KIA Corporation (KRX:000270) had its best quarter ever by volume and by revenue in the three months to June. It sold 851,639 vehicles and booked ₩33.04tn. Operating income came in at ₩2,628.6bn, down 4.9% from a year earlier.
Management's framing is that this is a recovery in progress. Operating margin has improved for three consecutive quarters from the 5.1% low in Q3 2025, when US automotive tariffs first bit hard. That's accurate. The margin sequence reads 5.10%, 6.56%, 7.47%, 7.96%, and an executive quoted after the results contrasted this year's "recovery in core profit capacity" with last year's struggle.
Look at what's driving the improvement and the story gets less comfortable. The gross margin has not recovered at all.
Take the same four quarters and split operating margin into its two components.
Gross margin: 18.86% in Q3 2025, 18.31% in Q4, 19.67% in Q1 2026, 18.29% in Q2 2026. Selling, general and administrative expense as a share of revenue: 13.76%, 11.75%, 12.20%, 10.33%.
The gross margin is flat to slightly down across the period, and in the most recent quarter it fell 138 basis points sequentially. SG&A fell 187 basis points over the same three months. Net that off and operating margin rose 49 basis points, which is the improvement being described as recovering core profitability.
Against the prior year the picture is the same. Q2 2025 produced a 20.02% gross margin and a 10.60% SG&A ratio for a 9.42% operating margin. Q2 2026 produced 18.29% and 10.33% for 7.96%. Gross margin lost 173 basis points year on year. SG&A saved 27. The operating line lost 146.
So the cost of building and selling a Kia, relative to what Kia charges for it, has not improved. What has improved is the overhead line, and most of that is the unwinding of a spike. Q3 2025's 13.76% SG&A ratio is far outside this company's normal range, and the sequence since then looks like a provision quarter normalising rather than a business getting more efficient. Kia's Q3 2025 SG&A was ₩3,946.8bn on ₩28.69tn of revenue. Two of the four quarters before that ran between 10.6% and 11.8%.
Normalising an overhead spike is a real earnings benefit. It happens once.
The company gave a useful figure without quite spelling out its implication. It said the cost of sales ratio would be 79.2%, below 80%, excluding tariff effects.
Reported cost of sales in Q2 was ₩26,994.2bn against ₩33,037.1bn of revenue, or 81.71%. The gap between reported and the ex-tariff figure is 2.51 percentage points. Applied to the quarter's revenue that's roughly ₩826bn of tariff cost in three months.
Annualise it and you get something close to ₩3.3tn. Now compare that to the damage already visible in the annual accounts: FY2024 operating income of ₩12.67tn fell to ₩9.08tn in FY2025, a drop of ₩3.59tn on revenue that grew 6.2%. The two numbers line up well enough that the tariff explains most of what happened to this company's profitability last year.
I want to be careful about what that arithmetic is and isn't. It rests on the company's own ex-tariff figure, which is a hypothetical rather than an audited line item, and on the assumption that the entire 2.51-point gap is tariff rather than a mix of tariff and other adjustments. The half-year report doesn't break out a tariff cost line. Treat ₩826bn as an order of magnitude derived from management's number, not as a disclosed figure.
For a US reader the relevant structural fact is that Kia sells overwhelmingly outside Korea. Domestic sales were 18.2% of the quarter's volume against 81.8% overseas, and the United States is its most profitable market by a wide margin.
Kia builds vehicles in West Point, Georgia, and it also imports a large share of what it sells in America from Korea and Mexico. A tariff on imported vehicles and parts therefore lands on a big fraction of the US business, and it lands on the cost line rather than being recoverable through price without losing volume. That's the difference between Kia and a US-domiciled competitor: the same tariff that raises a rival's input costs at the margin raises Kia's cost of goods across a much larger share of what it sells.
The volume record tells you Kia has chosen to defend share. 851,639 units in a quarter, with electrified sales up 60% year on year, is not a company withdrawing from the market. It's a company absorbing cost to keep the dealer network full.
Three counters are worth taking seriously.
First, the gross margin comparison is contaminated by mix. Electrified vehicles rose to 35.3% of sales, and the reported gross margin includes whatever those units earn. If the mix shift is dilutive to gross margin but accretive to volume and to future service revenue, judging the business on this quarter's gross margin understates it.
Second, an 18.29% gross margin and a 7.96% operating margin are still good numbers for a volume automaker. Kia earned a 7.95% operating margin for all of FY2025, a year it describes as damaged. Most Western mass-market manufacturers would take that. The comparison to FY2024's 11.79% flatters the past rather than condemning the present, because FY2024 was an exceptional year of pricing power that has now normalised across the whole industry.
Third, tariffs can change. They are a policy variable, not a law of physics, and the ₩3.3tn a year this appears to cost would come straight back if the policy shifted. A company earning ₩9tn of operating income with a ₩3.3tn policy cost embedded is a different asset from one earning ₩9tn structurally.
Valuation reflects some of this already. At the August 27 close of ₩126,100 across 390,412,998 shares, Kia's market capitalisation was ₩49.23tn, roughly $33bn. That's against total equity of ₩64.69tn at June 30, so the shares trade at about 0.76 times book, and against first-half net income of ₩4.16tn, which annualises to somewhere near six times earnings. Dividends paid in Q2 alone were ₩2.64tn. The market is not pricing this as a company whose margins are about to recover.
Gross margin in Q3, reported in late October. That's the line to watch, not operating margin, and specifically whether it holds above 18.29% or slips further.
Two things make Q3 the right test. It laps the tariff-affected quarter from 2025, so the year-on-year comparison stops being distorted by the transition. And it's far enough from the Q3 2025 provision that the SG&A line should be at a normal level, which removes the cushion that flattered Q2. If gross margin improves in a quarter where SG&A can't fall much further, the recovery is real. If operating margin stalls or reverses because there's no overhead left to save, then what happened in the first half of 2026 was the reversal of a one-time charge, and the tariff cost is still sitting exactly where it was.
The second marker is any disclosure of localised production plans for the US. Building more of what Kia sells in America inside America is the only durable answer to the cost, and a capital commitment to that would say more about the next three years than any quarterly margin will.
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.