PUBG Mobile's store revenue halved in September for the second year running. Krafton's books and Tencent's release calendar say that's the wrong alarm. The one worth hearing comes from players, and it's about whether the matches work.
PUBG Mobile's App Store and Google Play revenue fell 50% in one month this September, from $109.9m to $55.2m, knocking it from third to tenth in AppMagic's global grossing chart (GameDevReports). Twelve months earlier the same thing happened on the same page of the calendar: $132m in August 2025, $77m in September, its lowest month since December 2024 (Pocket Tactics). Both times the read was identical. The veteran is fading, and Tencent's own Delta Force and Valorant Mobile are eating it.
The second half of that sentence is partly true. The first half keeps failing its follow-up test. After the 2025 slump, Krafton reported KRW 702.7bn of mobile revenue for Q1 2026 (Krafton), up 32% on the KRW 532.4bn it booked in Q1 2025 (Krafton Q1 2025 release).
So here is the claim, and it is narrower than "PUBG Mobile is fine." For this game, store-tracker revenue is a seasonal operating rhythm seen through a keyhole, which makes it a poor gauge of decline. The gauge that is moving the wrong way is match quality: crashes, frame drops, queue times, bot lobbies and cheaters. No revenue chart measures it, and it is the failure mode that actually ends long-lived shooters. Worse, the same content machine that produces the revenue spikes is a plausible cause of it.
1. The keyhole: what the trackers see
Start with what a "PUBG Mobile revenue" headline measures. AppMagic's estimates cover App Store and Google Play purchases only. They leave out China's Android stores and direct-to-consumer web shops (PocketGamer.biz). In China, where the game runs as Peacekeeper Elite, that leaves iOS carrying the whole story. PocketGamer.biz said China made up "the lion's share" of the autumn 2025 fall, and Niko Partners' Daniel Ahmad told the outlet that Apple devices account for about a quarter of mobile devices used in China, so a drastic iOS swing "may not be completely representative" of the game's performance (PocketGamer.biz).
Then the outlets print different things. PocketGamer.biz publishes AppMagic gross revenue. GameDevReports and Mobilegamer.biz publish AppMagic figures net of store commission (GameDevReports, Mobilegamer.biz). For August 2025, that is $187.3m in one headline (PocketGamer.biz) and $132m in another (Pocket Tactics). For February 2026, PocketGamer.biz printed $184.9m (PocketGamer.biz), while Mobilegamer.biz said "$129m+" in May (Mobilegamer.biz) and "$133m" in July (Mobilegamer.biz). PocketGamer.biz's own October 2025 estimate moved from $59.9m to $60.2m between reports (November 2025, February 2026).
None of that is anybody's mistake. Estimates get revised and gross differs from net by the platform cut. It does mean a month-on-month swing in a single headline is a weak basis for a structural call about a game whose largest market is mostly outside the panel.
2. The books: what Krafton reports
Krafton's mobile segment is the nearest thing to audited ground truth, with a large caveat up front: it is not PUBG Mobile alone. It bundles Battlegrounds Mobile India (BGMI), PUBG Mobile, and the technology-service fees Krafton earns from Tencent's Peacekeeper Elite (Mobilegamer.biz). It is still the only line that captures China Android.
Here is what it shows.
- FY2025: mobile revenue of KRW 1.7407tn, up 3% year on year (Krafton), or up 11% once you strip out PUBG IP incentive payments recognised from partners in 2024 (Q4 2025 call transcript).
- Q3 2025, the quarter that contained the September 2025 "collapse": mobile revenue of KRW 488.5bn, up 14.8% year on year (Krafton, release PDF).
- Q4 2025 was a real down quarter: KRW 292.2bn, down 19%. The CFO blamed "seasonal off-peak effects and traffic-driven content updates", said technology-service fees from partners had declined "due to traffic-oriented operations", and added that they "have already shown solid recovery in January" (transcript).
- Then the rebound: KRW 702.7bn in Q1 2026, then KRW 451bn in Q2, up 5.5% year on year and down 36% on the quarter, a drop the CFO called seasonal (Krafton, Q2 2026 call transcript). First-half mobile revenue reached KRW 1,153.7bn, up 20% (Mobilegamer.biz).
