ₩5.99tn. Before that, from FY2015 through FY2019, the line was positive every year, in a range of ₩0.42tn to ₩3.92tn.
Now put the financing line next to it. FY2023 brought in ₩9.39tn, FY2024 ₩19.49tn, FY2025 ₩15.42tn. In FY2022, the one year operating cash was strongly positive, financing was negative ₩1.32tn.
The two lines move in opposite directions with a consistency that isn't coincidence. When operating cash goes out, financing cash comes in. Net change in cash across FY2025 was negative ₩654bn on a balance of ₩18.36tn, which is to say the company's actual cash position barely moved.
The balance sheet explains it. At June 30, 2026, Hyundai carried ₩394.51tn of total assets, of which ₩265.47tn were non-current. Property, plant and equipment accounted for ₩49.92tn of that and intangibles for ₩10.09tn.
Which leaves something over ₩205tn of non-current assets that are neither factories nor patents. For a company that also holds only ₩22.50tn of inventory and ₩6.90tn of trade receivables, that is not a manufacturing balance sheet. It's a lending one.
Hyundai consolidates a large captive finance operation that writes loans and leases against its own vehicles. When that book grows, the new loans are an operating outflow, because originating receivables is the operating activity of a finance business. The debt raised to fund those loans is a financing inflow. Same transaction, two lines, opposite signs, and the accounting standard puts them in different sections of the statement.
On the liability side the mirror image is visible: ₩259.10tn of total liabilities against ₩135.42tn of equity, with ₩161.96tn of it non-current. An automaker does not need ₩162tn of long-term debt to build cars. A lender needs it to fund a loan book.
I should be clear that the summary financials I'm working from don't itemise the finance receivables line, so the ₩205tn figure is a residual rather than a disclosed number. The notes to the half-year report break it out properly, and anyone doing real work on this company should go there.
The cleanest proof sits inside the same group.
KIA Corporation (KRX:000270) sells roughly three million vehicles a year to Hyundai's four million, in the same markets, under the same accounting standards, sharing platforms and purchasing. Its total assets at June 30 were ₩106.83tn against Hyundai's ₩394.51tn. Its total liabilities were ₩42.14tn against equity of ₩64.69tn, a ratio of 0.65 times, where Hyundai's is 1.91 times.
Kia's operating cash flow was positive ₩9.05tn in FY2025 and positive ₩12.56tn in FY2024, the same two years Hyundai posted negative ₩5.99tn and negative ₩5.66tn.
Two companies, three-quarters the volume, and balance sheets that look nothing alike. The difference is that the group's captive finance business sits under Hyundai Motor rather than under Kia. Once you know that, the comparison stops being alarming and becomes useful: Kia's statements show you roughly what Hyundai's industrial business looks like, and Hyundai's show you the industrial business plus a bank.
US readers will recognise the structure from Ford and General Motors, both of which run captive lenders. The difference is presentation. Those companies report their credit arms as separate segments and typically publish an adjusted free cash flow measure that excludes them, precisely because the consolidated statement is misleading. Hyundai's consolidated IFRS statements make no such adjustment on the face of the accounts. The information is in the segment note. The screen doesn't read the segment note.
None of this means the finance arm is harmless, and the fact that the cash flow statement is misread doesn't make the underlying business safe.
A captive lender is exposed to credit quality. Loans written to buyers of Hyundai vehicles perform well when employment holds and badly when it doesn't, and a consumer downturn shows up in provisions before it shows up in vehicle sales. The book is large enough relative to equity that a deterioration in loss rates would matter to consolidated earnings even if the factories kept running.
It is exposed to funding cost. ₩162tn of non-current liabilities has to be refinanced continuously. A lender borrows at wholesale rates and lends at retail rates, and the spread compresses when funding costs rise faster than it can reprice new originations. Rate moves hit this part of Hyundai before they hit the car business.
And it is exposed to residual values. Leases end with a vehicle coming back, and the gain or loss depends on what a three-year-old Hyundai is worth. Used-car values have been volatile since 2021, and a captive lender that underwrote leases at optimistic residuals eats the difference.
Those are the reasons to worry about this structure. Negative operating cash flow is not one of them.
The segment disclosure in the FY2026 annual report, due next March, and specifically two numbers inside the financial services segment: the provision for credit losses, and the segment's operating profit relative to the size of the book.
A growing loan book with stable provisions is the finance arm doing its job, which is moving metal and earning a spread. A growing book with rising provisions is the finance arm subsidising vehicle sales by lending to weaker buyers, which flatters volume today and costs money later. Both look identical in the consolidated cash flow statement.
The second thing to check is whether Hyundai starts publishing an industrial-only cash flow measure. The August 26 Investor Day showed a company willing to correct its own metric definitions, renaming what it had been calling total shareholder return to total payout ratio because the first term was misleading. Separating automotive cash generation from finance-book growth would be the same kind of housekeeping, and it would remove the single biggest reason foreign screens misprice this company.
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.