People tend to forget that both devices (Vive wands and Oculus Touch) contain high speed IMUs. The optical part of the tracking systems, alone, is not fast and fine-grain enough to track smoothly.
Both also have haptic "buzzers", which should draw quite a bit of juice.
(For that matter; A LED, (which emits light just as the designation says. :P), should use a good bit more power than a photo diode, that just detects it. Those diodes, on lighthouse-tracked devices, are connected to circuitry that makes their effects more distinct, and so on, of course, but that should be negligible, and I doubt there is a great deal more processing done on a Vive wand, at the moment, than in a Touch controller; If nothing else, the latter needs, in addition to piping a steady flow of IMU data (just like the wand), to modulate the LEDs, for identification, so it's not an entirely passive thing.)
I have no idea what it is that sucks so much juice in the Vive wands, but overall I can't help but think that contrary to what use shows, they should bloody well not be significantly thirstier than their Touch counterparts. (...nor do I know how small the accumulators are, nor how large the actuators... )
Reminds me of how my old comfy Logitech MX1100 wireless mouse would last months on an AA battery, but the newer and less ergonomic MX Master that replaced it, when I accidently dropped its predecessor into a cup of tea (sigh - yeah, yeah, go on, laugh), demands weekly recharges.