Building robots or other physical devices is all the rage, but I have a few concerns. 

• Is it a necessity for covering the Computing curriculum? No! It might be a "nice to do" (might be), but there are other objections (see below). I get that sometimes concrete learning is more effective than abstract learning, but I don't see why screen-based cause and effect scenarios are regarded as inadequate, especially at secondary (high) school level. You might want to make your  Computing lessons more interesting, but you don't need robotics to achieve that. I give a talk, which I update every so often, on ways to make your Computing curriculum more interesting. People always leave bursting to try out a few of the suggestions -- and there is no mention of robots anywhere!

• What about time? When I taught Computing I had between one and three hours a week with classes to teach it. Given the amount of other stuff which needs to be covered in the syllabus, I couldn't afford to spend hours and hours building and programming robots.

• Money? One device I looked at at the 2019 Bett show costs £200-ish. It is brilliant, but if you have classes of 30, and you buy just one set of 15, that would probably blow your entire year's budget. I know it would have blown mine, leaving no money for licence renewals, printer paper and assorted other things.

• Technician help? Unlike science teachers, teachers of Computing tend not to enjoy the services of a technician who can put out all the stuff they need for the lesson, then clear it away afterwards.

• Collaboration? A massive waste of time in my experience. I've taken part in a few training sessions in which I was partnered with others, fiddling about with a device. On one occasion, I was in a group of three. One person got the required bits out of a box. Another person fitted them together. The third person programmed the finished contraption. So after half an hour, we learnt that if you enter the word "Fwd 10" into the the program, this thing would move forward 10 units. The same thing could have been learnt in ten seconds by each person in the class using Logo or similar.

All of the above points to the question of ecological validity. That means, that when you hear people saying how marvellous physical computing is, maybe they have a lot more resources (time, technician help, money) at their disposal than you do.