Just two words, capital costs, suffice to throw that entire line of thought to the ground.
If you go through the long list of farming equipment companies, they're all gone now. Defunct, or conglomeration.
Caterpillar? In agriculture? Do you mean Challenger (AGCO) or CLAAS, both of whom have used Caterpillar branding at points in the past? Although neither of them have recently.
Curiously, you're the second person to have mentioned Caterpillar being involved in agriculture in this discussion. Which surprises me. They are not a player at all, outside of some overlap use of their industrial products (skidsteers, payloaders, etc.)
Economics isn't a great description of reality. Its a just so story that is made to make sense after the fact because it lacks the predictive power that defines science.
Messy, stupid, corrupt, emotional, irrational complicated is the rule as opposed to the exception in real life.
For instance no such animal as the free market you describe exists. Just because hard/impossible to repair tractors is a bad thing doesn't mean that anyone is inspired to fix this, that such a company would succeed in the fact of competition ethical and not so ethical.
The extremely probable failure of such an effort doesn't imply that the world wouldn't be a better place if this was so and so if this is important it falls on us as citizens to outline to the players what is acceptable and not via regulation.
"Car company" and "rocket company" are interesting examples, since the "smart, passionate and dedicated entrepreneurs" in those areas needed billions in government subsidies to get off the ground []. And "computer company" includes the software side, with its low capital costs, and the hardware side, where the competitors are firmly entrenched (what's the last laptop/desktop upstart to hit it big?).
[] http://www.latimes.com/business/la-fi-hy-musk-subsidies-2015...
That is patently incorrect. There are many examples of market failures that produce incorrect, poor quality, or even predatory results. A recent example is the total disregard for security in many (most?) "IoT" products, and the market for lemons[1] is the canonical example of a market that specifically selects for bad results ("lemons"). Note that the latter example of markets not producing "better products" was worth a Nobel Prize.
In a truly free market, the product that is most efficient for the market wins. This is often (but not always) the most profitable product. Sometimes it may also be the "best" product, depending on your point of view, because profit (or price) is not the only metric that is used in determining which product is "best".
