In order to do that at scale, they would need for example an optical sorter or something to differentiate for the size distribution which i don't believe they have ever thought about. I also believe they have always had an issue with size distribution which causes a major problem. For example, just imagine that 30% of all shrimp in the final grow out tank have not reached their target size, do you harvest that 30% or leave them in the tank.
Market Data 
Markets 
AI
