Wasn't google tango their solution? (but requires custom android hardware for depth sensor)

https://get.google.com/tango/

I haven't heard anything about them for a few years now.

Visual-inertial SLAM has been working for a few years on mobile phones: https://www.youtube.com/watch?v=e7bjsIqlbS0 So there should be plenty of third-party libraries you can use.

The beauty of it is that ARKit also requires somewhat specialized hardware. But since apple controls the hardware as well, they can pull it off much faster than the android eco-system ever will.

They solve the chicken and egg problem, having phones with the needed hardware before the apps/demand even exists.

Apple already has the first mover advantage on this. Will be interesting to see how Google/Android will overcome it especially given they don't entirely control the hardware. They should at the very least add the necessary hardware in the next Pixel phone.