The idea is old and the obstacle has always been the same. Running a game remotely and sending back video is straightforward. Doing it with little enough delay that a player cannot feel the gap between pressing a button and seeing the result is not. Several services attempted it before household connections were widely good enough, and their difficulties had less to do with server capacity than with the physical distance between a player and the nearest data center. Latency is a property of geography as much as of engineering.
Google's Stadia, launched in 2019, was the most visible attempt to sell streaming as a product in its own right, running games on remote servers and delivering them to browsers and phones. The technology largely worked and the business did not. A thin library, an arrangement in which players bought games outright in order to stream them, and public doubt about how long the service would last all kept adoption low. It closed a little over three years later with purchases refunded, which is not how most shutdowns go.
Streaming raises a question the rest of the archive keeps circling: what it means to own a game. A closed service takes its catalog with it, and that closure became the standard illustration of the risk even though buyers got their money back. The approach that lasted treats streaming as one feature inside a broader subscription rather than a separate product. Entries here cover the services, the technical milestones, the closures, and the moment when cloud catalogs became a formal question in competition law.