Alteogen Inc. (KRX:196170) told the Korean market on September 2 that it had signed an exclusive option and license agreement with Novartis Pharma AG covering subcutaneous versions of multiple Novartis biologics, built on Alteogen's ALT-B4 enzyme. Maximum value if everything goes right: US$3,223 million, or about ₩4.4165tn at the 1,370.30 won-per-dollar rate the filing itself applies. Sales royalties are on top of that and are not counted in it.
Then, near the bottom of the same disclosure, there is a small box most summaries skipped. Deferred item: the contract details, meaning which products, which milestones, what royalty. Reason: contractual confidentiality. Deferral deadline: January 3, 2053.
Twenty-seven years.
That box is the more interesting half of the announcement. The $3.22bn is a ceiling, and ceilings in licensing are cheap to print. What Alteogen has actually done is move its single largest 2026 event into a form where the only number an outside investor can verify is the one that eventually shows up in the revenue line, quarters from now, without a label attached to it.
The word order in the filing matters. Option, then license. Novartis holds multiple options to take exclusive rights to develop and commercialize subcutaneous formulations for several of its own products; it has not exercised them. The $3,223 million assumes all of them are exercised and all development and commercial milestones are hit, and the filing says so directly, adding that the total may shrink if trials fail or if sales come in small.
Alteogen also flagged that this is a conditional contract. Revenue recognition depends on clinical trials and marketing approvals, and the agreement can be terminated by the counterparty for convenience, for material breach, or on insolvency. Read that clause slowly. Termination for convenience means Novartis can walk away from the option chain without alleging anything went wrong, at which point the ceiling collapses to whatever has already been paid. The filing does say those payments are non-refundable, which is the meaningful protection here.
So what did Novartis commit to on September 2? A signing payment, due within thirty days. Its size is not disclosed.
American readers know the shape of this. Big pharma deals lead with a biobucks total and the upfront turns out to be a small fraction of it. The structure is familiar. What isn't familiar is the seal.
Korea's exchange disclosure regime lets a listed issuer defer parts of a material disclosure it would otherwise have to publish, but requires it to say three things on the face of the filing: what is being withheld, why, and the date the deferral ends. Alteogen used it and named 2053-01-03, which reads like the end of the royalty term rather than a real review date.
In the United States, a company signing an agreement of this size would generally file the contract itself as a material-agreement exhibit and redact the commercially sensitive terms. The redaction is narrow and the document is public. Investors argue over what's under the black bars, but the bars sit inside a document they can read.
Korea's version substitutes a promise. You get the maximum, the counterparty, the term, and a date decades out. You do not get the contract. For a company whose entire equity story is the quality of its licensing terms, that difference is not procedural. It's the whole disclosure.
The practical consequence: a deal with excellent royalty economics and a deal with mediocre ones would produce an identical filing, distinguishable only by the ceiling. And the ceiling is the number a licensor has the most latitude to construct.
Alteogen's reported numbers have gone somewhere genuinely different in two years. Revenue was ₩28.8bn in 2022 and ₩96.5bn in 2023. It reached ₩102.9bn in 2024 and then ₩215.9bn in 2025. Operating income went from a ₩9.7bn loss in 2023 to ₩25.4bn in 2024 to ₩106.9bn last year. Retained earnings, negative as recently as 2021, stood at ₩378.7bn at the end of June.
The quarters underneath are not a business ramping. They're a series of payment events. The first quarter of 2025 booked ₩83.7bn of revenue; the second booked ₩18.6bn. Then ₩49.0bn, then ₩64.5bn. This year ran ₩71.6bn and ₩68.9bn, so the first half of 2026 came in at ₩140.5bn against ₩102.3bn a year earlier. Growth, yes, but from a base that jumps around by four times between adjacent quarters.
The cash line lags in a way that's worth understanding before the Novartis payment shows up. Operating cash flow in the June quarter was ₩16.6bn against ₩34.2bn of operating income. In March it was ₩21.0bn against ₩39.3bn. Trade receivables tell you why: ₩86.0bn at the end of March 2025, ₩14.3bn three months later, back to ₩49.2bn in September, ₩41.9bn this March and ₩18.6bn in June. Milestone and signing payments get billed, sit in receivables, and convert. Full-year 2025 operating cash flow of ₩124.1bn did exceed operating income of ₩106.9bn, so the conversion works — it just runs on the contract calendar rather than the reporting one.
The Novartis signing payment is due by October 2, which straddles the quarter end. Whether it lands as third-quarter revenue, a third-quarter receivable, or a fourth-quarter cash item is not something the disclosure settles.
Alteogen closed at ₩298,500 on September 2, against a 52-week range of ₩206,213 to ₩430,123, for a market capitalization of ₩20.79tn — roughly $15.2bn at the filing's own exchange rate. The trailing four quarters produced about ₩254bn of revenue, ₩119.8bn of operating income and ₩198bn of net income. That works out near 82 times sales, 174 times operating income and 105 times earnings. Equity of ₩547.2bn at June 30 puts the shares around 38 times book.
The gap between the operating-income multiple and the earnings multiple is finance income, which contributed ₩56.4bn in 2025 and ₩35.9bn in the March quarter alone before turning slightly negative in June. Earnings-based multiples here are partly a multiple on the balance sheet, not the platform.
Against those numbers, the Novartis maximum of ₩4.4165tn is about 21% of market cap. Undiscounted, contingent on options nobody has exercised, spread across a term that runs to royalty expiry, and net of nothing. It is not a small number and it is not a near-term one. A reasonable way to hold both thoughts is that the market is paying today for the rate at which Alteogen signs these — four ALT-B4 agreements this year, per the company — rather than for any single one of them.
The concentration is the obvious one. Nearly all of this rests on a single enzyme, and a single enzyme invites a single lawsuit. The patent fight with Halozyme has gone Alteogen's way so far: in May 2026 the US Patent Trial and Appeal Board declined to institute Halozyme's petition to invalidate Alteogen's PH20 manufacturing patent, and separately finalized the invalidation of one Halozyme patent that Merck had challenged. Those are real wins. They're also two decisions in a dispute that a platform licensor can expect to keep paying legal attention to for years, and partners weigh that when they price a deal.
The termination-for-convenience clause deserves more weight than the ceiling. It converts the option chain into something Novartis renews by choice each time, and each non-exercise is invisible — there's no filing that says an option lapsed quietly.
Then the seal cuts the other way too. If the royalty rate on the Novartis agreement is better than the market assumes, holders won't learn that from a disclosure either. Confidentiality is symmetric; the discount it creates is not.
Currency is the quiet one. Payments come in dollars, books are in won, and the 2025 finance income line shows how much of reported profit can come from that translation rather than from operations.
Three things, in order of how soon they arrive.
The third-quarter report, due in November, and specifically whether a Novartis signing payment appears in revenue, in contract liabilities, or not yet at all — and what trade receivables do alongside it. That is the first hard evidence of what Novartis actually paid to hold these options.
An amended version of the September 2 filing. Korean issuers correct and reissue the original disclosure when a deferred item is released early, so any restatement of that document is where a real number would first appear.
Novartis's own reporting. A named subcutaneous program in its pipeline disclosures would tell you which product the first option covers, which is the fact that turns a ceiling into a forecast. Until one of those three lands, the $3.22bn is a headline, and the company's own filing is unusually honest about saying so.
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.