NCSoft Stock Trades on Profit Its Games Did Not Make

I remember the day Aion 2 launched, and not because of the game. Korean press reported 37.2 billion won of short selling on that single session, about $26.3 million, the largest one-day figure in the domestic market at that point. On November 19, 2025, in the session the new title went live, NCSoft stock fell 14.61% and closed at 191,700 won, roughly $135.37 (reported by The Elec and Newsis, both Korean outlets, and I am paraphrasing their wording into English). I wrote in my file that the market was not judging the game, it was judging the monetization model, and then I stopped opening the ticker for nine months.

At the August 10, 2026 close the shares finished at 243,500 won, about $171.95. Against the 191,700 won launch-day close, that is 27.02% higher, a move of 51,800 won. Pull up any screen and the ticker shows a trailing price-to-earnings ratio of 15.2 times, which for a game publisher reads as unremarkable. I went to check what that 15.2 was dividing, and the answer stopped the rest of the work.

Contents15 min read

A note on where NCSoft stock trades and how a foreign account reaches it

NCSoft, which shortened its corporate name to NC during 2026, is listed on the KOSPI under ticker 036570. The KOSPI is the senior board of the Korea Exchange, the venue for large-capitalization Korean issuers; the KOSDAQ is the separate junior board where most smaller technology and biotech names sit. NCSoft sits on the senior board.

There is no American depositary receipt for this company, so a US-based account reaches it one of three ways: direct Korean market access through a broker that supports it, such as Interactive Brokers; indirect exposure through a Korea country fund like EWY or FLKR, where the weight is small and diluted by semiconductors and banks; or not at all. Foreign ownership stood at 37.98% on the August 5, 2026 screen, so the register is genuinely international even without a US listing. I am writing this note as an observer, and everything below is arithmetic I ran myself on published Korean disclosures.

The trailing multiple on NCSoft stock divides profit the games did not earn

Price-to-earnings divides the share price by earnings per share, so the question is which earnings. The company reported its 2025 full-year results on February 10, 2026: revenue of 1.5069 trillion won ($1.064B), operating profit of 16.1 billion won ($11.4M), and net income of 347.4 billion won ($245.3M).

Operating margin for the year was 1.07%, which is 16.1 billion divided by 1,506.9 billion. Net income came in at 21.58 times operating profit, which is 347.4 divided by 16.1. Selling games produced 16.1 billion won of operating profit, and the figure that reached the bottom line was twenty-one times larger. The 331.3 billion won difference came from somewhere other than operations.

This is where I stopped, because I could not settle where. One Korean outlet attributed it to acquisition consideration booked as intangible assets, but that explanation addresses why operating profit was small, not why net income was large. What the company itself said on the call is a separate item: per Dealsite, excluding roughly 20 billion won of one-off retirement payments lifts 2025 operating profit to 36.9 billion won. That adjustment sits inside operations and does nothing to explain the 331.3 billion won outside it.

So I am not writing a breakdown of the 347.4 billion won. Leaving a blank is better than filling it plausibly. One fact does survive: the 15.2x multiple is not dividing profit that the game business generated. Anyone using it as evidence that the shares are cheap for a publisher is standing on the wrong number.

Two multiples sit on one screen and describe two different companies

The Korean company data screen dated August 5, 2026 carries both: a trailing multiple of 14.67 times at that day’s 235,000 won mark, and a forward multiple of 10.90 times built on a 2026 consensus EPS of 21,561 won. Recomputed at the August 10 close of 243,500 won, those become 15.20 times and 11.29 times.

The gap between them is not a growth premium. The numerator is the same price. The denominators are earnings from two different years, and those two years genuinely belonged to different companies.

The quarter that proved the game, and the NCSoft stock price that did not move

Segment revenue in the first quarter of 2026 ran PC online 318.4 billion won, mobile 182.8 billion won, mobile casual 35.5 billion won, and other 20.7 billion won. Those four add to exactly the 557.4 billion won reported. Aion 2 contributed 136.8 billion won ($96.6M), or 42.96% of PC online revenue, and Lineage Classic added 83.5 billion won ($59.0M), according to Korean coverage from PNN and eFocus.

The geographic mix moved too: 58% domestic, 27% Asia, 15% North America and Europe, taking overseas revenue from 35% to 42% year over year. Total revenue rose 54.7% against the same quarter a year earlier.

So the title that the market marked as unproven on launch day became 43% of PC revenue two quarters later. And here the second reversal arrives, the one I find more interesting. The share price barely moved once that proof landed. Vendor price history shows a one-month return of -0.6% and a three-month return of -1.59%. Stretch to twelve months and it is +28.11%, six months +15.69%. The rally happened before the evidence, not after it.

