One week at summer's end compressed a whole season of signals into a single industry column: image-generation models went mainstream under the nickname "Nano Banana", Adobe quietly retired Aero, its AR authoring tool, and XR studios kept consolidating.
The Nano Banana moment — Google's image model earning a consumer meme name — matters to XR more than it appears. The most requested XR experiences have always been generative: worlds, props, characters, interfaces. If creation collapses to a prompt, the bottleneck of spatial content stops being 3D artists.
Adobe retiring Aero says the same thing from the other direction: dedicated AR-authoring pipelines never found their audience, in part because the audience for AR remained smaller than the tools built for it.
The layoffs were the mundane part of the cycle: a market segment maturing from its investment phase, with studios consolidating around the products that actually monetize.
Read together, the week sketched the industry's next composition: fewer bespoke authoring tools, more generative content; fewer generalist XR studios, more focused teams shipping on the platforms that survived. The tools to build a world are becoming ordinary software — and the interesting question moves from how to build it to what is worth building.
The week also landed near the end of a decisive summer for XR strategy: with Meta, Samsung and Google's roadmaps all public, the fall's remaining question was less about whether spatial AI products would ship — and entirely about who would build the experiences that made them indispensable.