This article has enough completely-wrong aspects that exempts it from the concept of "not even wrong" I suppose.
1) The watch does not engage in covert traffic. It's the pairing app for the watch that a user installs on a phone that does the communication.
2) What on earth does the redundant phrase "covert communications behind the users' back" even mean? Have you looked at network traffic when *any* application has been launched? If you think that any app talking on the internet without explicitly asking the user first counts as "covert communications", then I think you can label just about all of the software out there (esp. in the mobile space) as engaging in "covert communications."
3) The phrase "random IP address" used by the speaker is slang meant to convey that he didn't know what it is. In this case, it's a system referred to by its IP rather than its DNS name. So rather than looking up who owns the IP address, he says it's "random" and shrugs.
4) To give up and say that it's "very difficult to determine" what is being sent over the network because it's over an encrypted channel is ridiculous. For all we know, it's just talking to the software vendor via HTTPS. In which case it would be trivial to inspect by using MITM.
I'm not saying that there's nothing sketchy going on here. But to provide zero evidence of what's actually happening and just speculate and spread FUD is irresponsible.