Nexon Q3 2026: Cancel six games, bench 300 devs, and watch ARC Raiders fall out of Steam's Top 100 anyway

Back in August, when I wrote about Nexon's Q2 numbers, the whole joke was watching Nexon hand Patrick Söderlund a freshly invented Executive Chairman seat off the back of ARC Raiders' launch window and then pretend one extraction shooter's honeymoon period was a repeatable business blueprint. At the same time, they put The First Descendant and Dungeon&Fighter Mobile on a slide labelled "What Did Not Work" for failing to hold players long term, even as ARC Raiders had already lost nearly 90% of its Steam concurrency. Three months later, Nexon's Q3 2026 report landed on 10 November, and the follow-up is even funnier. MapleStory is carrying an even larger share of the entire business, Dungeon&Fighter is bleeding out in China, subsidiary studio Nexon Games swung into a quarterly loss after Blue Archive imploded over a greed-driven gacha rewrite, ARC Raiders has already fallen back out of SteamDB's Global Top 100 Sellers despite a half-price sale and a free weekend, and Söderlund spent the earnings call contradicting both himself and his own CEO inside the same 45-minute window.

The Q3 headline numbers look great until you check Q4:

On paper, Nexon had a strong third quarter, beating the top end of its own operating profit guidance by nearly 30% and sitting on enough cash for another huge shareholder payout:

Q3 revenue: ¥138.4 billion (£654m / USD 866m), up 17% year-on-year

Q3 operating income: ¥41.8 billion (£198m / USD 261m), up 11% year-on-year

Q3 net income: ¥38.1 billion (£180m / USD 238m)

Full-year 2026 revenue outlook: ¥519.4bn to ¥530.6bn (£2.46bn–£2.51bn / USD 3.25bn–3.32bn)

Full-year 2026 operating profit outlook: ¥135.2bn to ¥142.9bn (£639m–£676m / USD 846m–894m)

Cash and special dividend: ¥622.4 billion (£2.94bn / USD 3.89bn) in cash, with a ¥415 per share special dividend worth roughly ¥324 billion (£1.53bn / USD 2.03bn) paying out on 25 November

Nexon's Q3 Business Wire release and investor relations portal lay out the headline results. The trouble starts the moment you scroll down to the Q4 guidance:

Q4 revenue outlook: ¥107.7bn to ¥118.9bn (£509m–£562m / USD 674m–744m), down 4% to 13% year-on-year

Q4 operating profit outlook: ¥4.0bn to ¥11.7bn (£19m–£55m / USD 25m–73m), collapsing from ¥41.8 billion in Q3

That profit drop is what happens when two of last year's big Q4 launches, ARC Raiders and MapleStory: Idle RPG, hit their one-year comparisons right as marketing spend ramps up for late-November releases like Azur Promilia in Korea and TEMPPAL: OVERGEARED. Worse for margins, Nexon is losing ultra-high-margin royalty income from China and replacing it with lower-margin, fee-heavy titles like MapleStory: Idle RPG, MapleStory Worlds, and FC ONLINE. When the easiest profit stream shrinks, a modest revenue dip takes most of the operating profit with it, as shown in Quartr's Q3 and Q4 summary.

MapleStory is still carrying the whole company, and that is a problem:

In Q2, MapleStory jumped 63% year-on-year to save the quarter. In Q3, it had to work even harder, surging 72% year-on-year to set a second consecutive all-time quarterly revenue record. Global MapleStory roughly doubled thanks to MapleStory Classic quadrupling revenue in China and doubling in Taiwan, Western revenue grew 33%, MapleStory Worlds roughly doubled, and MapleStory: Idle RPG continued printing cash despite eating a refund scandal earlier this year over an undisclosed coding error. FC also rebounded to double-digit growth thanks to a record quarter for FC ONLINE. Together, MapleStory, Dungeon&Fighter, and FC accounted for ¥108.5 billion (£513m / $679m), or roughly 78% of Nexon's entire Q3 revenue.

If a 23-year-old 2D MMO is the only reason your record-breaking year has not turned into an embarrassing slump, you have a concentration problem. Even analysts on Tuesday's call pressed management on how MapleStory can possibly keep growing next year against a +72% outlier baseline, and CEO Junghun Lee openly admitted Nexon has no new MapleStory titles scheduled for reveal in 2027. They have already spun out MapleStory, MapleStory M, MapleStory Worlds, MapleStory Classic, MapleStory: Idle RPG, and a crypto spin-off, leaving nothing left in the tank to clone. Relying on one legacy brand to bail out every misstep elsewhere works right up until the comparison gets too big to clear.

