economies of scale taper off because other effects (like coordination problems) tend to overwhelm it with scale. also, capital becomes less efficient at scale due to opportunity costs and the innovator’s dilemma, for instance. there are many more reasons why economies of scale aren’t the be all and end all of market economics.

I agree, my point was it depends on industry. You can’t build a modern Fab for a 100 million it simply takes billions to be competitive. Further, you need a pipeline of Fab’s to afford the R&D etc. At the other end a McDonalds franchise can be as profitable as their corporate restaurants.

there's no intrinsic reason a capital-intensive industry can only be winner-take-all. capital-intensiveness is primarily a financing issue, and assuming capital finance has been corralled into being a competitive market, it isn't a particularly insurmountable problem. tokyo public transportation is an example of a capital-intensive market that still has competitive dynamics.

It’s not just about capital, my point was the benchmarks where different between industries. You can scale a soda bottling plant in the way you can’t scale a single sit down a restaurant simply because you can ship soda hundreds of miles without issue.

Fab’s are built at such scale in part because finished chips are really easy to ship being both durable and value dense. They also benefit from that scale because of the nature of the lithography process which has many individual steps which need to be done in the same factory, but each step is very fast and operates on a full wafer at the same time. Millions of chips per year is basically the minimum scale that’s in any way reasonable.

sure, the scales of various industries can be very different, but that doesn't imply that economies of scale (or other economies, like scope) will necessarily dictate winner-take-all dynamics in any of them.

A competitive market sure, but you can’t have 100+ competitive and independent companies running Fab’s worldwide while you can have literally 100,000+ competitive farms, restaurants, etc.

That’s a small enough pool that individual failures starts to seriously consolidate the market. It’s already down to a handful of companies that can currently manufacture modern chips and if anything that number is likely to shrink rather than grow.