My reading is that the first quarter is already in the price and the market is now marking a quarter nobody has seen. The FnGuide consensus carries second-quarter revenue of 683.3 billion won ($482.5M) and operating profit of 132.0 billion won ($93.2M), as cited by SBS Biz. If that holds, first-half revenue reaches 1,240.7 billion won ($876.1M) and first-half operating profit reaches 245.3 billion won ($173.2M), which alone would be 15.24 times the whole of 2025. Against the upper end of the company’s own 2 trillion to 2.5 trillion won revenue target, given on the February 10, 2026 call, the first half would fill 49.63%.

A detail from the company call that the quarterly headline hides

One line from the February 10, 2026 call is worth pulling out because it changes how I read every Aion 2 revenue figure printed since. On that call the company gave two numbers for the title’s fourth-quarter contribution: 94.1 billion won of gross bookings against 77.4 billion won recognized as accounting revenue, a 16.7 billion won gap created by its deferred recognition policy, per the Dealsite coverage linked above. Management also described monthly revenue running around 70 billion won from January onward, and membership purchases climbing from one million characters in early January to 1.5 million by February 9.

The practical consequence is that reported revenue for this title lags cash collected, in both directions. When the title is scaling, the printed line understates it; when engagement rolls over, the printed line will hold up for a period after the cash has already turned. That is why my thesis-breaking condition below is written against Aion 2 revenue specifically instead of against consolidated operating profit. Consolidated profit can be carried for a quarter by recognition timing on a title whose live spending has already peaked, and I would rather be wrong early on the title line than late on the consolidated one.

Market value for NCSoft stock still contains 2,071,063 treasury shares

This is the part that made me write the note.

The August 5, 2026 screen shows 21,544,022 shares issued, of which 2,071,063, or 9.61%, are treasury stock. Yet NH Investment and Securities analyst Ahn Jae-min raised his valuation to 330,000 won ($233.03) citing a board resolution to cancel 9.9% of held treasury shares, in Korean coverage dated February 26, 2026 (BizTribune). More than five months after that resolution was reported, the August 5 screen still carries the block at 9.61%.

I am not merging those two numbers. The reported 9.9% and the displayed 9.61% may rest on different reference dates and different bases, and I could not confirm whether the cancellation was executed. The one thing I confirmed is that the share count on the screen has not come down.

So I ran the arithmetic twice. A separate vendor screen shows 21,544,148 shares, 126 more, a difference of 0.0071% that does not move market value and that I therefore did not use.

One correction I want on the record, because I have seen it stated the other way. Treasury shares are already excluded from the weighted average count used for earnings per share, so executing the cancellation does not mechanically raise EPS and does not change the price-to-earnings ratio. What changes is the market value line and any multiple built on market value. Reading a treasury cancellation as an automatic lift to earnings per share conflates it with a capital reduction.

Five brokerages covering NCSoft stock leave a 208.9 billion won hole

Five named Korean brokerages have published a forward view. Every one of them was dated before the May 13, 2026 first-quarter release, and I could not find an updated note reflecting that quarter. Read the table with that limit attached.

What holds my attention is not the level of those valuations but the spread underneath them. Yuanta’s 584.9 billion won ($413.0M) and the market consensus of 376.0 billion won ($265.5M) that Yuanta itself cited are 208.9 billion won apart, a ratio of 1.56 times. When professional estimates of the same year’s operating profit differ by more than half, any valuation resting on those estimates carries the same width. I treat the spread itself as the finding. The August 5, 2026 screen showed a consensus valuation of 367,500 won ($259.52) and an average analyst rating score of 4.00, and that blend almost certainly still contains the pre-quarter notes above.

Yuanta’s Lee Chang-young flagged something else worth repeating: despite the commercial success of Aion 2, the prevailing view of this company in the market stayed rooted in negative sentiment instead of a reading of the numbers, as reported by Money Today in March 2026. I kept the ticker off my list for nine months, so I am one instance of that sentiment.

What Take-Two shows about NCSoft stock that the Korean screen does not

For a global comparison I wanted a publisher whose valuation rests on a launch instead of a run rate, and the cleanest case is Take-Two Interactive on the Nasdaq. My reason for choosing it is specific: Take-Two has no trailing multiple at all. The market is forced to price it on a forward number because there is no trailing one to misuse. NCSoft has both, and having both is exactly what creates the confusion in this note.

