Life-Cycle Water in Materials and Fuel
A complete water account includes the water used to mine, process, and manufacture a plant's materials and fuel, not only its cooling.
Operational cooling water is the visible term, but a full life-cycle account also includes the water embedded in building the plant and supplying its fuel. Mining, ore processing, metal refining, concrete production, and semiconductor-grade fabrication all consume water. A fair comparison acknowledges this upstream term even when it is harder to quantify.
Where fusion's upstream water sits
- Superconductor and magnet materials (REBCO tape, copper, structural steel) carry manufacturing water.
- Blanket materials for the breeder — lithium, beryllium — carry mining and processing water.
- Fuel: deuterium is separated from water; tritium is bred rather than mined; helium-3 supply is a separate, constrained question.
- Construction concrete and steel carry embodied water like any large facility.
The key structural point is that fusion fuel is not a continuously mined, water-intensive feedstock the way some thermal-fuel chains are. Deuterium is abundant in ordinary water; the breeder makes its own tritium; the burner's helium-3 constraint is about supply existence, not water intensity.
Honest limits
We do not publish a full life-cycle water inventory for an unbuilt machine. The defensible statements are directional: the burner shrinks the operational term via direct conversion, and fusion's fuel chain is not a large ongoing water consumer. Both are stated as design-and-simulation reasoning, not measured totals.
The reason fusion's fuel term is small deserves emphasis. A thermal-fuel chain consumes water continuously to extract, process, and sometimes cool its fuel over the plant's whole life. Fusion's deuterium is separated once from abundant water, its tritium is bred on-site, and its helium-3 constraint is one of existence, not water intensity. So the recurring, life-long water draw that burdens some fuel chains is largely absent.
We still decline to publish a full inventory for an unbuilt plant. The defensible contribution is structural: identifying which life-cycle terms are large, which are one-time, and which are simply not water-intensive for fusion.