Nuclear power for infrastructure planners: fission now, fusion on the horizon

I am not a reactor physicist and this is not reactor physics. But nuclear stations are, from the fence line inward, immense water and civil engineering projects — cooling flows that dwarf a city's supply, seismic structures, marine intakes of the kind I have spent a career around — and the region I work in is now a nuclear operator. Planners deserve a plain account of what these machines ask of their surroundings.

Fission: the machine we have

A pressurised water reactor is thermodynamically a steam plant with a 300 °C heat source: roughly one-third of thermal power becomes electricity and two-thirds must be rejected. A 1,400 MWe unit on once-through seawater cooling moves on the order of 50–70 cubic metres of seawater per second — an intake, screening and outfall problem larger than most desalination plants, and the reason coastal nuclear and large desalination pair naturally. Barakah's four units next door to this market demonstrated both the model's credibility and its appetite for disciplined delivery.

• Small modular reactors (50–300 MWe) trade economy of scale for economy of series — factory fabrication, shorter sites, passive safety. For desalination and industrial steam they are the most interesting fission format in decades.
• Cooling remains the siting decision: once-through, evaporative towers or dry cooling — the last costing efficiency exactly where ambient temperatures are highest.
• The civil package — nuclear island concrete, seismic qualification, marine works — is where schedules are historically lost; the QA culture is the one I described for nuclear wastewater, applied to everything.

Fusion: real physics, unbuilt engineering

Fusion joins light nuclei — deuterium and tritium — at plasma temperatures above 100 million °C, confined magnetically (tokamaks, stellarators) or inertially (lasers). The physics milestone is real: the National Ignition Facility achieved ignition in 2022, and ITER is assembling the largest tokamak yet to demonstrate sustained burning plasma. The engineering milestones are not yet: no fusion device has produced net electricity, and between plasma gain and a power station stand tritium breeding blankets, plasma-facing materials surviving years of neutron flux, and the balance of plant nobody has built.

When it comes, a fusion station will still be a steam plant to the grid and a water plant to its site — heat rejection of the same order as fission, tritium handling as its radiological signature, no fission products and no meltdown pathway. Private ventures target the 2030s for pilots; a planner should treat commercial fusion as a 2040s-plus entrant, welcome it if it arrives early, and size no reservoir on its promises.

The planning posture

Fission is bankable now and pairs well with desalination and hydrogen production at night-time surplus. SMRs deserve site-banking studies today precisely because their civil and cooling footprints are modest. Fusion deserves attention and scepticism in equal measure — the correct hedge is coastal land, grid corridors and cooling-water rights that serve any thermal machine, whichever physics wins.