Dungeon&Fighter is sinking fast in China:

While MapleStory prints money, Nexon's longtime cash cow in China keeps sliding down the drain. Dungeon&Fighter franchise revenue fell another 33% year-on-year in Q3, following drops of 26% in Q1 and 44% in Q2. Back in April, DnF Mobile had already fallen so hard from its 2024 launch high that Nexon handed primary development of the Chinese mobile version over to Tencent for hyperlocalisation. That panic move did not save Q3 either, as September's National Day updates and new raid failed to drag players back on either PC or mobile.

The breakdown inside Nexon's own China PC metrics is textbook live-service decay: monthly active users and paying users both fell quarter-on-quarter and year-on-year, while average revenue per paying user went up. Fewer people are playing, fewer are buying, and Tencent and Nexon have simply been squeezing the remaining whales harder while the in-game gold economy collapses, pug raid requirements turn hostile, and China's wider market floods players with newer PC and mobile alternatives for their time. It got bad enough that Nexon spent Q4 shifting China DnF toward prioritising engagement over monetisation, forced Junghun Lee into crisis meetings with Tencent, and delayed Arad: Idle RPG out of Q4 into the first half of 2027. Because Tencent pays for servers and operations in China while Nexon collects clean royalties, every yen lost on China DnF hits Nexon's profit margin twice as hard as a miss anywhere else.

Nexon Games took the hit on Blue Archive and The First Descendant:

Over at subsidiary developer Nexon Games, Q3 was downright ugly. The studio behind Blue Archive and The First Descendant reported quarterly revenue of ₩38.6 billion (£21.9m / $28.9m), an operating loss of ₩22.1 billion (£12.5m / $16.5m), and a net loss of ₩28.8 billion (£16.3m / $21.5m), according to DigitalToday's coverage of its Q3 filing. Some of that swing came from heavy spending on upcoming projects like Dungeon&Fighter: ARAD, Fareidolia (formerly Project RX), and Woochi the Wayfarer, plus the exact accounting reversal I flagged in Q2 coming back the other way as a negative base effect on stock-based compensation. The rest came straight from Blue Archive and Hit 2 declining.

Blue Archive is a case study in what happens when Patrick Söderlund's March "profit floor" mandate collides with an actual otaku fanbase. On 26 July, with three days' notice and voice actors made to read the announcement off a script, the Japanese service scrapped its five-year-old 200-pull Spark system and replaced it with a HoYoverse-style 50/50 pity that resets the moment you pull a featured unit. Steam reviews promptly collapsed to Overwhelmingly Negative, Blue Archive's Japanese revenue dropped year-on-year in Q3, and Yostar quietly pinned the change on Nexon Games in customer support replies. By 25 October, Nexon announced it is taking 100% of Japanese publishing back from Yostar starting January 2027, shutting down the current JP PC client for a migration so Nexon can pocket Yostar's publisher cut after Yostar spent nearly six years turning Blue Archive into a cultural phenomenon in Akihabara. Then, on 8 November, Nexon pushed that exact 50/50 system onto KR and Global anyway for the 5th anniversary, forcing producer Kim Yong-ha onto a stage to insist it had nothing to do with revenue while handing out 200 free pulls and a Fest 3-star selector so Global wouldn't torch the Steam page a second time. Inven Global covered July's gacha rewrite, Blue Archive's Steam page still wears the backlash, Nexon's JP transfer portal and Inven Global's October report cover the Yostar split, and Gamemeca tracked November's 5th-anniversary rollout.

The First Descendant, meanwhile, survived Q3's project cull largely by staying small and cheap. Korean financial press noted that TFD maintained steady revenue from updates during the quarter, though parent Nexon still singled it out for declining year-on-year in Japan and quietly tucked it into third place inside a North America and Europe PC/Console total that collapsed from ¥40.5 billion (£191m / $253m) in Q1 and ¥21.5 billion (£102m / $135m) in Q2 down to just ¥9.7 billion (£46m / $61m) in Q3. With Steam concurrency hovering between 1,800 and 3,300 players and Season 4 Episode 2 shifted to a Tuesday just to dodge Grand Theft Auto VI, TFD isn't being shuttered tomorrow, though its lifespan now looks strictly tied to keeping maintenance cheaper than Söderlund's margin floor. SteamDB tracks TFD's player curve, and Director Minseok Joo's November Dev Note handles the pre-patch damage control.

