Hanmi Pharm. Co., Ltd. (KRX: 128940) posted an operating margin of 28.1% in the second quarter of 2026, on revenue of ₩467.2 billion and operating profit of ₩131.1 billion. Operating profit rose 116.9% year over year. That's a striking number for a company whose three most recent full years ran operating margins of 16.7%, 14.5% and 14.8%, and whose own first quarter this year came in at 13.7%.
It didn't come from selling more medicine. About ₩112.9 billion of it was the recognition of an upfront payment from Eli Lilly for sonefeglutide, a GLP-2 candidate for short bowel syndrome. My view is that Hanmi is a genuinely good research shop attached to a mediocre and currently shrinking commercial business, that quarterly earnings here are a licensing schedule rather than an operating trend, and that anyone valuing it on a trailing P/E is measuring the timing of contract signatures.
The clean way to spot a licensing lump is to check whether cost of goods moved with revenue. It didn't. Revenue grew ₩106.0 billion year over year, from ₩361.3 billion to ₩467.2 billion. Cost of revenue grew ₩7.5 billion, from ₩156.5 billion to ₩164.0 billion. Gross margin jumped from 56.7% to 64.9%.
Manufactured pharmaceuticals don't behave that way. Hanmi's product gross margin has sat in a tight 55% to 60% band for years, so a nine-point jump in one quarter means a chunk of revenue arrived with no factory attached to it. Contract income has exactly that shape.
There's a second confirmation further down. Hanmi paid ₩36.1 billion of income tax on ₩120.1 billion of pretax income in Q2, an effective rate of 30.0%. The four quarters before that ran 7.8%, 7.6%, 13.1% and 13.2%. Korean pharmaceutical companies shelter a lot of ordinary income behind R&D tax credits. A one-time licensing receipt is much harder to shelter, and a tax rate that quadruples in the same quarter as a margin that doubles is consistent with income that isn't ordinary.
This is the part that doesn't show up in the headlines. Take the ₩112.9 billion out of ₩467.2 billion of revenue and you get roughly ₩354.3 billion. The comparable figure a year earlier, with no upfront in it, was ₩361.3 billion. So on this arithmetic the non-licensing business was down about 2% year over year, in a quarter reported as up 29.3%.
The profit version is harsher, and I'll flag that it needs an assumption. Upfront payments rarely carry zero cost, so the ₩112.9 billion probably didn't all drop to operating profit. But cost of revenue only moved ₩7.5 billion, which caps how much of it could have been consumed at the gross line. Even assuming a generous fifth of the payment was absorbed somewhere, the underlying operating profit lands well below the ₩60.4 billion Hanmi earned in Q2 2025 without any deal income at all.
Meanwhile the research spending that generates these deals keeps rising. Gross profit less SG&A comes to ₩183.2 billion, against operating profit of ₩131.1 billion, and the ₩52.1 billion in between is where the separately reported R&D line sits. That figure was ₩44.0 billion in Q2 2025 and ran ₩195.8 billion across FY2025, about 12.7% of revenue. R&D grew 18% while the commercial base didn't grow at all.
To be fair about the domestic business: Korean prescription drug sales are a slow, price-regulated market, and Hanmi has never claimed otherwise. Low single-digit growth is the normal state. The problem isn't that it's slow. It's that the market is currently paying a growth multiple for it.
One thing that's improved without anyone announcing it. Hanmi spent heavily on plant between 2015 and 2019, with capex peaking at ₩247.1 billion in FY2017, and carried ₩516.7 billion of non-current liabilities as recently as FY2020. That's now ₩109.6 billion. Capex ran ₩43.0 billion in FY2025. Total equity has climbed to ₩1.59 trillion against ₩744.6 billion of total liabilities, and cash sits at ₩209.0 billion.
So the capital-intensive phase is finished and the debt that funded it is largely gone. Whatever you think of the earnings quality, this is a company that can now fund a decade of R&D out of a mediocre operating business plus the occasional deal, without going back to the market. That matters for a research-led model, because the alternative is signing deals from a position of needing the cash.
At ₩509,000 as of August 28, Hanmi is worth ₩6.52 trillion on 12.81 million shares. Trailing four-quarter net income is about ₩234.6 billion, giving 27.8x. Book value is ₩1.59 trillion, so 4.10x.
The trailing earnings include the Lilly payment. Tax it at the 30.0% rate Hanmi actually paid in the quarter and roughly ₩79 billion of after-tax income comes out, leaving something near ₩156 billion and a multiple north of 40x. That's before considering that Q4 2024 posted an actual net loss of ₩4.8 billion, which tells you how thin the base gets when no deal lands.
I don't think 40x is automatically wrong for this company. It's the wrong statistic, though. Hanmi's value is a probability-weighted stack of pipeline assets and the terms it can extract for them, and no trailing earnings number captures that in either direction. Price to book at 4.10x is at least measuring something stable.
The strongest argument against everything above is that I'm treating recurring events as one-offs. Hanmi has signed the Aptose deal for HM43239, the Lilly deal for sonefeglutide, and on August 26 an exclusive license to Roche's Genentech for HM17321, an obesity candidate, worth up to $2.3 billion with $190 million payable upfront. That's not a company that got lucky once. That's a business model, and a working one: discover long-acting peptides in Korea, hand global development to someone with a $60 billion commercial organization, keep Korea.
If deals arrive every year or two, then normalizing them away understates the company rather than the reverse. The HM17321 structure is particularly favorable, with Hanmi finishing Phase 1 and Genentech taking over from Phase 2, which caps Hanmi's spend while preserving milestones and tiered royalties.
The obesity positioning also deserves credit rather than skepticism. HM17321 is a UCN2 analog with a non-incretin mechanism aimed at preserving lean mass, which is the specific weakness of the GLP-1 class that Novo Nordisk and Eli Lilly currently dominate. Roche paid for a differentiated shot, not a me-too.
And the governance overhang that dominated English-language coverage of this company for two years has faded from the filings. What's in DART now is ordinary: a treasury share disposal in late July, routine insider reports in August.
The counter to the counter is simply timing. Deal income is real money, but it arrives when a partner decides it arrives, and a company whose base business shrinks between deals has to keep signing them. The $190 million from Genentech is comfortably larger than the Lilly payment, at any recent exchange rate, so Q3 will look spectacular and tell you even less.
The number that settles this is in the Q3 2026 report, and it isn't operating profit. It's cost of revenue. If Q3 revenue jumps on the Genentech upfront while cost of revenue stays near the ₩160 billion to ₩175 billion range it's held for eight quarters, then the base business still isn't growing and the whole result is the wire transfer again. If cost of revenue climbs meaningfully alongside revenue, the commercial side is finally moving and the thesis here weakens.
Two supporting checks. Watch the effective tax rate: another jump toward 30% is a marker that the quarter's profit is deal income rather than product income. And watch the separately disclosed R&D line against the ₩52.1 billion Q2 figure, since Hanmi's ability to keep signing $2 billion deals depends on that number staying high, not on it being cut to flatter a quarter.
The semiannual report for the period ended June 2026 was filed with DART on August 14, and the segment and licensing detail behind these figures is there.
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.