That's related to the advancement of the underlying technology powering computers, transistors. It was a young a quickly improving technology scaling on both effectiveness and efficiency at every new improvement. I don't see this for AI until new computer technology is available, and I especially don't see it with robots.

Furthermore, just like OP, you're only thinking and you bring an example related to a big company, a government company from the wealthiest country in the world no less, as representative of a pervasive reality in the upcoming decades. I know the US is heavily stardardised and every industry is dominated by one or more of such gargantuan companies, but that's not the reality everywhere else in the world, nor is it likely to be an applicable model in other countries apart from a few extremely densely populated areas.