Look at the bars and the pattern is plain: Q4 is the trough and Q1 the peak, in both years shown. The trackers picked up the same rebound. PocketGamer.biz's gross estimate went from $76.1m in December 2025 to $147m in January (PocketGamer.biz), and Mobilegamer.biz put Q1 2026 net revenue at $353m, up 128% on the quarter, "driven by a major eighth anniversary update" (Mobilegamer.biz).
The payer base kept growing through the scare. Krafton said paying users rose 5% for PUBG Mobile and 27% for BGMI in 2025 (Krafton), and BGMI paying users were up 17% year on year in Q1 2026 (Krafton). On the Q4 call, CEO Kim Chang-han said Peacekeeper Elite's average DAU grew by double digits in 2025 versus 2024 even as revenue growth softened, according to the call's interpreted transcript (transcript).
The bear case deserves a fair hearing. DS Investment & Securities cut its Krafton target price from KRW 400,000 to KRW 370,000 in December, arguing that Delta Force and Valorant Mobile were hurting PUBG Mobile and that, even allowing for Q4 being an off-season, "the downward slope is not small compared to last year" (Chosunbiz). Delta Force took $492.5m on the App Store and Google Play in 2025 up to 21 December, and Tencent said Valorant Mobile reached 50m MAU in China (PocketGamer.biz). Ahmad said Delta Force was out-earning Peacekeeper Elite at that point (PocketGamer.biz). Competition is real. The argument here is that the monthly tracker line exaggerates its size and misreads its timing.
One gap is worth stating plainly. There is no reliable public DAU or MAU series for PUBG Mobile. Krafton offers directional statements; the MAU figures that circulate online carry no stated methodology, so they are not used here.
3. The calendar: a revenue metronome
PUBG Mobile has shipped nine major versions since May 2025, one roughly every eight to nine weeks. Version 3.8 arrived on 8 May 2025 with the Steampunk Frontier mode and Attack on Titan content (Liquipedia); 3.9 on 8 July with Transformers (Liquipedia); 4.0 on 4 September with Spooky Soiree (Liquipedia); 4.1 on 6 November with Frosty Funland (Liquipedia), followed by a Balenciaga collaboration from 14 November (Swipe Right / Mynewsdesk). In 2026: 4.2 on 7 January with Primewood Genesis (Liquipedia); 4.3 on 12 March for the eighth anniversary, with Jujutsu Kaisen, MrBeast and OneRepublic content (Liquipedia); 4.4 on 12 May with Hero's Crown (Liquipedia); 4.5 on 9 July with Naruto Shippuden and Ferrari (Games Press); and 4.6 on 8 September with Midnight Hunters plus One-Punch Man, Lincoln, Alan Walker and B.Duck tie-ins (Level Infinite).
Each two-month version is a monetisation curve built from a themed mode, premium skins and collab bundles, and Krafton's commentary tracks the spikes to specific items. In Q1 2025 it credited "high-end, high-value skins for cars" and IP collaborations, plus an April progressive outfit that set a daily sales record (Q1 2025 call transcript). In Q1 2026 it pointed to an Apollo Automobil collaboration that "drove demand among high-spending users" (Krafton). In Q2 2026, a new premium item set a single-day revenue record (Q2 2026 call transcript). Asked about dependence on PUBG in April, Kim said: "we do have some ups and downs across some seasonalities where there's some peaks and off seasons depending on the quarters" (Q1 2026 call transcript).
Line the two Septembers up against that calendar. Both slumps come right after a summer version built on a big licensed IP (Transformers in 2025; Naruto and Ferrari in 2026) hands over to an autumn version led by an original Halloween-season mode. Summer 2026 was unusually heavy. Version 4.5 carried what the publisher called PUBG Mobile's "biggest giveaway to date", over $600 of free rewards, plus an original anime short from Studio Pierrot (Games Press). A September comedown from that peak tells you about August more than about the game's trajectory.
Six seconds of the Studio Pierrot collab short. Source: PUBG MOBILE on X, 8 July 2026.
