a Graveyard
Now
11 of 19 major live-service games released in 2025 lost more than 90% of their players within months. Concord shut down in eleven days. A CEO resigned over one launch. The genre publishers keep betting the company on has become, by its own 2025 numbers, the least reliable business model in modern gaming.
2025 Was the Year the Numbers Stopped Lying
As of January 2026, the picture is unambiguous. Researchers tracking 19 of 2025’s biggest live-service and live-service-adjacent titles found that 11 of them lost more than 90% of their all-time player count on Steam, according to SteamDB data. This is not a story about one or two high-profile disasters. It is a structural pattern across an entire release year, spanning multiplayer shooters, cooperative action games, and online RPGs.
Supervive stepped out of early access in 2025 and was cancelled less than five months later, having lost 99% of its players. FragPunk retained just 2% of its launch player count. Mecha Break, InZoi, Skate, Killing Floor 3, and Blue Protocol: Star Resonance all retained fewer than 5% of their players. Four months after Remedy’s live-service launch underperformed, the company’s CEO resigned. This is the toll the model is currently taking — not in projected risk, but in confirmed, measured outcomes.
“You only have to look at some of 2025’s best-performing games — Clair Obscur: Expedition 33, Kingdom Come: Deliverance 2, Schedule 1 — to see that a large percentage of audiences don’t want this type of game. Yet developers persist, chasing a dream that comes from out-of-touch executives who see Fortnite’s success and assume the formula transfers.”
The Gamer, January 2026 live-service player analysisWhy the Same Mistakes Keep Repeating
Live-service games demand perpetual profitability or they die. Unlike a traditional game you can return to anytime, an abandoned live-service title becomes an unplayable void the moment support stops. This single structural fact — this is a service, not a product — is the root of nearly every failure pattern documented across the genre. Daily logins, battle passes, seasonal cosmetics, and constant FOMO mechanics are not bonus features layered onto a good game. For the live-service model to be commercially viable, they are the business model itself.
The pattern of failure repeats with remarkable consistency: poor launch quality with bugs and missing features killed momentum for Anthem, Redfall, and The Day Before. Market saturation meant LawBreakers, Concord, and Rumbleverse simply could not compete in crowded genres already dominated by entrenched leaders. Aggressive monetisation alienated players in Chocobo GP, Evolve, and Marvel’s Avengers. And studios built around single-player design — forced into live-service mandates by executives chasing recurring revenue — consistently struggled, as seen with Redfall and Marvel’s Avengers.
The Sony case study: In 2022, Sony announced plans to release a dozen live-service games by 2025. By 2024 that number had shrunk to six. The only unambiguous success has been Helldivers 2 — which nearly failed anyway after Sony attempted to force a PSN account requirement on PC players. The Last of Us Online and a God of War live-service title were both cancelled mid-development. Concord launched to a peak of under 1,000 concurrent players and shut down within two weeks, having cost a reported $200 million to develop. Sony has since enacted what it calls “structural reform” across its live-service ambitions — corporate language for admitting the strategy mostly did not work.
What 2025 Actually Did to the Genre
| Game | Studio Background | What Happened | The Lesson |
|---|---|---|---|
| Concord | Firewalk Studios, owned by Sony, single-player pedigree | Launched at $40, peaked under 1,000 players, shut down in 11 days | Premium price + generic hero shooter = death |
| Supervive | Theorycraft Games, MOBA-battle royale hybrid | Exited early access 2025, lost 99% of players within 5 months, cancelled | Early access hype does not equal launch demand |
| FragPunk | NetEase, competitive FPS with card mechanics | Retained just 2% of launch player count | Novelty mechanics wear off fast without depth |
| FBC: Firebreak | Remedy Entertainment — known for single-player narrative games | Underperformed badly; CEO resigned 4 months post-launch | Forcing single-player studios into GaaS rarely works |
| Marvel Rivals / Apex / Fortnite | The exceptions executives point to when greenlighting the next attempt | Genuine sustained success — among only 5-10 live-service titles globally that work | Survivorship bias drives the entire industry strategy |
Who Keeps Greenlighting This?
The uncomfortable truth buried in the 2025 data is not that live-service games are hard to make. It is that the executives approving these projects have access to exactly the same failure data the rest of us do, and keep approving them anyway. Publishers prioritise monetisation over experience, engineering games around spending rather than fun — battle passes locking cosmetics behind paywalls, events disappearing if missed, artificial scarcity and player anxiety baked into the core design. These are not accidents. They are deliberate choices made with full knowledge of the genre’s failure rate.
The financial reality is stark: the live-service model works for maybe 5 to 10 titles globally. For everyone else attempting it, the outcome is closer to financial ruin. Publishers close underperforming live-service games that fail to meet profitability targets within 12 to 24 months, which makes the entire business model structurally unsustainable for anything except a handful of blockbuster outliers. And yet the announcements keep coming. Marathon and Highguard both launched into 2026 carrying significant skepticism, with Highguard not even cracking Steam’s top 100 most-wishlisted games before release.
A single successful live-service title can generate revenue for a decade, far exceeding what a traditional premium release achieves. Fortnite, Apex Legends, and Marvel Rivals prove the model can work at extraordinary scale. The potential upside is large enough that the high failure rate is treated as an acceptable cost of pursuing the few outsized winners.
The 80-90% failure rate is not a calculated risk borne by executives. It is borne by the developers who get laid off when the project inevitably underperforms, and by the players who lose access to a game they invested time and money in when servers shut down. The people making the greenlight decisions are rarely the people who pay for the decision going wrong.
The Graveyard Keeps Growing. Nobody in Charge Seems to Notice.
Live-service games are not inherently broken as a concept. Helldivers 2, Fortnite, and Apex Legends demonstrate that the model can produce genuinely great, sustained experiences when the game has a clear identity, generous content cadence, and design built around player enjoyment rather than retention metrics. The problem is not the format. The problem is an industry that keeps approving new entries into an oversaturated, structurally hostile market on the strength of survivorship bias, while ignoring its own published failure data.
2025’s numbers should have ended this conversation. 11 of 19 major launches losing over 90% of their players is not an anomaly requiring further study. It is a recurring industry pattern that the people responsible for greenlighting these projects continue to treat as bad luck rather than predictable outcome. As 2026 brings Marathon and Highguard into an already crowded graveyard, the only honest question left is not whether live-service games can work. It is why an industry with this much evidence keeps betting against its own data.
The solution was named clearly by the people studying this collapse: the question is not whether to add more seasons or cosmetics to a failing live-service title. It is whether every game needs to be a live service in the first place. For most of them, demonstrably, the answer is no.
0 Comments