Financial results season is upon us, and as is so often the case, it is a generous benefactor to anyone with an interest in seeing how the industry’s winds are blowing. It throws up numbers, narratives, and the occasional brutally expensive admission of failure, and asks only that we do the courtesy of joining the dots between them.
This particular cycle has been more generous than most; it’s not often that the headlines line up so perfectly to tell a coherent story.
Exhibit A: Sony, which in its FY2025 results booked a total impairment of $765 million against its 2022 acquisition of Bungie, and confirmed what the Steam concurrent-user chart has already shown – that Marathon, despite generally positive sentiment, is underperforming.
Exhibit B: Sega Sammy, which took a $200 million impairment against its 2023 purchase of Rovio, slid into a net loss for the year, and – in a single dry line on a slide titled “Review of the GaaS Strategy” – announced that it had cancelled its so-called “Super Game.” That was an initiative announced five years ago with talk of gigantic investment; it was never described in concrete terms, but was clearly an attempt to create something that could dethrone Fortnite.
And then there’s Exhibit C: Capcom, which behaves more or less like a traditional mid-2000s games publisher. It mostly makes single-player games with a beginning, a middle, and an end, occasionally reanimates a beloved franchise, sells things at full price, and gets on with making the next one. It’s not unaware of microtransactions and games-as-a-service (GaaS) and the like; they’re just not its core business and show no signs of ever being its core business. Capcom has just posted its ninth consecutive year of record profits.
I’m not the only one seeing how these stories line up; Sega’s results presentation was surprisingly candid about a significant pivot back towards an older way of doing business. Aside from the “Super Game” cancellation, free-to-play (F2P) titles generally are being deprioritised, with developers being moved off those titles and onto premium games built around the company’s major IPs. It’s not that there’s no money in F2P or GaaS titles – Sega’s own F2P revenues grew 14% year on year, so there’s still growth there. It’s just not enough to justify the resources, attention, and executive bandwidth that had been poured into the effort, to the detriment of the company’s actual core business.
Sony, unsurprisingly, did not say anything quite so blunt. I’m not sure it ever will, but actions speak louder than the careful phrasing of an earnings call, and a $765 million write-down of the studio that was meant to fuel their GaaS future speaks eloquently.
Sony’s attempt to pivot hard into GaaS titles has cost it far more than $765 million, or Bungie’s $3.6 billion price tag. While the PS5 is a success by many metrics, it’s notable that despite facing extremely diminished competition from its most direct rival in this generation, it has not meaningfully grown its install base versus its predecessor. That hasn’t been helped by pricing, for sure, but it’s hard to ignore how thin its first party game slate has been compared to the PS4 era, even if what has been released has mostly been of excellent quality.
The gaps in the release schedule mutely point to the enormous quantities of talent, time and money that were redirected into the GaaS push during the twilight of Jim Ryan’s tenure, most of which never produced a shipping product.
Sega’s U-turn and Sony’s quiet dropping of its once-shouted GaaS ambitions will not be the last signs we see of the pendulum swinging back. A retreat from these kinds of projects was probably inevitable from the moment the cheap-money era ended. Throwing enormous sums at very-high-risk, very-high-return swings, which GaaS titles ultimately are, is an appealing tactic only when interest rates are so low that investments with casino-like odds look appealing.
“The twisted high-risk economics of live service efforts both fuelled and masked a fundamental structural problem”
This is a markedly less attractive proposition when the cash in question is your own rather than someone else’s – as has increasingly been the case in recent years. The rush to GaaS, companies falling over one another in their haste to launch live service games, even knowing that so many of them would be dead on arrival, was always to some degree a low-interest-rate phenomenon dressed up as a product strategy.
The costs weren’t just financial. You could argue that a misguided product strategy has cost Sony vital opportunities to grow its market in this generation; across the industry overall, the GaaS detour did not merely waste money and developer-years, although it certainly did both. The twisted high-risk economics of live service efforts both fuelled and masked a fundamental structural problem: the relentless climb in the cost and timescales of making big games.
AAA budgets have drifted into Hollywood territory – often measured in the hundreds of millions of dollars – and they have done so without any corresponding expansion in the audience willing to pay $70 a copy for the result.
The incentive structure that creates is a mess to say the least. If a game costs $250 million to make and market, it’s almost certainly impossible to recoup it through ordinary sales alone. You need a hook, a hope, a prayer, a service layer, an in-game economy, and a moonshot at becoming the kind of cultural phenomenon that prints money for years. So you make it a GaaS game, and you cross your fingers, and most of the time the universe declines to oblige. The failure to control AAA costs made the wild gambles on live service look appealing precisely because it has totally undermined the economics of just selling a game to people who want to play it.
There’s a thread of thought in parts of the industry which says that AI tools of various kinds are the panacea we’ve been waiting for, finally promising the productivity boosts to finally curb these budgets. That might even work to some degree, for a while at least; although certainly not to the extent the technology’s most devoted evangelists claim. The logic of it – controlling budgets by cutting the cost of labour – is nothing new. A decade ago the magic answer was outsourcing to studios in Eastern Europe or South-East Asia. Now, the magic answer is outsourcing to a data centre in Oregon.
“It speaks to a lack of capability not at the coalface of development, but in management tasks”
The underlying instinct of those trends is identical. The pursuit of cost reduction by pruning labour costs in this way is an attempt to punt the buck away from where the problem more commonly lies – not in the workforce, nor with the technology in use, but at a management level.
An industry in which it is common for major projects to sit in development for four, five, six, or even seven years, hoovering up budget and headcount the entire time, is an industry whose issue is not insufficient throughput in its asset pipeline. It speaks to a lack of capability not at the coalface of development, but in management tasks like scoping and planning a project, controlling a budget, holding a timeline, or – crucially – prototyping extensively at a small scale and making good early decisions about whether a given creative idea is actually ready to be scaled up into full production.
Which brings us back to Capcom. A nine-year run of profitability surely means there must be something there for the rest of the industry to learn from. The lesson, I’d argue, isn’t actually about the kind of games it makes – it’s about how it makes them.
Capcom, for the most part (there are certainly exceptions), keeps tight control of its budgets and its timelines. It helps that Japan is a relatively cheap place to hire experienced development staff, of course, but effectively managing the efforts of those staff is also vital. Capcom has had many misses, but it’s rare for the company to see huge budget and timeline overruns, let alone late-stage cancellations, on projects that have come out of the creative incubation phase and into full production.
It helps that Capcom is relatively small, too; it just doesn’t have the scale to go around ramping up huge studios to work on vague and shifting concepts on the assumption that they’ll work it out as they go along. It iterates on the technology and tools so that successive games benefit from compounding investment rather than starting from scratch. To some degree, at least, it has internalised the idea that disciplined production is itself a competitive advantage.
None of this is glamorous. None of it makes for a punchy earnings-call slide about the next great paradigm shift. It is how you build nine straight profitable years, though.
The GaaS hangover the industry is now experiencing is real, and the strategic retreat under way is sensible. But nobody can lose sight of why the industry was so desperate to create GaaS hits to begin with; easy money made the high-stakes gambling attractive, certainly, but soaring budgets and stagnating audiences had also wrecked the economics of the traditional games business for many companies. That has to be fixed, and neither a clever monetisation strategy or a shiny AI pipeline will be the magic wand to fix it. The actual fix is the often thankless task of improving project management, adjusting company cultures to align incentives around scope and delivery, and taking on the behind the scenes grind required to actually finish a game on time and on budget.