Patrick Söderlund is selling Nexon its own pipeline as a visionary breakthrough:

Listening to Söderlund talk on Tuesday's call, you would think Nexon had never heard of sequels or legacy IPs before he took the Executive Chairman chair in February 2026. Since March, he has toured investor briefings pitching his "Four Expansion Vectors," the "MapleStory Playbook," and the "Classic Playbook," lecturing shareholders on how every player treasures a different nostalgic era and reassuring them that Nexon already has "relevant blueprints in place." Nexon's February appointment notice and the March Capital Markets Briefing lay out the whole pitch.

Of course those blueprints are already in place, because Nexon started building almost all of them years before Söderlund got promoted. Mabinogi Eternity, the Unreal Engine 5 rebuild, was unveiled in June 2023. Vindictus: Defying Fate was revealed in February 2024 and ran public alpha tests across 2024 and 2025. Mabinogi Mobile launched in Korea in March 2025. MapleStory Worlds soft-launched in the Americas in October 2024, and MapleStory: Idle RPG arrived in November 2025. Even Project OVERKILL and Dungeon&Fighter: ARAD were playable at G-Star back in November 2024. Söderlund didn't invent a nostalgia strategy for Nexon; he walked into a Korean publisher that has been repackaging MapleStory and Mabinogi for a decade, renamed the existing release calendar in Western MBA jargon, and presented himself as the architect.

His one big nostalgia pitch for Dungeon&Fighter, Dungeon&Fighter Classic slated for 2027, manages to look even sillier in China. China already has an entire underground ecosystem of 60-cap, 70-cap, and 86-cap private servers built specifically for veterans who quit official servers over predatory holiday bundle stacking. Better yet, Nexon and Tencent already played the "2009 60-cap nostalgia" card when DnF Mobile launched in May 2024, which Söderlund himself mocked back in March for lacking the retention mechanics to stop players from leaving once the nostalgia wore off. Rebooting 2009 DnF a second time while delaying an idle spin-off isn't a masterplan; it is hoping lightning strikes the same crater twice while slapping the same pay-to-win systems underneath.

Profit floors, quiet headcount shrinkage, and shameless contradictions:

Söderlund's management rhetoric ages about as gracefully as his product pitches. On the Q3 call, he proudly told analysts that while the rest of the industry is making cuts, Nexon "chose to keep our people," adding that Nexon does not grow by cutting staff. On the exact same call and shareholder letter, Nexon announced it had cancelled six more projects in Q3, part of an emergency restructuring drive that Korean business media report has shut down eight or more projects and businesses across 2026. That cull includes Nexon Games pulling the plug on Durango World (Project DX) on 8 September nearly four years after its 2022 G-Star reveal, cancelling Project EL after two years, and closing subsidiary studio Buttons entirely in April after sinking roughly ₩52 billion (£29.5m / $38.9m) into Project Everett. Meanwhile, Korea Economic Daily and PocketGamer.biz report that more than 300 employees are sitting on internal standby waiting for reassignment, CFO Uemura openly credited Q3 HR cost savings to headcount tracking below expectations while keeping annual labour spend frozen flat, and management already has deeper restructuring targets lined up for 2027. They don't need noisy Western-style layoff headlines when natural attrition, benched staff, and killing multi-year projects under a contribution-margin floor quietly do the job for them. Inven covered Durango World's September cancellation.

His AI talking points perform the same U-turn depending on who is listening. Back in March, after Embark caught flak from players for using AI voice acting in THE FINALS and ARC Raiders, he claimed very little of ARC Raiders' budget efficiency came from AI and framed it as reconfiguring old workflows instead. Eight months later, talking to Wall Street on the Q3 call and earnings letter, AI is suddenly turning designer sketches into pixel art "more than 10 times faster," cutting localisation timelines, and letting teams prototype "five versions of a mechanic in the time it used to take to build one," all plugged into Nexon's Mono Lake analytics platform to tweak design and monetisation. The moment Goldman Sachs pressed him on whether surging AI compute costs will actually pay for themselves, he slipped right back into PR mode, reminding everyone that players care about how their games are made. GamesIndustry.biz has Söderlund's March interview.

