Isn't that more a Windows problem? An OS should never crash entirely because a device was plugged in. Fitbit probably should've caught that in testing though.
Device drivers run in kernel mode and can bugcheck the machine. Microsoft can certainly provide tooling and help device driver writers write reliable drivers, but the device is still responsible for it's driver.
My understanding is that Windows supports user-mode drivers and Microsoft strongly encourages not writing kernel-mode drivers unless absolutely necessary. If Fitbit did that anyway though then it's definitely their fault.
Yeah exactly I've had Linux drivers cause kernel panics a number of times.
It also could be a defect in the host computer's USB controller. Sometimes those issues have to be fixed in firmware; the OS isn't necessarily to blame.