Unfortunately, at least for a lot of manufacturers, the decision process is going to go something like, "Should we make sure this can operate even if that domain goes offline?" "Nah, if that happens, the company will be belly-up, so why should we care?"

(And yes, that is assuming that they'll never accidentally—or intentionally—fail to renew their domain name...)

Because people keep falling for it, so manufacturers do not build independence features.

IMO, it's because most manufacturers' main goal with "smart" devices isn't making better products; it's designing products that are less expensive to manufacture, but bring in at least as much revenue.

A device that's controlled using the owner's phone (or a web application, etc.) is cheaper to produce in bulk than one that has a built-in interface to interact with.

A security camera system that stores all its data in cloud-hosted storage is also cheaper to produce in bulk, because it doesn't need to include storage hardware of its own. That has the added benefit (to the manufacturer) of tying the customer into ongoing service fees.

In other words, while I agree with you that they're bad design decisions from the consumer perspective, they're good design decisions (at least in the short term, which is all most corporations care about now) from the manufacturer's perspective.

For many of these products it was a very deliberate design decision, both in order to lock-in the user and also to be able to collect the data on them.