Loss prevention using this tech is tough, especially since it's trivial to defeat UHF RFID tags' detectability -- on purpose or by accident, by placing the tags against your body (fundamentally different from EAS). Plus, the things you're most-keen to track (high-value items) are also the most likely to contain metal; metal mucks w/ RF propagation and energy harvesting, and therefore requires special (more-expensive) tags.

This idea dates back to the early days of UHF RFID... it didn't work back then, and not much has changed in the technical fundamentals that would make me think it has suddenly become possible. There are some new uses cases in retail... but they revolve around new (non-checkout) business plans.

I would LOVE nothing more than to be proven wrong... so I hope they find a way to make it work!

(I happen to be an expert in this field... happy to talk more, as it's still a major interest area.)

I'm not sure if it will save stores money or not. A lot of stores don't want it to be too easy to check out and leave. Getting to the cashier and waiting in line increases the chances that you'll make another purchase. Stores focused more on high volume convenience-driven purchasesprobably think about things differently though, and maybe they'd be more interested in something like this.

Maybe the bigger possibility is that it allows stores to present and sell their goods in afar more decentralized way that gets them closer to where consumers are. If on-demand workers, or drones, could restock inventory, and shoppers just needed an app to make purchases, then shopping malls could disolve into the surrounding cityscape.