It is possible that people in relevant positions are clueless cowards or calculating psychopaths.
My initial impression leans towards the first option.
Taking measures needs courage. You have to take very disruptive steps in the environment where popular opinion still calls it not worse than flu and before that you have to be smarter than popular opinion.
Actually 21 million or 45 million does not make much difference. There will be still order of magnitude between influenza and covid-19 (CFR 0.2% vs 3.4%, and mind you, 3.4% is an early estimate and subject to change in the situation where new cases are still coming in and people are still dying, if you observed then CFR for covid-19 was around 2% when it was rapidly growing in China and did rise to current level when the new cases started to slow down; many people warned that it will be much higher actually but these people were hushed down; you have to use your own analytical mind to choose what is more closer to the truth).
Problem is that there exists people who try to convince us that we need a middleman between our networked devices and us - that we need to connect our devices not into our home network but with their servers.
The reality is that we do not need middlemen, the middlemen needs us such that they could steal the data, repackage it and sell it.
I will give one example that I have already given before.
Suppose you have an IoT device in your bedroom that uploads CO2 values (of course it may also do something useful too such us showing this data to you or regulating the ventilation).
Now the rate of CO2 production depends on bodily activities. More activity means more CO2 production. I suppose that you agree that having this information you more or less are capable to deduce when a couple has sex.
This means that we could calculate a metric how much sex is going on and relate it to specific user.
Suppose that we detect that rate of sex has gone down by 20 (a made up number dependant of our model). We can now sell this information.
This information is for example useful perhaps for psychologist who want to sell you counselling.
But it is even more interesting for divorce lawyers who now could prey on couples having period of difficulty in their sex lives.
Even more, this information could be made more valuable by some additional influence. For example feeding the user with articles related to marital happiness and sex life.
I leave the rest for you imagination.
I think that such devices are very usable, but the data should not leave our home.
I think it is an overestimate but not more than by 1 order of magnitude.