XBOX’s restructuring is no longer cost-cutting; it is a fundamental redesign of how Microsoft makes games
Microsoft has announced another 268 Xbox job cuts, but the organisational changes are considerably more important. Activision will develop the next Halo and take responsibility for Rare and World’s Edge; Bethesda adds Obsidian; Playground and Turn 10 merge; King takes over Microsoft Casual Games; Undead Labs has been divested; and Ninja Theory faces potential closure after two sale attempts failed.
The pattern is clear: Microsoft is replacing a collection of relatively autonomous studios with fewer operating groups centred around publishers that already know how to run large franchises. Even Halo development moving to Activision would have sounded almost absurd a few years ago.
Why it matters professionally: This is portfolio management overtaking studio identity. Microsoft appears to be asking less “which studio owns this IP?” and more “which operating organisation is best equipped to commercialise and maintain it?”
Industry implication: If the structure works, other large publishers may follow with fewer standalone studios and more shared production, technology, marketing and live-ops organisations. The downside is equally obvious: distinctive studio cultures become harder to preserve.
dig deeper
Meta wants games to become something you encounter in a social feed rather than install
Meta unveiled Horizon Create and Horizon Studio. Create lets users describe a game through prompts on a phone; Studio provides more detailed browser-based editing. The resulting games can be published across Horizon, Facebook and Instagram.
The genuinely important part is distribution. Meta says an Instagram user could tap a game clip and enter its multiplayer session within seconds, without downloading an app or leaving the feed. Monetisation details are still forthcoming.
This creates an intriguing loop:
see game → play instantly → remix game → publish variation → distribute through social graph.
Why it matters: Roblox has spent years proving that creation, social discovery and play can form one ecosystem. Meta is attacking the same problem from the opposite direction: it already owns enormous social discovery surfaces and is now adding game creation.
What to watch: not whether AI can generate good games. That remains highly questionable. But whether instant social distribution produces a viable new discovery channel.
Casual watch: Merge-2 is having a monster year, and Gossip Harbor is now a genuine global heavyweight
One of this week’s most useful datasets concerns the game Gossip Harbor.
Sensor Tower placed the game second globally by mobile-game revenue in August, behind Honor of Kings and ahead of Whiteout Survival, Royal Match and Candy Crush Saga. Separate Attn Economy estimates put August IAP revenue at $111 million, with its user base up 64% YoY and ARPDAU up 17%.
More interestingly, the broader merge-match-2 category reportedly grew 84% YoY, with Gossip Harbor accounting for about 40% of the segment. That’s no longer a niche casual phenomenon.
Design lesson: Merge-2 has found an unusually effective compromise between low cognitive friction and long-term content depth. The board interaction is extremely understandable, while narrative, orders, events, collections and progression provide the retention machinery.
It also reinforces an important casual trend: the core mechanic does not need to become substantially more complicated for the metagame to become substantially deeper. For anyone working around match, merge, solitaire or mahjong, Gossip Harbor is currently one of the products I’d examine most closely.
dig deeper
YouTube Playables may quietly be becoming a meaningful web-games platform
This one has flown surprisingly far under the radar. YouTube Playables has expanded from seven original launch countries to more than 50 markets, including much of the EU and Nordics. Google also intends to test IAP during 2027.
More importantly, actual usage is becoming meaningful. YouTube reports that titles such as Build A Queen, Snake Clash and Stealth Master have exceeded 90 million plays, while Supersonic’s Tall Man Run has reached 106 million. Developers represented include SayGames, Supercent, Supersonic, Voodoo, TapNation and CrazyLabs.
Why it matters: Web gaming is acquiring distribution surfaces that traditional browser portals never possessed.
YouTube combines:
video discovery + creators + recommendation algorithm + instant gameplay + potentially IAP.
Industry implication: HTML5/web distribution deserves another look. Between YouTube Playables, browser instant-play initiatives and social platforms, “no-install gaming” could become a meaningful secondary distribution strategy rather than merely an old browser-games niche.
dig deeper
Supercell doing the unthinkable with Mo.co is a fascinating live-service experiment
Supercell has effectively unlaunched mo.co. The company says the game is not achieving the mass appeal it expected and has moved it back into beta. The existing game will remain online in low-maintenance mode, premium seasonal tracks and new weapons are stopping, while the development team experiments more aggressively with new versions.
AppMagic estimates put lifetime performance at roughly 9.2 million downloads and $7 million in IAP. The interesting thing here is not another struggling live service. It’s the product-management decision.
