Rising energy demand, from electrified transport to AI data centres, is forcing a harder conversation about firm low-carbon capacity than the one the sector has been having.
Why the question changed
For a decade the nuclear debate was framed around cost per megawatt against a falling solar curve. That framing quietly assumed demand growth would stay linear and that storage would arrive on schedule. Neither assumption survived contact with the last three years of load growth.
What has changed is not the physics but the load profile. Data-centre clusters and electrified industry want power that is firm, sited near demand, and contracted over decades. That is a different procurement problem from adding intermittent capacity to a grid with spare balancing headroom.
Where advanced designs actually differ
- Small modular reactors trade unit output for factory fabrication and a shorter construction path. The bet is on learning rates, not on any single plant.
- High-temperature designs open industrial process heat, which is the part of decarbonisation that electrification alone does not reach.
- Fuel-cycle choices determine the regulatory and non-proliferation conversation more than reactor size does.
For African markets the binding constraint is rarely technology. It is the regulator’s capacity to license, the grid’s ability to absorb a large single unit, and the sovereign’s appetite for a thirty-year liability.
What we advise clients to do now
Treat advanced nuclear as an option to be kept open rather than a decision to be made. That means investing in the regulatory and human capability that any future pathway requires, while making near-term capacity decisions on grounds that hold regardless of which technology matures first.
The countries that will have a genuine choice in 2040 are the ones building licensing capability in 2026.
Renewables plus storage remain the fastest route to added capacity in most of the markets we work in. The argument for engaging with advanced nuclear now is about optionality on firm capacity, and about not discovering in fifteen years that the constraint was institutional all along.
