The Big-Budget Failure Canon

Every big-budget failure gets flattened into the same headline the week it happens — over budget, under-loved, shut down fast — and the flattening is the problem, because it teaches nothing. A studio spending years and tens of millions on a game that doesn’t land is rarely failing for one reason, and the reasons are almost never the reason the discourse settled on by the time the servers went dark. This is a list of five that are worth returning to specifically because each one collapsed for a genuinely different structural cause, and together they make a better argument about how AAA development goes wrong than any single postmortem could.

Concord — the market that had already been answered

Firewalk Studios’ hero shooter launched in August 2024 and was pulled from sale within two weeks, its servers shut down permanently a month later — one of the fastest deaths of a full-price, years-in-development title in the genre’s history. The failure here wasn’t a broken game; by most contemporary accounts the shooting felt competent. The failure was a market that had already been thoroughly answered by Overwatch years earlier and reanswered by half a dozen free-to-play competitors since, entering as a $40 premium title into a genre where the audience had already settled on its free incumbents and had no unmet need the new entrant was solving. The full account of the two-week collapse is the cleanest case study in this list of a game failing on positioning before a single player judged its mechanics.

Redfall — the genre mismatch inside one studio

Arkane Austin’s Redfall (2023) is the harder case, because Arkane’s design pedigree — Dishonored, Prey — made the studio’s name synonymous with exactly the kind of single-player immersive sim craft that Redfall was supposed to bring to a co-op vampire shooter. What shipped struggled with enemy AI, technical performance and a structure that split the difference between immersive-sim systems depth and live-service co-op looter design without fully committing to either one. The full read on Arkane’s worst day at the office makes the case that the failure wasn’t a talent problem — the studio’s pedigree was real — but a brief mismatch: a team whose entire design language is built around solitary, systemic problem-solving was asked to build a genre that depends on the opposite instincts.

Suicide Squad: Kill the Justice League — the live-service tax on a story studio

Rocksteady built its reputation on the Batman: Arkham series — tightly authored, single-player, story-driven combat games with no ongoing service layer at all. Suicide Squad: Kill the Justice League (2024) asked that same studio to deliver a always-online, seasonal, loot-driven live-service shooter, a genre with entirely different pacing, monetisation and content-cadence requirements than anything Rocksteady had built before. The full account of how live service ate Rocksteady traces a studio whose core competency was authored narrative spending its resources instead on systems built to sustain player attention indefinitely — a completely different design discipline that its prior body of work had never asked it to practise.

Skull and Bones — the decade that ate its own premise

Ubisoft Singapore’s naval combat game holds the unfortunate record for development time on this list: originally announced in 2017 as a spinoff of Assassin’s Creed IV: Black Flag’s ship combat, Skull and Bones didn’t ship until 2024, seven years and multiple full reboots later. The full account of the eleven-year gestation covers how a game can lose its own reason for existing across a development cycle long enough for the genre around it, the technology underneath it and the creative leadership steering it to all change multiple times before launch — by the time it shipped, the pitch that justified greenlighting it in 2017 had been answered by other games several times over.

Anthem — the flight model with nothing built around it

BioWare’s 2019 looter-shooter is the canonical case of a single genuinely excellent system shipping inside a game that didn’t build enough around it to justify the whole. Anthem’s Javelin flight — freeform, physical, genuinely satisfying traversal through open skies — was, by wide critical agreement, one of the better feeling movement systems of its console generation. The full read on why the flight model deserved a better home covers the actual mechanics of that traversal in detail; the failure sat almost entirely in the loot, mission structure and endgame built around the flying, which never gave players enough reason to keep flying once the novelty wore off.

The tell that shows up before launch, if you know where to look

Each of these five had a visible warning sign well before the release-day numbers confirmed the problem, and the signs were different enough from each other that no single pre-launch checklist would have caught all five. For Concord, the tell was a closed beta drawing modest numbers against a genre already saturated with free alternatives — attendance data that a healthier positioning strategy would have treated as a stop signal rather than a scheduling detail. For Redfall, the tell was press previews describing combat and AI behaviour that didn’t match the tension Arkane’s prior games were known for, months before release, in previews that read as polite rather than enthusiastic. For Suicide Squad, the tell was as far back as the initial reveal trailer’s mixed reception, years before launch, when a character-action fanbase built on the Arkham games first saw the shift to a loot-shooter structure and reacted with open scepticism rather than excitement. For Skull and Bones, the tell was the reboot count itself — multiple public restructurings of a single game’s design over seven years is never a healthy production signal, no matter how it gets framed in each new re-reveal. For Anthem, the tell was BioWare’s own internal culture, later reported publicly, of building the flight tech late and iterating on the endgame loot loop even later still, leaving a gap between the game’s best system and everything meant to sustain it that no marketing push could close in the run-up to launch.

None of these signs guaranteed the eventual outcome on their own — plenty of games survive a rocky beta or a mixed reveal trailer and go on to succeed. But in each case here, the sign was visible, and in each case the studio or publisher proceeded on the original timeline anyway, which is itself a pattern worth naming: an organisation large enough to have spent this much is also often large enough that the sunk cost of reversing course looks more frightening on a spreadsheet than the cost of shipping into a market that’s already telling you no.

The pattern underneath the headlines

Line these five up and the differences matter more than the similarities. Concord failed on market positioning before launch. Redfall failed on genre mismatch inside a studio with real talent. Suicide Squad failed on asking a narrative studio to become a live-service one overnight. Skull and Bones failed on a premise that expired during an unusually long development cycle. Anthem failed on building too little around one genuinely good system. None of those five diagnoses transfer cleanly to any of the other four, which is exactly the point: “it was a live-service game and those always fail” or “the studio wasn’t good enough” explains none of these cases individually and both explanations are demonstrably wrong for at least two of the five, whose underlying studios had shipped genuine successes both before and after.

The survivors that make the same bets

It’s worth being honest that identical structural bets have paid off elsewhere, which is the strongest evidence that none of these five failed simply because the genre or the ambition was inherently doomed. Live-service hero shooters aren’t a dead category — plenty of successful ones exist alongside Concord’s failure, which only sharpens the point that Concord’s specific problem was market timing and differentiation rather than the format itself. Studios pivoting genres mid-career have also succeeded before, which means Redfall’s difficulty wasn’t simply “an immersive-sim studio can never make a shooter” but something more specific to that project’s execution and brief. The uncomfortable truth a failure canon has to sit with is that there’s rarely a rule general enough to have predicted all five outcomes in advance without also incorrectly predicting failure for games that went on to succeed under near-identical conditions. That’s precisely why the specific, individual diagnosis matters more than a general theory of “live service is bad” or “genre pivots don’t work” — both of those theories are falsified by counter-examples the moment you look for them.

What the list is for

A failure canon isn’t assembled to mock the people who worked on these games — the actual cost of each collapse landed hardest on developers who had no say in the market positioning, genre brief or greenlight timeline that ultimately sank the project, and the wider studio-closure wave carries its own separate accounting that this list isn’t trying to relitigate. It’s assembled because each of these five names a specific, avoidable structural mistake clearly enough that the next studio staring down a similar brief has an actual precedent to study rather than a vague cultural memory of “that one failed.” That’s worth more than a highlight reel of bad numbers, and it’s the only honest reason to keep this list around once the news cycle around each individual collapse has long since moved on.