VR headsets and the research behind them have made it possible to accurately track the position of the device with very little processing cost from a single camera. Additional cameras are simply for occlusion and field of view at this point. A coupled depth sensor handles any needed resolution of positional conflicts between real world objects and digital objects.

That tech wasn't there for 3DS, even with 2 cameras it wasn't stable or accurate, and it was pretty low res and low angle, and it couldn't tell if the digital content should be occluded by real world objects at it's perceived depth. Plus there are actual AR games now already established, and the framework and proof of concept to easily on-board new ones. Also, by the time the Switch 2 is released, the mainstream AR headsets and content will be even more established. Currently the best AR content is either on expensive headsets or in limited form on iPhones. But in less than a month the Quest 3 is out, an accessible mixed reality headset. It will have been out for a while by the time the switch 2 would come out.

This is very much one of the directions gaming is going. It doesn't have to replace all gaming, and never will. But it will be a pretty big part of gaming. Especially once there is enough public trust behind it for people to play augmented reality games outdoors. Us nerds already do so, we know we won't walk out into traffic even if the headset malfunctions. But the hypothetical "everyman" is apparently worried that they might? That's just not how anything would work... but whatever. It's safer than playing a phone game and wearing headphones, since letting the real world in and paying attention to both is the whole point, sight and sound.