This is indeed possible. The US spent a fair amount of money in the 60s and 70s on beamed power. In recent years it has gotten some more interest for lunar rovers that can survive the 14 day lunar night (without RTGs). The losses are not negligible, and the pointing requirements are rather extreme but in practice there's no reason why this couldn't happen.
Losses: Initially the losses are driven by the solar panels on the satellite, if we assume they are 25% efficient then you've already lost a big lump of your 1300W/m2 (at 1 AU). Then you have the internal storage and transfer loses, maybe a few more percent ~2%. Then the beaming losses, creating the beam might be as inefficient as 10% or 20%, lets be optimistic and say 20%. There are potential losses based on the footprint size of the beam, but we'll (very optimistically) assume we can point just right. The receiver at the lander might manage another 25% efficiency, plus a few extra % for inboard storage etc ~ 2%. Total power efficiency: 0.25*0.97*0.20*0.25*0.98 = 0.012 = 1.2%.
So about 1% of the power you get in orbit will make it to the lander. Of course all of these efficiencies can be improved upon, but it gives you an idea of the sort of problems. For 100W on the surface you would need 100W/0.012/1300W/m2 = 6.4m2 of solar panels in orbit.
Once I get on my laptop (and off my mobile) I'll pepper this with links and correct typos.