Gravity is a Seoul based mobile and online game publisher whose entire earnings stream traces back to the Ragnarok Online franchise, and the company has only just begun returning its dormant cash to minority holders after a long silence on distributions.
The board approved a first ever interim dividend on August 7, three and a half months after the annual report stated the company had no intention to pay dividends in the near future. That reversal of a stated policy, taken in a single quarter, is the event that redefines the investment case. The per share payout of 4,400 won followed an interim shareholder register closure in mid June, and it arrives alongside a record half for the franchise with first half revenue near 323,762 million won.
The tension is structural, and it sits in three places. One franchise, one dominant parent in Tokyo that is also a licensee, and a platform commission stack that absorbed 58 percent of revenue in the first half all cap what the cash can buy minority investors.
The catalyst is the dividend payment date in September, and the question the next earnings cycle answers is whether the payout is a policy or a one time gesture.
Gravity develops and publishes online and mobile games, but the honest description is narrower than the product list suggests. The company is a franchise operator for Ragnarok Online, the action adventure MMORPG it launched in August 2002, and every significant revenue source since then is a regional relaunch, a platform port, or a spin off of that single intellectual property. The relevant peer set for comparison is other Korean live service publishers such as NCSoft, Webzen and Com2uS, and Japanese controlled game subsidiaries, yet Gravity is the purest single IP expression of that group, with no second franchise anywhere near the revenue scale of the first.
The business model splits into two engines. One engine is direct operation, where Gravity or a subsidiary runs servers and collects micro transaction revenue, now dominant at 81.2 percent of 2025 revenue from mobile games. The other engine is licensing, where third party publishers such as GungHo in Japan, Xindong in China and local operators in Europe and Vietnam run the game and pay royalties and license fees. The mix between these two engines determines the margin profile, because the direct operation engine carries platform service fees and commissions, while the licensing engine is high margin but exposed to the licensee spending and renewal decisions of a partner.
Governance sits on top of both engines. GungHo Online Entertainment, the Japanese company that operates Ragnarok in Japan, beneficially owns 59.3 percent of Gravity and controls the board, having taken majority control in 2008. Gravity is a dual listed issuer, trading in Seoul and on Nasdaq as American depositary shares, each representing one common share, with roughly 6.95 million shares outstanding and a float small enough that single launch headlines move the stock several percent. The combination of a controlled company structure and a concentrated franchise means minority holders are long a cash flow whose allocation decisions are made substantially by a parent with its own balance sheet needs.
The strategic question for the thesis is whether the franchise can keep funding itself through regional re entry without dilution of the margin. The answer so far is yes on revenue and no on profit, a gap that the next sections quantify.
The moat is a brand that survived its own developers leaving. Ragnarok Online has run for two decades and holds 91 commercial markets, and the franchise now spreads across a long ladder of titles, each a variant of the same world. Ragnarok M: Eternal Love launched in 2017 with Dream Square as co developer and remains the franchise's global anchor. Ragnarok M: Classic, the direct sequel, launched in Southeast Asia in early 2025 and reached Taiwan, Hong Kong and Macau shortly after, and in the fiscal year that followed it became the single largest revenue driver of the company at 27.6 percent of total revenue. A year earlier the largest single title was Ragnarok Origin at 42.1 percent of revenue, so the franchise's internal hierarchy is itself in flux as each new variant absorbs and then passes through the spending of a mobile game audience.
The technology story is that the franchise outlived the companies that built it. The development partnership with Dream Square expired in March 2021 without renewal, and Gravity now holds direct contractual relationships with Xindong, the Chinese developer behind the Eternal Love and Classic lines, and with Shanghai Rexue. The latest annual report discloses that 81.3 percent of revenue came from titles developed by third parties, of which a third came through Dream Square workstreams, so the operating company increasingly resembles a rights holder and regional distributor for games it does not fully build. The moat in that structure is the intellectual property itself and the regional service network, not an in house production engine.
