This is a good summary of why it is so hard to make a wrist-watch device to monitor glucose (without the little needles of the current commercial, accurate, continuous monitoring devices.) I remember people talking about optical and IR detection in the 1990's, and it was an old problem even then. Non-invasive glucose monitoring is a great example of a problem that can be explained to nearly anyone (so relatively simple), but a practical solution is very hard. Maybe mid-ir (8-12 um) technology will improve to the point that it makes sense in a consumer device. That will open up a lot of other possibilities in chemical detection, too.

Not mentioned is https://en.wikipedia.org/wiki/Google_Contact_Lens, another technology failure that sounded great at first (the glucose level of tears is unfortunately poorly correlated with blood.)

I cannot begin to tell you all how much of a difference it would make to people with either type of diabetes to have a non-disposable device to monitor blood glucose. CGMs are great, I use a Libre, but it's $$$ every other week, and sometimes the damn things just break when you're putting them on.

I thought I read that there are actually other things going on with blood sugar, not just glucose, but maybe fructose? Wonder if this can be measured.