Why does it have to be Boston Dynamic? There are lots of robot dogs out there now. I also agree with all the good points made by others already about low gravity (these dogs took a long time to get trained well against Earth's gravity, they won't "just work" on Mars or the moon). Wheels have their problems but are also very simple and hence fewer ways to fail.
That's more a question for NASA than HN, don't you think? But does NASA have an actual use case for something like this, rather than just rover wheels being a very difficult thing to tackle?
Good luck testing low gravity jumps on earth. If it falls down, literally nothing can help it. Also the pipeline for space tech is decades long.
What about SpaceX? They send a car to space as a publicity stunt... I guess the difficult part is to land the robot-dog in one piece in the Moon.
each wheel has one point of failure whereas those moving limbs have at least two if not more points of failure