We can't even get interop between instant messaging platforms, and video calling platforms. Why would we anticipate that companies manufacturing $200,000+ pieces of equipment are going to work together when they have the option of lock-in?
I would never do that, unless I was actually renting the land. Some people do rent their land, and perhaps committing to such a contract could make sense to them.
The alternative is to simply buy a tractor and then hire someone that does many other things, like driving other tractors, cleaning them and do some repairs and maintenance. (If I could do that to my own AI-powered tractor, maybe then I would buy one). This is about who controls your business.
I believe there's future in modularity, but I also believe some kinds of equipment won't fall into the automation ratchet or the "evolve to AI or die" category.
If you are developing an AI sophisticated enough to pilot a tractor, would it only be able to drive one model? If John Deere ever achieved this goal, the AI should be adaptable to other brands as well.
Car manufacturers are doing this already with Android. They do not develop the OS, they simple accommodate their cars for it.
I have a working GPS system that belonged to my father's John Deere tractor. I cannot use it anymore, even in other John Deere tractors. This is wasteful and makes no sense. People working in farms, close to nature and subject to profit margins they do not completely control feel very uncomfortable with this.
In which case you still won't be locked out of your machine, in the sense that you can choose the company with the best farming AI.
The only way for a quick automation scenario is to make crops a lot more predictable. In our days, this means linearly sown and harvested, artifact free crops.
If we choose a more hands-on approach to farming (and food production in general for that matter), then automation ceases to be as interesting.