You place a virtual chair in your living room using an AR app. You close the app, do something else, then reopen it. The chair should still be exactly where you left it.
That persistence creates a problem. Each AR platform uses different methods to map physical space and remember locations.
What Spatial Anchors Do
A spatial anchor is a bookmark in physical space. When you place a virtual object, the system scans nearby physical features like walls, furniture edges, and floor patterns. It creates a unique fingerprint of that location.
When you return, the system scans again and matches the current view to its stored fingerprint. It repositions the virtual object based on recognized physical features.
Cross-Platform Anchor Systems
ARCore on Android, ARKit on iOS, and Azure Spatial Anchors on HoloLens all work differently. They scan at different resolutions, store different amounts of data, and have different accuracy levels.
Cross-platform AR development requires using an anchor abstraction system. You create an anchor through a unified interface. The underlying platform handles the actual scanning and storage in its native format.
Shared Experiences
When multiple people need to see the same virtual objects in the same physical location, spatial anchors become critical.
Ingrid Thorsen built a multiplayer AR puzzle game where players on different devices worked together in the same room. She used Azure Spatial Anchors to create cloud-stored reference points. When a player on an iPad placed a puzzle piece, a player on HoloLens saw it in the exact same spot because both devices referenced the same spatial anchor data.
Understanding spatial anchors means knowing how AR platforms remember physical locations and how to make that memory work across different devices.