Water is tightly conserved, the main loss is through plant respiration, but even that could be reduced in a greenhouse. Then the main water loss would be either the occasional emergency flushing, when a tank of fish get sick, or through exporting the product itself.
Nutrients are greatly simplified by the fish and insect side of the system. Many of the vegetable byproducts can be converted into feed pellets, or used to grow grubs as feed. Some input may still be needed, beyond atmospheric CO2, in the form of supplemental fish feed.
Automation in aquaponics is unheard of, but it should be much cheaper, because the plants are not embedded in soil. The presence of healthy, voracious predators (fish, frogs, geckos) reduces the insect-related problems. A tolerance for insects reduces fungus problems. There is little or no need to spray the crops.
Are we not scaling up because of a lack of research? Are there just too many unknowns for anyone to invest in the $100M range? Have there already been large scale failures where we were unable to analyze the root cause, due to the biological complexity of the system? This all looks so good on a small scale.