The pipeline is the franchise being re mined for every region and platform. Ragnarok: The New World, the first open world MMORPG built on the Ragnarok IP for both mobile and PC, launched in Taiwan, Hong Kong and Macau in January 2026 and in Southeast Asia in July 2026, with a wider global rollout planned for the fourth quarter. Ragnarok Origin Classic launched in Korea, Taiwan, Hong Kong, Macau and Southeast Asia in March 2026. Ragnarok Zero, the PC revival of the original game with a subscription model, opened beta in May 2026. A new title, Ragnarok M: Eternal Love 2, co developed with Xindong, received a Chinese ISBN in July 2026, the regulatory gate that unlocks the mainland market, the single largest gaming market in Asia.
The durability question is whether a 24 year old brand can keep generating re entry spending from a generation that never played the original. The concurrent user data in the annual report is the telling detail. Peak concurrent users for the original PC game in Japan declined from roughly 7,300 a year ago to roughly 6,260 in the most recent quarter, while Thailand held near its prior levels, so the long tail is stable in some markets and slowly thinning in others.
Full year revenue grew 11.9 percent to 560,548 million won. Profit attributable to the parent fell 20.6 percent to 67,464 million won, and that divergence between the top line and the bottom line is the load bearing fact of the year. The growth came from the Classic relaunch in new regions and from Ragnarok X: Next Generation entering the Americas and parts of Europe, while the profit decline came from the cost of buying that growth. Commission costs rose 19.3 percent, now running near three hundred twenty six thousand million won. The gross margin line is where the answer sits, and it fell from 38.7 percent to 35.0 percent as the new regional launches carried heavier platform service fees, royalty payments and outsourcing fees. Selling and administrative expenses rose 15.6 percent. Advertising spending rose by about fifteen and a half percent and drove most of that increase across the launch slate.
The first half of 2026 shows the pattern continuing, and the two quarters carry different stories. One quarter ran at record revenue, and the other faded against the year ago base. Revenue in the first quarter was 161,878 million won, up 17.8 percent year over year. The second quarter revenue of 161,884 million won ran flat to that level and down 5.2 percent year over year. The cost line barely moved, with cost of revenue near 109,541 million won in the second quarter, and that held gross margin compression intact from the prior fiscal year. Net profit of 24,341 million won in the second quarter rose 83.8 percent year over year, but the annual report's own explanation is that the increase came from lower advertising expense year over year rather than from top line strength, a profitable quarter built on cost restraint against a declining revenue base.
The balance sheet is the counterweight to that revenue story. Cash and short term financial instruments totaled roughly 611,615 million won at the end of the first quarter, against no meaningful financial debt and a market capitalization near 500 million USD. The balance grew further into the second quarter, and the company generated 76,092 million won of operating cash flow in the most recent full year, parking the surplus in money market instruments with maturities of one year or less, a defensive posture that has accumulated for years without a single distribution to shareholders.
The tax line deserves a sentence of its own because it affects the cash the minority can expect. The effective tax rate rose from 20 percent to 26 percent year over year, driven by deferred tax liabilities related to the investment in subsidiaries. Income tax expense rose 10.2 percent to 23,629 million won. Even as profit before tax fell 14.5 percent, the tax line is now a mild headwind to the free cash flow that funds any dividend policy.
The forward revenue case rests on three named thesis variables: the regional launch cadence, the China gate, and the capital allocation decision. The launch cadence is the most visible variable. Ragnarok: The New World reaches a wider global audience in the fourth quarter, Ragnarok Zero: Global launched in Southeast Asia, Europe and Oceania on August 18, and Ragnarok M: Eternal Love 2 moves toward a Chinese release with Xindong as the local partner. The China gate is the highest leverage item in the entire thesis, because the mainland market is the largest gaming audience in Asia and the ISBN approval of July 23, 2026 removes the single biggest regulatory barrier that has kept the franchise out of that market for its entire mobile history.
