Yes and no. If we were to somehow develop a way to power robots w/o having to attend to them constantly, maybe, but that's unlikely. Solar seems unlikely except in the most passive or sedentary application, I don't need to mention nuclear, batteries raises the question of how to recharge them, and the most likely (some sort of combustion) requires refueling and maintenance. Mechanical failures would likely require maintenance by people, and the more complicated and cumbersome the system then the more difficult it will be to recover and then repair those robots. Then you have ammunition concerns...how do you resupply 50,000 robots in the field? If your goal was to keep soldiers out of harms way then forcing soldiers to constantly repair/supply/recover robots in the field brings the entire thing into question.
It's going to either have to be a stupendously gigantic undertaking where almost everything is automated (currently not plausible) or you're going to have to use robots in specific roles (EOD, recon, etc) where they're used only as a tool rather than being used as a wholesale replacement for soldiers across the board. Which is the gist I get from the article, but I imagine even that is probably far too optimistic.
The real rub for me isn't that "people are expensive", though that's true. If that was the only concern they could address that without having to bring robots into the mix. The reality is that in this day and age a draft will almost never happen and they can't get enough people to sign up. Not only that, but when we draw-down between conflicts we then need to recruit unbelievable amounts of people when we do go to war, and as Iraq and Afghanistan showed us, it's impossible to do. When at one point the DOD was forced to make up for the military manpower shortage by hiring so many civilian contractors that there were actually
more contractors in theater than combat troops, you know that there is a critical issue with recruitment in general and upsizing/downsizing specifically. And bear in mind that if you think that soldiers are expensive, just imagine how much contractors cost.
Costs are somewhat a secondary concern, as bizarre as that sounds. The gov't doesn't have an issue with spending trillions of taxpayer dollars. They just need to address our endemic manpower shortages that we have when we go to war and they're hoping that robots can do what they last did with hiring contractors. Except instead of hiring contractors to move soldiers out of the rear and into frontline roles, they want to try using robots to move soldiers out of front line roles and into the rear. I don't think it's going to work as well as they hope, though.
I'm still laughing that the OP article mentioned wanting to control costs...ludicrous.