At the same August 10, 2026 close, Take-Two finished at $253.57 with a market value of $47.41B, a trailing price-to-earnings ratio shown as not applicable because the company is loss-making, and a forward ratio of 30.42. Trailing revenue was $6.69B, trailing net income was negative $320.40 million, earnings per share negative $1.73, share count 186.98 million, and the 52-week range $187.63 to $265.94, per StockAnalysis. I checked the screen against itself before using it: 253.57 multiplied by 186.98 million reproduces $47.41B, and negative 1.73 multiplied by the same count gives negative $323.5 million against the reported negative $320.40 million, a 0.97% gap that rounding in the EPS figure explains.

Set the two side by side and the size difference does most of the talking. Take-Two carries 12.80 times the market value of this company on 6.3 times the trailing revenue, while running a loss. A US investor pricing Take-Two has to underwrite a launch that has not happened. A US investor pricing this one has to do something harder: decide which of two published multiples is describing the business, when one divides last year’s non-operating windfall and the other divides a half-year nobody has seen.

The other side of NCSoft stock, and what I could not verify

Here is the material that works against my interest.

First, the three-year revenue growth rate runs at negative 16.32% annualized on the vendor’s calculation. One strong quarter is one observation, not a reversal of that path.

Second, the dividend per share fell from 8,550 won in 2020 to 1,150 won in 2025, down 86.55% across five years on the regulatory filing history. Whether the payout follows recovering earnings will not be visible until the 2026 annual settlement.

Third, the August 10, 2026 close sits 29.11% below the 250-day high of 343,500 won ($242.57) and already 38.27% above the 250-day low of 176,100 won ($124.36). Measured from the low, this is not a cheap zone.

Fourth, the sources disagree on the Aion 2 global launch date. Coverage of the company call points to the third quarter of 2026, an April 9, 2026 summary in The Elec put testing in the third quarter and launch in the fourth, and an early-August summary named September. I am not picking one. The fact that half of the next quarter’s case rides on a date that appears three different ways is itself the clearest description of where this name stands.

Fifth, several vendor fields would not reproduce from primary figures, so I dropped them all: operating cash flow, free cash flow, EBITDA, and total equity. Book value per share also splits between two screens at 187,076 won and 173,514 won, which puts price-to-book at either 1.30 or 1.40 times. I could not settle it and therefore did not build any argument on book value.

My position is watching without a holding. Market value of 5.246 trillion won places this outside the top hundred Korean issuers by size, and I do not take trading positions in that band. Separately from that rule, the reason I am not buying is not that the shares look expensive but that neither printed multiple is usable. The 15.2 divides profit the games did not earn. The 11.29 divides six months nobody has reported.

Where this connects to what I have already written

I asked a version of this question on another Korean publisher. In my note on Krafton and the bill that grew behind a hit, the cost arrived after the success was confirmed. Here the order is reversed: the cost passed through first and the success followed, while the multiple is still dividing the old year.

On multiples that split inside a single sector, my piece on KakaoBank carrying a peer’s multiple without the peer’s growth reached the conclusion I keep reusing. A multiple does not describe a company. The profit that built the multiple describes the company.

On the ownership side, my Coway note covering Netmarble’s stake purchase connects directly, since Netmarble also sits on this company’s register as a substantial holder.

Questions and notes on NCSoft stock

Basis, and the shelf life of this note

Every price and multiple here rests on the August 10, 2026 close of 243,500 won, and the indicator set was pulled from Kiwoom data refreshed at 20:16 Korean time on August 10, 2026. Share count and treasury ratio come from the Korean company data screen dated August 5, 2026. Dollar conversions throughout use 1,416.10 won per dollar, the Seoul foreign exchange market close of August 7, 2026; the August 10 closing rate was not available to me at the time of writing, so the currency reference sits one trading day behind the price reference and I have used a single rate across the whole note instead of mixing two.

When the second quarter is confirmed, every multiple above has to be rebuilt. Half of the 245.3 billion won first-half operating profit figure is still an estimate, and executing the treasury cancellation would rewrite the market value section outright. I could not establish the reporting date.

What I missed on November 19, 2025 was not the fate of a game but the kind of question being asked. The market that day was not asking whether the title would work, it was asking when anyone would know how well. I collapsed those two questions into one and closed the file. Nine months later, in front of the same screen, I ran the multiple three separate ways to avoid repeating that. All three came out different, and that is the honest description of where this name sits.

Related reading

• Cheil Worldwide Stock Is 58% of the Ad Index It Just Beat
• KEPCO KPS Stock Made Money Every Quarter Its Cash Did Not