The mechanism of the China play deserves explicit statement. Under the historical arrangement, Xindong holds the Chinese distribution rights for the Dream Square era titles, and the revenue Gravity records from China arrives as royalties and revenue share rather than as operated micro transaction income. That structure keeps the asset light, with Gravity not bearing the server and marketing costs of the mainland release, but it also means the Chinese contribution is a slice of another company's economics, and the quality of the Xindong release, its traffic allocation, and its pricing decisions all sit outside Gravity's direct control.
The third variable is the one the market has been waiting on for a decade, capital allocation. The annual report filed in April 2026 stated that Gravity has not declared or paid any dividends on its common shares since inception and had no intention to pay dividends in the near future. The first interim dividend of 4,400 won per share, totaling roughly 30.6 billion won, reverses that statement in a single quarter. The company framed the move as part of a capital allocation strategy that divides surplus among growth investment, strategic investment, liquidity reserves and shareholder returns, and it disclosed a total growth and strategic investment plan of roughly 200 million USD through 2028, so the dividend is not presented as liquidation of the treasury but as the first tranche of a return policy.
Execution risk on the outlook is the hit driven nature of the top line. Second quarter 2026 revenue already fell 5.2 percent year over year while the launch slate was at full tilt, and the same quarter showed the older titles, Ragnarok Origin and Ragnarok M: Classic in their mature markets, declining as the newer variants absorbed the audience. The risk is that the next cohort of relaunches, The New World and Origin Classic, follows the same arc, spiking on day one and then decaying faster than the regional gaps can fill, leaving the revenue line flat to negative even as the cost base stays loaded with advertising and commissions.
The single franchise risk is the most obvious and the hardest to hedge. One intellectual property, launched in 2002, now funds the entire company, and the internal evidence shows the long tail is not infinitely elastic. Peak concurrent users on the original PC game in Japan declined through the quarters disclosed in the annual report, and the franchise's answer to decay has been to port and relaunch the same world rather than to build a new one. The non Ragnarok catalog, Pigromance, White Chord, Twilight Monk, Schop Hero and the JLPGA Heroine Collection, generated enough of a footprint to warrant impairment charges, but not enough revenue to stand alone, so the franchise concentration is not a temporary phase but the business model.
The platform stack risk is the quiet one. The structure is that commissions to Apple and Google, plus royalty and outsourcing fees to regional partners, sit between the franchise and the free cash flow. Those charges consumed 326,579 million won in the most recent fiscal year, roughly 58 percent of revenue, and the same structure applies into the current year. The gross margin slid from 38.7 percent down to 35.0 percent across that same span. A further shift toward operated markets, or a change in platform fee policy, would move the free cash flow line by tens of millions of won annually, which against a cash balance near 650 billion won is manageable for the company but material for a minority holder who can only see the cash leave through distributions.
The governance risk is the controlling parent. GungHo owns 59.3 percent and is simultaneously the Japanese licensee, a competitor for the same franchise's spending, and a company with its own capital needs, having funded a 5 billion yen buyback of its own shares in the first half of the current year. The first dividend at Gravity followed the closing of Gravity's shareholder register while GungHo faced pressure from an activist shareholder of its own, and the plausible reading is that the distribution is partly a response to the parent's need to show cash yield from its most valuable asset. For the minority, that cuts both ways, the parent has incentive to make the payout visible, but the parent also controls the timing, the size, and the growth spending that precedes it, and the related party flows between the two companies are not fully priced into the minority's economics.
The downside scenarios are therefore not a game failure but a cash trap. In the bear path, the relaunch cadence decays, revenue trends below the 500,000 million won level, and the dividend stays a one time gesture as the treasury quietly funds the 200 million USD investment plan and the parent's needs. In that path the equity is a 500 million USD claim on an operating business the market values near zero after the cash, a position that earns a dividend yield on a shrinking base. In the base path, the cadence holds, the China title contributes a royalty stream, and the dividend becomes a modest annual policy that returns a fraction of the roughly 76 billion won annual operating cash flow. The base path is the one that makes the current price defensible, and it depends on all three thesis variables moving in the same direction at once.
