Certainly the
average loss of range due to severe cold weather isn't "over 40%". But the saying "Your mileage may vary", or YMMV, certainly does apply here, in spades!
DaleL already made the points I was going to; that pre-conditioning your car by leaving it plugged in overnight, and the car's computer set to warm the car up before the driver expects to drive the next morning, certainly does reduce the range loss. And there are a lot of variables, such as how long (in minutes or hours) the drive is, and how much the driver uses cabin heat. The longer the drive, the more energy is going to be used heating the cabin.
So it really depend on your driving pattern. I remember a report from some years back, where a Model S driver in Norway said he only lost 20% range on bitterly cold days.
Regarding this recent spate of reports we're seeing repeated all over the internet, of Model 3 drivers losing up to 40% of range, and some even more, up to nearly 50%: I wonder if these are just Tesla bashers spreading FUD, or if there really is something about how the Model 3 is engineered which makes it more susceptible to losing range in very cold weather.
Certainly there are some differences from previous Tesla models. The Model 3 does not have a separate battery heater; it is engineered to intentionally generate waste heat in the motor when needed (and of course
only when needed), and that heat is used to warm the battery pack. But I have no idea how much that does or does not impact range.