What to learn from it: live-service launches may not need to be irreversible transitions. For sufficiently well-funded studios, “return to beta” could become a useful framework for escaping the obligation to continuously feed an underperforming content treadmill.
dig deeper
Design watch: Dreapeline is a clever collision of roguelite optimisation and raising-sim attachment
A smaller release worth studying this week is Drapeline, which left Early Access on September 24. Its premise sounds eccentric: raise an endlessly hungry dragon girl over a year, deciding what she eats and how she trains, then see whether your build survives approaching calamities. But structurally it’s smart.
Runs last roughly one hour, providing the optimisation and build-crafting loop of a roguelite. The full release adds 25 endings determined by how the dragon was raised, turning repeated runs into narrative outcomes as well as optimisation exercises. Before 1.0 it had already sold more than 50,000 copies, accumulated 100,000+ wishlists and received 96% positive Steam reviews across 1,700+ reviews.
The loop becomes:
raise → specialise → fight → discover outcome → understand consequences → raise differently.
Why it’s interesting: roguelites normally motivate repetition through mechanical mastery and unlocks. Raising sims motivate repetition through emotional attachment and curiosity about alternative lives. Drapeline layers the two motivations together.
That’s a neat example of genre hybridisation where each genre solves the other’s weakness, rather than simply adding more mechanics.
dig deeper
Casual design watch: Tactile is turning Simon’s Cat into a puzzle container, not another single-mechanic game
Simon’s Cat World launched this week with an interesting structure.
Instead of building the product around one dominant puzzle mechanic, Tactile Games is combining sorting puzzles, Sudoku-style challenges and additional minigames, with more formats planned over time. Events and leaderboards provide the engagement layer around them.
That may sound modest, but it’s worth watching alongside NYT Games’ experiments discussed in previous briefs.
Casual publishers have historically built:
one mechanic → enormous level pipeline.
An alternative model is emerging:
one audience/IP → rotating library of familiar mechanics.
Why it matters: content diversification may reduce dependence on continually inventing thousands of variations of one puzzle. It also lets publishers test mechanics against an established audience before investing in standalone products.
For established casual brands, the “game as puzzle hub” model could become increasingly attractive.
dig deeper
App-store supply is exploding much faster than consumer attention
A striking discovery statistic this week: Google Play submissions rose from 46,136 in March 2025 to 111,047 in July 2026. Apple’s App Store rose from 46,889 to 117,648 over the same period, according to 42matters data cited by Google’s CEE apps lead. That’s roughly a 2.5× increase in little over a year.
The obvious suspect is lower production barriers—including AI-assisted development—but the important issue isn’t why people can make more apps. It’s what happens when supply expands dramatically while human attention does not.
Why it matters: prototype production is rapidly becoming cheaper; proving demand is becoming more expensive. In other words, technical execution becomes less valuable as a funding signal while demonstrated distribution becomes more valuable.
This is probably one of the clearest structural consequences of AI-assisted development we’re seeing already: making games gets easier; getting anyone to care gets harder.
dig deeper
Build A Rocket Boy entering administration is a warning about AAA-scale startup economics
Build A Rocket Boy, the studio founded by former Rockstar producer Leslie Benzies and developer of MindsEye, has appointed Opus Advisory Group as administrators following further layoffs. The studio had attracted major backing from investors including NetEase, RedBird Capital Partners and Galaxy Interactive.
Established AAA publishers already possess most of that infrastructure. Startups have to finance all of it before discovering whether their first major product works.
Industry implication: the failure strengthens the case for smaller initial productions, external technology, staged headcount growth and proof-of-market before scaling toward AAA.
The “raise hundreds of millions, hire hundreds of people, then discover product-market fit” model increasingly looks structurally fragile.
dig deeper
A deceptively interesting casual release: Solitaire 1380 treats a classic game as a catalogue
Solitaire 1380 isn’t going to dominate revenue charts, but the product idea is worth examining.
Rather than offering another heavily metagamed Klondike implementation, CodeFarmer has packaged more than 1,400 solitaire and patience variants, including regional games, Hanafuda and Mahjong. Players can access three favorites for free and unlock the entire collection through a one-time purchase.
It’s almost the inverse of modern F2P casual design.
Instead of:
one mechanic + infinite progression + live ops + recurring monetisation
it offers:
enormous mechanical breadth + player-directed discovery + premium unlock.
The interesting question is whether older casual audiences—particularly solitaire, mahjong and card players—have been underserved by an industry obsessed with metagames and event calendars.
There may be room for “reference-library casual games”: polished collections that treat traditional games as a hobby to explore rather than a retention funnel to optimise.
That is a small release, but a product direction I’d keep an eye on.