ARC Raiders has the exact "no staying power" problem he mocks other teams for:

Come back around to ARC Raiders, the entire justification for giving Söderlund control of Nexon in the first place. When I looked at its numbers in August, it had already bled down to around 26,000 to 29,000 average Steam players and 50% Mixed recent reviews. In September, before Frozen Trail arrived, it slid further to a 21,256 monthly average on Steam, dipping close to 10,000 concurrents at daily lows. Steam Charts tracks that monthly slide.

To manufacture October's recovery spike for Q3 slides, Embark threw everything in the cupboard at 8 October at once: the overdue Frozen Trail 2.0 patch, a four-day Free Play Weekend, a 50% discount slashing the game to $19.99, and a limited-time PvE-only toggle test from 13 to 20 October. Even with all four stacked together, ARC Raiders peaked at 185,443 Steam concurrents and only climbed to #9 on SteamDB's Weekly Global Top Sellers for 6–13 October, slipping to #17 the week after. The instant the free weekend and PvE mode ended, roughly 100,000 peak players vanished within days. As of 13 November, live Steam concurrency sits back around 30,000 to 40,000, and ARC Raiders has plunged 37 places on Steam's live Top Sellers chart down to #79, landing at #103 on SteamDB's Global Top Sellers, outside the Top 100 altogether. Even the recent review bump back to Mostly Positive came largely from PvE players showing up for a one-week PvE test that Embark yanked out again, forcing Söderlund on 4 November to backtrack and admit adding permanent PvE would not even be difficult. SteamDB shows the live player and Top Sellers numbers, and ARC Raiders' Steam store page shows the recent reviews.

The sales and outlook numbers make that bounce look even thinner once you separate the reporting quarters. Nexon's Q3 accounting ended on 30 September, before Frozen Trail arrived, which explains why total North America and Europe PC/Console revenue dropped from ¥21.5 billion (£102m / $135m) in Q2 down to ¥9.7 billion (£46m / $61m) in Q3. Because Nexon did not report until 10 November, management still used the Q3 shareholder letter to boast that lifetime ARC Raiders sales had "now surpassed 17 million units" after October's patch. Considering the game was already at 16 million back in May and 16.3 million in Q2, five months, a 2.0 overhaul, a free weekend, and a 50% discount still added under a million copies. Worse still, Q4 is the quarter that actually captures Frozen Trail and the half-price sale, yet Nexon guided Q4 ARC Raiders revenue down year-on-year and only "broadly flat" quarter-on-quarter because slashing the price cut revenue per unit. Even with a major update and a 50% promo inside Q4, Nexon does not expect ARC Raiders to beat the update-free summer slump it just posted in Q3. Meanwhile, Embark's other flagship, THE FINALS, posted another modest year-on-year revenue drop in Q3 and expects a flat Q4, which somehow did not stop Söderlund from boasting about a Hollywood film and TV deal with Chernin Entertainment for both games. Insider Gaming broke down the sales slowdown and THE FINALS dip, alongside Steam Charts for THE FINALS.

Best of all is how Embark's famous "lean team" myth collapsed right before the call. Back in spring, Söderlund bragged about shipping games with significantly fewer people at a fraction of traditional AAA cost, telling investors that Embark peaked around 110 to 120 developers and proved you didn't need big teams to run hit games. By late September, ARC Raiders Executive Producer Aleksander Grøndal publicly admitted to Game Developer that Embark's monthly update cadence left the staff "exhausted, unhappy, and constantly overshooting our capacity," forcing them to abandon monthly content for two major patches a year. Then on Tuesday's Q3 call, when UBS analyst Yijia Zhai flatly pointed out that ARC Raiders players returned for Frozen Trail and immediately started dropping again, Söderlund claimed "two high-quality, large-scale updates per year is the appropriate cadence" based on competitor data. Earlier on that exact same earnings call, CEO Junghun Lee blamed Dungeon&Fighter's collapse in China on "largely relying on one or two major updates per year." Inside one 45-minute briefing, shipping one or two big updates a year is the fatal flaw ruining Nexon's legacy Korean studios and simultaneously the masterstroke cadence for Patrick Söderlund's studio. GameSpot recorded Söderlund's April bragging, and The Game Business interviewed Embark in November.