I'm working an an agricultural robotics startup and can confirm that if anything, it's challenging to find robotics use cases in specialty crops that have a large enough market to justify the R&D investment necessary to get the product working in the first place. Corn and soy add up to hundreds of billions of dollars of market value a year, while even a huge specialty crop like strawberries is ~1/100 of that. There's an argument to be made for building out a robotics platform in a crop with higher value per acre (i.e. a specialty crop), then moving to the larger markets (i.e. row crops), but the degree of specialization real world robotics requires is a counterpoint to that - making the jump is often nontrivial.

PS - if anyone reading this is a roboticist/mechanical engineer in the Bay Area and interested in working on something like this or even just talking more, let me know! My email is my username (at) stanford.edu