Uzbekistan farms in a double-landlocked basin fed by two rivers, and that single fact shapes the entire sector.

For most of the twentieth century the planting plan was set centrally and cotton took the best land and most of the water. Policy has moved deliberately away from that: area has been reallocated toward horticulture and grain, and the state procurement system that enforced cotton quotas has been dismantled.
Horticulture is where the growth is. Apricots, melons, grapes and tomatoes travel well and earn considerably more per hectare and per cubic metre of water than cotton does. The constraint has shifted from growing to handling: cold chain, grading and the paperwork of export markets.
Most of the canal network is old, unlined and loses a substantial share of what enters it. Modernisation competes for money with everything else and pays back slowly, which is the familiar reason it is perennially underfunded.
Any forecast for this sector is a forecast about water allocation upstream, and those decisions are not made in Tashkent.
Cotton picking was for decades organised as a mass seasonal mobilisation, and dismantling that system changed the economics of the crop as much as any policy on water or area. Mechanisation covers part of the gap, but machine picking needs level fields, specific varieties and dry conditions at the right moment, so the transition is uneven across regions.