SK biopharmaceuticals Co., Ltd. (KRX: 326030) earned an operating margin of 39.2% in the second quarter of 2026. For a company most American investors have never heard of, sitting in a country whose pharmaceutical industry is mostly built on generics and out-licensing, that number needs an explanation. The explanation is structural, and it's the reason this company is worth understanding: SK biopharmaceuticals discovered cenobamate itself, and it sells the drug in the United States under its own label, XCOPRI, through its own sales force. There is no partner taking a cut on either end.
That decision looked expensive for about four years. It's now producing the highest-margin income statement in Korean pharma. My argument is that the margin is real and durable, that the reported price-to-earnings ratio of 17.6x is not, and that the company just committed a large slice of the resulting cash to buying its next drug.
Start at the top of the Q2 income statement, because the first two lines tell you almost everything. Revenue was ₩247.4 billion, up 40.3% from ₩176.3 billion a year earlier. Cost of revenue was ₩15.0 billion. That's 6.1% of sales, leaving a gross margin of 93.9%.
Small molecules are cheap to make, so a high gross margin isn't itself remarkable. What's remarkable is that nothing sits between the gross line and SG&A. A Korean company that licenses its compound to a Western partner books a royalty, typically a mid-teens to low-twenties percentage of the partner's sales, and that's the end of its participation. SK biopharmaceuticals books the whole prescription. The difference between collecting a royalty on XCOPRI and collecting XCOPRI's revenue is roughly a factor of five on the same underlying number of prescriptions.
The cost of that choice is SG&A, and it's enormous in absolute terms: ₩135.3 billion in Q2, or 54.7% of revenue. A US neurology sales force is not cheap. But it's a fixed cost that was built once, and that's the part the market took years to price.
Look at what happened between FY2022 and FY2025. Revenue went from ₩246.2 billion to ₩706.7 billion, up 187%. SG&A went from ₩339.5 billion to ₩459.9 billion, up 35%. Operating income went from a loss of ₩131.1 billion to a profit of ₩203.9 billion.
That's the whole thesis in three numbers. The company hired a sales organization big enough for a drug much larger than XCOPRI was at the time, ate four years of losses carrying it, and is now growing into it. FY2020 is the low point worth remembering: ₩26.0 billion of revenue against ₩263.5 billion of SG&A and a ₩239.5 billion operating loss. The rep count was already mostly there. The prescriptions weren't.
The leverage is still improving. Q2 2026's 39.2% operating margin compares to 35.1% in Q2 2025 and 28.9% for FY2025 as a whole. Q1 2026 came in at 39.4%. So the first half of this year is running roughly ten points above last year's full-year margin, on SG&A that grew 29% against revenue that grew 40%.
One caveat on the pattern, since it isn't smooth. Q4 2025's operating margin dropped to 23.8% on an SG&A jump to ₩133.1 billion. Fourth quarters at this company have been heavy before, and anyone modelling the second half straight off first-half margins should expect some of that back.
Here's where the screen misleads. At ₩88,000 as of August 28, the market values SK biopharmaceuticals at ₩6.89 trillion on 78.31 million shares. Trailing four-quarter net income is about ₩391.9 billion, which gives the 17.6x that shows up on data services. For a company compounding revenue at 40% with a 39% margin, 17.6x looks cheap.
Now add the four quarters of pretax income instead: ₩93.4 billion, ₩100.0 billion, ₩33.8 billion and ₩62.1 billion, or ₩289.2 billion. Net income is above pretax income. That happens when the tax line is a credit rather than a charge, and here it was a net benefit of roughly ₩102.7 billion over the trailing year, almost all of it recognized in Q4 2025.
This isn't an accounting irregularity. It's what happens when a company that lost money for a decade finally becomes reliably profitable and gets to recognize deferred tax assets against its accumulated losses. FY2024 did the same thing on a bigger scale, with a ₩154.4 billion tax benefit turning ₩72.6 billion of pretax income into ₩227.0 billion of net income. Retained earnings have climbed from negative ₩861.6 billion at the end of 2023 to negative ₩164.1 billion at Q2 2026, and a chunk of that climb is the tax line rather than the business.
Tax a trailing pretax figure of ₩289.2 billion at even 20% and you get about ₩231 billion of normalized net income, which puts the shares near 30x. The multiple isn't crazy for this growth rate. It just isn't 17.6x, and the tax tailwind will run out. Price to book, at 6.74x against ₩1.02 trillion of equity, is arguably the more honest read on how the market sees it.
Four days ago the company disclosed an exclusive worldwide license from Biohaven for opakalim (BHV-7000), a selective Kv7.2/7.3 potassium channel activator in Phase 2/3 for focal epilepsy, along with Biohaven's broader Kv7 platform. The headline is up to $795 million. The part that matters for the balance sheet is $400 million of near-term cash, with $350 million due at closing and $50 million in 2027.
Set that against ₩380.0 billion of cash and equivalents at June 30. At any recent exchange rate, the closing payment consumes most of it.
Strategically I think this is the right kind of deal for this specific company, and it follows from everything above. SK biopharmaceuticals has built the single most expensive asset in US CNS commercialization, a working neurology sales force, and it currently carries one product. Every additional epilepsy drug pushed through that same organization arrives at close to zero incremental selling cost. Opakalim in the same indication, sold by the same reps, to the same neurologists, is the highest-leverage thing they could buy.
The risk is the timing and the price. Opakalim isn't approved. Topline data from the registration-enabling RISE3 study is expected in the second half of 2026, which means the company is paying $350 million at closing for an asset whose pivotal readout hasn't happened. If RISE3 disappoints, the cash is gone and the amortizing intangible follows it.
There's also a cash-conversion issue worth watching independently. Operating cash flow was ₩78.1 billion in Q2 and only ₩17.9 billion in Q1, so ₩96.1 billion for the half against ₩186.8 billion of operating profit. Trade receivables were ₩311.6 billion, up 72% year over year against 40% revenue growth. US pharmaceutical gross-to-net and wholesaler terms can explain a lot of that. It still means reported profit is arriving as cash more slowly than it's arriving as accounting.
The strongest counter to my valuation point is that normalizing the tax rate is the wrong exercise while real cash taxes stay low. The deferred tax assets don't just flatter net income once, they shelter actual cash payments for years, and a company that keeps most of its pretax income is worth more than one that doesn't, whatever the P/E prints.
The operating case is also getting stronger, not weaker. Margins expanded in both quarters of 2026, revenue growth hasn't decelerated, and the company is doing this with ₩13.2 billion of property, plant and equipment. There's essentially no capital intensity here. Incremental revenue converts to profit at close to the gross margin.
And on Biohaven, the bear framing assumes they overpaid for optionality. If opakalim works, SK biopharmaceuticals owns a second CNS product with worldwide rights and a sales force already calling on the prescribers. That's a much better use of ₩380 billion than a buyback.
The cleanest single marker is the accumulated deficit. It stood at negative ₩164.1 billion at the end of Q2 2026. At the ₩85 to ₩100 billion of quarterly net income the company has been producing, retained earnings should turn positive somewhere around the fourth quarter of 2026. When that happens the deferred tax story is finished, the tax line starts behaving like an expense, and reported earnings will step down even if the business does nothing wrong. That's the quarter where you find out what this company actually earns.
Two others. RISE3 topline is guided for the second half of 2026 and will decide whether the $350 million closing payment bought a product or a write-down. And in the Q3 report, check whether SG&A stays near ₩135 billion or resets higher as opakalim integration costs land, because the entire operating leverage argument rests on that line staying roughly flat while revenue climbs.
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.