This is a well understood problem. Method claims that require actions taking place in two or more locations are a problem in enforcement for two reasons - one is extraterritorially - the topic of the question. All actions need to be performed in the juristiction of a specific patent. The other is that all actions must be performed by a single infringing entity or at least controlled and commanded by a single entity. If actions are split they might be done by two different entities - then nobody infringes even if all action performed in the U.S.
Lets look at Claim 1 of the application. As a method or process claim, its required elements are not things, the elements are actions. The question is who performs the actions and where they are performed. To infringe, one entity needs to perform all positively recited actions.
in order to infringe that patent. The several actions claim 1 requires are "encoding; printing; capturing; storing; capturing; authenticating; selling; capturing; purchasing. Other than "storing said unique product identifier . . . in a cloud based data store", the actions all occur in the place the benefit is provided. The storing action kills the claim's attempt to have all actions take place in one local locality. A better wording would have been "sending said unique product identifier to be stored on a cloud-based data store". Then the required action is "sending" and the storing brought in inferentially - not a positivity recited required step. It can be a little difficult to see the distinction.
Imagine a system with a transmitter and a receiver. Some claims should be all about the transmitting and others all about the receiving.