Coding Crops: Farming as a Service
By Sam Arbesman
We are seeing growth in a number of fantastic technologies that are poised to cause great change in agriculture. We have sensors embedded in fields, huge amounts of data and analysis available through satellite and drone imagery, and the increased use of robotics in the process of agriculture itself, from drone cropdusters to automated harvesting (my colleagues Bilal Zuberi and Shahin Farshchi are actively investigating these many areas). All of this has been bundled under the term of precision agriculture, using technology and information to change and improve how we grow and cultivate food.
While many of the advances in precision agriculture are still underway and will take awhile to be adopted, I’m already thinking about the next step: coding crops. Coding crops is essentially farming as a service. It is the process of combining all of these technologies into a single automated process for farming, one that can seamlessly scale up and down, and all operating remotely through code. I envision this as a sort of AWS for agriculture, where the various pre-existing and embryonic technologies in satellites, drones, data analysis, sensors, and more could all be plugged in via an API and provide reproducible results for farming. Just as my colleague Adam Goulbourn has mapped out a similar vision in science — remote control biology — we will increasingly be able to do the same thing for farming as a service (see also a16z’s When Bio Meets Computer Science).
So what does exactly mean and what will it look like? Let’s say you want to test a new farming technique such as farming 500 acres of soybeans in a very precise manner under specific conditions. Code up the recipe for it and send it along to a farming service. It will be done automatically, with you monitoring the process remotely in great detail, every step of the way. And when you’re done, the soybeans can be processed or sold in whatever way you want (with the service of course costing a small amount for this, likely as a combination of the total yield and the amount of time for the yield to be completed).
How would this work? Coding crops would likely stitch together pre-existing technologies, in a large-scale partnership with farmers, essentially leasing and using their fields for whatever tasks are needed (Uber meets Ag).
Obviously, there is a lot left to figure out. How would this work? Which technologies still need to be developed? Are individual farmers willing to work with such a service? What are the margins?
But I can see both a need and a potential. Thinking along the same lines? I’d love to help partner in building a powerful company in this space. Please reach out to me at sam.arbesman@luxcapital.com.
