I ran the numbers on a small sugar cane setup: 16 rows with observer-piston modules are easier to expand than one oversized machine. My rule is to price the farm by output per hour after storage losses, then compare that return with simply trading the materials. A farm is only profitable when the stock actually sells.