The valuation framework starts from the balance sheet rather than the income statement, because the market capitalization sits close to the cash balance. At a share price near 72 USD the market values the company near 500 million USD. The share count is small, near 6.95 million shares outstanding, and the treasury stands out against it. Cash and short term instruments stood at 649,231 million won, roughly 419 million USD, at the end of the second quarter. Subtracting the treasury leaves an enterprise value near zero, and the operating business, with its 13.8 percent operating margin and 76 billion won of annual operating cash flow, is effectively being sold at no premium to the cash that funds it. The framework therefore reduces to two questions, what multiple is fair for the operating franchise, and how much of the treasury belongs to the minority on a sustainable basis.
On the earnings multiple, the trailing price to earnings of roughly 8.2 times sits at the low end of the global live service gaming peer set. Established franchises with demonstrated longevity typically trade above 10 times normalized earnings, and the gap between the two readings is the space the controlled company structure and the single franchise risk occupy. The full year profit, near 67,000 million won, is the denominator that holds the multiple down. That discount to peers is the price the market charges for the concentration in a single franchise and a board that answers to a controlling parent.
The quantified scenarios run across three cases, and the bridge from earnings to equity value is the same in each. Bear case, revenue drifts toward the low five hundred million won range, the tax and commission lines hold, and profit settles near fifty five thousand million won. At 8 times earnings the operating business is worth roughly 440 million USD. Adding the treasury, the equity supports roughly 80 USD per share, but only if the dividend continues at the first interim run rate and the parent does not redirect the treasury to its own needs. Base case, revenue holds near five hundred sixty million won, the China title adds a royalty line, and profit normalizes near seventy thousand million won. At 10 times the operating business is worth roughly 485 million USD. Adding the treasury back, the equity supports roughly 130 USD per share. Bull case, the China release converts into a meaningful royalty stream, the New World global launch lands, and profit extends toward eighty five thousand million won. At a dozen times earnings the operating business is worth roughly 700 million USD. Adding the treasury back, the equity supports 170 USD per share.
The explicit counterargument to this framework is that the treasury is not a permanent asset of the minority. The same board that approved the first dividend controls a parent that has its own buyback program, and the 200 million USD investment plan through 2028 is a disclosed claim on the surplus. If the market is correctly pricing the cash as partly committed, the zero enterprise value reading overstates the minority's claim, and the relevant multiple applies to a smaller earnings base. The dividend yield of the first interim payout, roughly 4.4 percent annualized at the current price, is the cleanest test of which reading is right, because a continuing policy validates the base case and a one time payment confirms the bear case.
Gravity is a franchise operator being valued like a dormant cash account, and the first dividend changes the frame without resolving it. The judgment on the equity is that the current price is justified only under the base case, where the launch cadence holds, the China gate converts into a royalty line, and the payout becomes a modest recurring policy, and it is cheap under that case because the treasury is sitting near the entire market capitalization with the operating franchise attached at no premium. The judgment is also that the bear case is more likely than the multiple suggests, because the revenue line already turned negative year over year in the second quarter while the cost base stayed loaded, and because the parent's structural incentive is to treat the treasury as a group asset rather than a minority asset.
The dividend decision is the strongest single data point in the current cycle, and it deserves weight in both directions. It is positive because it proves the board can move the treasury, and it is a warning because the same annual report had stated no intention to pay, which means the allocation stance can reverse as quickly in the other direction. The 4,400 won per share payout against 76 billion won of annual operating cash flow is a real but small return, and the size of the gap between the payout and the cash balance is exactly the space where the parent's needs and the minority's expectations collide over the next two years.
The data signals that resolve the thesis over the coming quarters are the third quarter revenue trajectory, the size of any second interim dividend announcement, the first disclosed revenue contribution from the China release, the gross margin line as the New World global launch moves from a regional spike to a mature run rate, and the treatment of the 200 million USD investment plan in the 2026 annual report. Those five items, in that order of importance, carry the entire difference between the bear and bull readings, and none of them has yet produced a reading that favors one side decisively.