The cost of America’s nuclear revival

On an industrial site on the shores of Lake Michigan, hundreds of workers
are racing to do something that has never been achieved before: restart a
US nuclear power plant scheduled to be decommissioned.

The revival is being
driven by power demand from artificial intelligence data centres and
reshoring of manufacturing, which is straining energy grids and pushing up
electricity prices. Big Tech is attracted to nuclear because it provides
[?] clean power that does not suffer from the intermittency that is a feature
of solar and wind energy.

Concerned that a power crunch could delay the
roll out of data centres, Microsoft and Google have signed long-term power
deals to support the reopening of shuttered nuclear plants in Iowa and
Pennsylvania.

President Donald Trump has also embraced atomic energy,
pledging to slash regulation and invest tens of billions of dollars to
reopen and build new fleets of reactors to provide the power needed to
“win” the global AI race. In May he set a goal of quadrupling US atomic
energy capacity by 2050, a massive undertaking that would involve building
hundreds of new plants.

And in October Washington inked an $80bn
partnership with private equity group Brookfield and reactor designer
Westinghouse, which aims to kick-start construction of eight large-scale
nuclear plants.

Last year private investors funnelled a record $3bn into
start-ups building smaller modular reactors, technologies that provide a
third or less of the power of large-scale nuclear plants. In 2024, Amazon
bought a stake in X-energy, a reactor developer with power supply
agreements to support the construction of 144 small modular reactors (SMRs)
in the US and UK.

But in an industry subject to high financing, regulatory
and reputational risks, not everyone is convinced the resurgence will be
sustained. Only a handful of mothballed nuclear plants in the US are
capable of being restarted, critics warn, and building large-scale reactors
from scratch is a highly complex undertaking prone to delays and cost
overruns in western nations.

None of the more than 50 SMRs under
development in the US has proved they are commercially viable yet nor
obtained an operating licence from the US Nuclear Regulatory Commission.

“The industry is boasting about how things are different now and they are
going to build fleets of new reactors, large and small,” says Edwin
Lyman, a physicist at the Union of Concerned Scientists and critic of
nuclear energy. “But the fundamentals haven’t changed: nuclear is more
expensive than other forms of energy and still poses a risk of accidents
and proliferation. You can’t wish these problems away, so I think
Trump’s so-called nuclear renaissance is built on a house of cards.”

FT 7th Jan 2026,
https://www.ft.com/content/9f6c4db1-559f-48e1-8c21-ac0bc1a1237c