A blogger created a calculator that allows you to determine (over a 5-year time frame) whether it is cheaper to buy an expensive robot and do automation or to just use manual labor and to save the excess in Treasury bonds. He found that when the minimum wage is at $7.25, it would be slightly cheaper to hire humans, and if the minimum wage is raised to $10.10, then it would be slightly cheaper to hire robots.

The blogger assumes that it would cost $150,000 to build the robot, including R&D and maintenance. This comes out to $30,000 per year (while your assumptions claim that the bots only cost $20,000 per year to replace one human).

http://politicalcalculations.blogspot.com/2014/01/business-m....

I don't know which assumption is correct, though I would guess that eventually bots will be cheaper as we get better making them.

I'm thinking specifically in the case of McDonalds and other large chains where the costs of R&D and dedicated maintenance staff would be spread over thousands of restaurants.

You're forgetting Onion2k's point that it's cheaper for companies to hire robot-replacable employees because it loweres the wages of the non-robot-replaceable labor. Under that theory, the reason they can't bring in the robots is that their remaining labor costs would increase.

Personally, I don't know why that would be, given that companies pay what they need to get employees to work there and nothing about the NRR's job would change if the robots came in.

Even if the robot costs more today, that wouldn't change the inevitability of it happening. A $10k robot this year will be a $5k (in today's value) robot in 5 years. While wages go up, tech costs go down. No one will escape that.