Skip to main content
EU_ECONOMICS06 / 08 · story of the day3 min · 524 words · 149 sources

Italy ends 40-year ban on nuclear energy

Written by AIto brief AI · 5 June 2026, 03:50
How it was written

A modular nuclear promise sits on a pallet, awaiting an infrastructure yet to be built.

Image composition · tobrief
the text · 3 min read

Italy's lower house voted on June 4 to reverse a four-decade ban on nuclear energy. The 155-86 vote overturns referendums from 1987 and 2011 that made atomic power politically untouchable. The Meloni government wants small modular reactors (SMRs, compact nuclear plants designed for factory assembly) running by 2034. The technology doesn't exist commercially anywhere in the West, and France is already using its nuclear fleet to capture the energy-intensive investments Italy needs.

Twelve months before the real decisions

What parliament approved is a legge delega, a framework law that authorizes the government to draft detailed implementing rules within 12 months (Camera dei Deputati). No reactor technology has been selected. No sites have been chosen. Italy dissolved its nuclear safety regulator in the 1990s and hasn't rebuilt one. The bill still needs Senate approval before summer recess.

The price motivation is real. Italian electricity costs above €105/MWh, nearly double the roughly €60/MWh France pays thanks to its 56 operating reactors (Élysée, Editoriale Domani). Meloni's own government acknowledges Italy is "completely out of the market on energy prices." That gap hits factory costs, household bills, and Italy's ability to attract data centers that demand massive, cheap, stable power.

SMRs are supposed to fix all of this. But no Western country produces commercial electricity from one. The most advanced Western project, NuScale's plant in the United States, was cancelled in 2023 after costs nearly doubled. Large European reactor builds in France and Britain have consistently run billions over budget and a decade behind schedule. Nuclear construction in complex regulatory environments costs far more and takes far longer than projected. This is the record Italy is betting against.

France converts reactors into AI investment

The divergence is happening now, not in 2034. At the Choose France 2026 summit, SoftBank committed €45 billion for AI data centers in northern France, powered by cheap, round-the-clock nuclear electricity.

The talent follows the infrastructure. Newcleo, a nuclear energy startup founded by Italian physicist Stefano Buono, relocated from Italy to France, where it received backing through France 2030, the country's industrial investment program. Italy is exporting its own nuclear expertise to the competitor that already has reactors.

The World Nuclear Report's assessment is blunt: Italy's nuclear plans are "irrelevant" for European power flows until at least the mid-2040s.

The financial gamble

Italy's energy cost problem is genuine. Its factories pay nearly twice what French competitors do for electricity. Diversifying away from imported gas makes economic sense.

The question is whether SMRs are the right vehicle. New nuclear power typically costs several times more per megawatt-hour than new solar or wind. For a country carrying one of Europe's highest public debt loads, underwriting an unproven technology at scale is a financial gamble layered on top of an energy strategy.

The June 4 vote buries the referendums of 1987 and 2011. Politically, that matters. Economically, it changes nothing for at least a decade. The investments Italy needs — AI infrastructure, advanced manufacturing, electrified industry — are being placed now, in countries that already have the power to supply them. Italy voted to want nuclear energy. Building it requires 15 years of engineering, regulation, and spending that no parliamentary vote can shorten.

How was this article?

Help us get better

Details about this article
Model:
claude-opus-4-6
Generated:
6/5/2026, 3:13:41 AM
Pipeline run:
eu_pipeline_20260605_015006
Watermark:
SynthID (Google's invisible watermark)
Human review:
None before publication
Learn more about our methodology