Watch:
The Official NARUTO x PUBG MOBILE Collab Anime (PUBG MOBILE) · Official Naruto Collaboration Gameplay Trailer (IGN)
V4.6 Midnight Hunters Themed Gameplay (PUBG MOBILE) · 8th Anniversary Version available now (PUBG MOBILE)
4. The lobby: where the erosion is
If wallets were the problem, it would show up in reviews. It doesn't. Glotier's automated read of 900 store reviews on 5 July 2026 (700 from the App Store, 200 from Google Play) found pay-to-win mentioned in about 1% of the 341 negative ones. The leading complaints were crashes, freezing or lag (11%), matches that feel rigged or unfair (8%), hackers (7%), account theft or login loss (6%) and regional matchmaking (4%) (Glotier). The sample is small and the classification is automated, so treat it as directional. The same page lists ratings of 4.32 on the App Store from 1.6m ratings and 4.33 on Google Play from 48.5m.
The version 4.6 cycle adds anecdotes that rhyme with that list. These are individual Reddit threads with modest engagement, a signal rather than a survey:
- Auto-glide. Version 4.6 added a parachute auto-glide that locks forward input when you push the stick up (PUBG MOBILE Japan patch notes). Players in the version 4.6 Classic Mode feedback thread and in separatethreads want a way to switch it off.
- Performance. A Snapdragon 8s Gen 4 owner reported that since async rendering arrived, frame rate drops "down to 40 in one go" when jumping from the plane and "down to 30" in fights on a phone that used to hold 110, with delays on scoping and weapon swaps (Reddit, archived snapshot). The only reply in the snapshot was a link to the bug-report form.
- Queues. Late-September posts describe TPP squad matchmaking in Europe taking up to two minutes at Diamond (Reddit, Reddit).
- Bots. Returning players describe long stretches of bot lobbies before real matchmaking kicks in (Reddit). Filling lobbies is a deliberate design tool: version 3.8 gave World of Wonder creators a toggle "to fill a match with fake players" when not enough humans show up (Liquipedia).
Now the less obvious part. On the Q2 2026 call, the CFO said that in Korea and Japan, where Krafton runs the game directly, UGC modes account for a mid-teens share of total time spent and classic battle royale for about 20% (Q2 2026 call transcript, interpreted). World of Wonder alone has passed 5.4m user-created maps and 33bn matches played (same call). If those shares are even roughly representative, the content that drives the revenue spikes (themed modes, World of Wonder, Metro Royale, Arena, Home) also splits players across a growing number of queues. Thin queues produce longer waits, wider skill bands and more bot fill, which is the complaint list above. Public data can't prove the causation. The mechanism is straightforward, though, and it fits what players report.
Tencent appears to know where the pressure sits. Version 4.4 extended 120fps support to 26% more devices and added Flash Crew, a system that recommends compatible teammates (Liquipedia). In August it announced PUBG Mobile Light, a second client for the same game with an initial download of about 1GB on Android and under 2GB on iOS, shared accounts and inventories, and crossplay with the main app (Mobilegamer.biz, Dot Esports). Dot Esports notes the standard iOS package is already 4GB before in-game downloads. Krafton India is preparing BGMI Lite for the end of 2026. These are infrastructure bets rather than content bets, and they are the most interesting thing the game has announced this year.
From the PUBG MOBILE LIGHT announcement video. Source: PUBG MOBILE on X, 16 August 2026.
5. What would prove this wrong
The thesis is testable within months.
- Krafton's Q3 2026 results, due in early November (Q3 2025 results came on 4 November). If mobile revenue falls by double digits year on year despite a Naruto and Ferrari summer, the September slump is more than a calendar effect.
- Q1 2027. If the Lunar New Year and ninth-anniversary window fails to produce a rebound like Q1 2026's, the seasonal reading is wrong.
- The complaint mix after Light ships. If crashes, queues and unfair matches keep gaining share of negative reviews once a lighter client is available, the lobby problem is outrunning the fixes.
The industry keeps asking whether PUBG Mobile's wallet is shrinking. Krafton's numbers say not yet, with the caveat that its segment mixes three products. The better question is whether someone who opens the app in November on a mid-range Android phone gets a smooth, fair match against humans in under a minute. Nobody publishes that number, and it will decide the next eight years of this game more than any September chart.
Follow BitTopup for more PUBG Mobile UC news~
