Big Tech got so big because digital technology is different from preceding industries in that increasing (not decreasing) returns to scale prevail. In particular, in software where marginal costs are near 0. It's a different ballgame. I think regulators have figured this out but for the most part it appears that they don't know how to act upon it.
Traditionally, you had _decreasing_ returns to scale. What's new with digital technologies is _increasing_ returns to scale. Very counter-intuitive for traditional economists - or even business people. People like Gates, Bezos, Brin & Page, Zuckerberg and so on have ridden this for all its worth. Let's out distance them as much as we can before they figure this out. Whether competitors or regulators.
That said, I like the story about someone observing Gates at, I think, a graduation ceremony where Gates had brought along a biography of Rockefeller (John D.). Did Gates know before or after the fact that MS had in fact used a competitive strategy comparable to Standard Oil?
Meaning, unlike increasing returns to scale that was something for which there was an historical precedent. The basic idea being: when my competitors revenues go up, my revenues also go up. A fatal long-term proposition for those competitors.
GP is talking about the marginal benefits of scale. In manufacturing, there is a floor at which increasing scale doesn't decrease the cost of making each additional unit. No matter how much a widget company automates their factories, they can't bring down the cost of a widget below the cost of the raw materials. At some point, investing into making widgets becomes a negative ROI.
That often never happens with software because network effects allow each unit of raw material (servers, data, etc.) to produce more value as part of the whole. It would be as if a widget factory could make 2 pounds of widgets out of 1 pound of raw material when it double its sales.
https://en.wikipedia.org/wiki/Returns_to_scale