The editorial frames Meloni's pitch as a pragmatic response to the 2022 gas shock, arguing that Chernobyl-era safety fears have been superseded by the concrete costs of running an industrial economy on imported Russian methane. The counter-argument that gas dependency is more dangerous than nuclear risk carried parliament comfortably.
By surfacing the AP story to 804 points, the HN submitter and upvoters signal that Italy's reversal is a significant energy-policy story worth engaging with — treating the return to nuclear as consequential news rather than a fringe political move.
The editorial emphasizes that the bill deliberately centers small modular reactors and Generation IV designs rather than the gigawatt-scale PWRs that Italy abandoned in 1990. With an 11%-by-2050 target and first SMRs online in the early 2030s, this is presented as a technology bet distinct from the reactors that Chernobyl-era voters rejected.
The opposition and green groups have announced they will push for another referendum, arguing that a party-line parliamentary vote should not be allowed to overturn the 1987 popular mandate against nuclear power. Their position is that direct democracy, not coalition politics, must decide a question this consequential.
The editorial argues that this Italian bill matters to a developer audience because hyperscaler site selection in Europe has become a power-siting problem. Nuclear capacity — especially dispatchable SMR capacity — directly shapes where AI training clusters and data centers can plausibly land over the next decade.
On a party-line vote led by Prime Minister Giorgia Meloni's coalition, Italy's parliament passed a bill establishing the legal framework for a return to nuclear power generation. Italy became the first G7 country to abandon nuclear energy after the 1987 referendum triggered by Chernobyl; it is now the first to formally reverse course. The last of the country's four operating reactors shut down in 1990. Everything since — the imported French nuclear electricity, the gas dependency, the winter price spikes — has been the consequence of that 1987 vote.
The bill does not authorize construction of any specific plant. It establishes the regulatory scaffolding: a national safety authority, siting criteria, waste management responsibilities, and — critically — a defined role for small modular reactors (SMRs) and Generation IV designs rather than the gigawatt-scale pressurized water reactors that dominated the 1970s. Energy Minister Gilberto Pichetto Fratin has publicly targeted 11% of national electricity from nuclear by 2050, with first SMRs online in the early 2030s. Environmental groups and the opposition Five Star Movement have already announced they will push for another referendum.
The political frame matters. Meloni's government pitched this as energy sovereignty, not climate policy — a response to the 2022 gas shock, when Italy discovered exactly how expensive it is to run an industrial economy on imported Russian methane. The Chernobyl-era arguments ("nuclear is dangerous") were met with 2020s-era counter-arguments ("gas dependency is more dangerous") and the counter-argument won by a comfortable margin in parliament, if not yet in the polls.
For a developer audience the obvious question is: why does an Italian energy bill belong on a tech news site? Because the constraint on where you can put compute has quietly become power, not silicon. Hyperscaler site selection in Europe now looks less like a real-estate problem and more like a grid-capacity problem, and Italy just changed its answer to that problem.
The numbers are stark. Ireland's grid operator has effectively frozen new data center connections in the Dublin region through 2028. The Netherlands imposed a moratorium on hyperscale builds in 2022 and only partially lifted it. Frankfurt is running into transmission bottlenecks. The UK is negotiating dedicated nuclear PPAs for AI campuses because there is no other way to get 500 MW of firm 24/7 power on a two-year timeline. Meanwhile Italy — the fourth-largest economy in the EU, sitting on excellent submarine-cable landings from North Africa and the Middle East — has been effectively off the map for large training clusters. Not because of latency. Because of electricity.
The SMR framing is the tell. If Italy were serious about big baseload reactors, this bill would look different — you don't need a modular-friendly regulatory regime to permit a single 1.6 GW EPR. The SMR bet is aimed at exactly the customer profile that has emerged in the last three years: industrial and data-center loads that want firm, carbon-free power on a colocated or near-sited basis. Amazon signed with X-energy in Washington state. Google signed with Kairos. Microsoft signed to restart Three Mile Island Unit 1. Meta is shopping. None of these deals existed in 2022. Italy is not writing this legislation in a vacuum; it is writing it in response to a market that materialized 18 months ago.
The skeptical case is worth stating clearly, because it's serious. No SMR has yet been commercially deployed at scale anywhere in the West. NuScale cancelled its Utah project in 2023 after costs ballooned to $89/MWh, well above market. Rolls-Royce SMR is still in UK generic design assessment. TerraPower broke ground in Wyoming but faces HALEU fuel-supply questions. The gap between "parliament passed a framework" and "a reactor is generating electrons" is measured in decades, not news cycles, and Italy's regulatory apparatus for nuclear was literally dissolved in the 1990s. A generation of Italian nuclear engineers has retired or emigrated. Rebuilding that capability is not a line item; it is a project.
Still, the direction of travel is unambiguous. Belgium extended its reactors instead of closing them. Sweden legislated for new build. The Netherlands is planning two large reactors at Borssele. Even Germany — the country that made "Atomkraft, nein danke" a global brand — is watching its industrial base bleed to countries with cheaper, firmer power. Italy joining that reversal moves the EU's political center of gravity another notch.
If you are architecting anything power-sensitive in Europe over a five-to-ten-year horizon, Italy just became a location worth modeling. Not tomorrow — the bill is a framework, not a groundbreaking — but on the timelines that actually matter for real estate, transmission upgrades, and PPA negotiations, this is a supply-side signal. Genoa, Trieste, and the industrial corridor around Milan all have decent fiber and, more importantly, historical industrial power infrastructure that never went away when the reactors did.
For AI infrastructure teams specifically, the implication is narrower but sharper. The cost curve for training frontier models is increasingly dominated by the electricity bill, not the GPU depreciation. A cluster running on €0.35/kWh Italian retail power is uneconomic against one running on €0.08/kWh French nuclear or €0.05/kWh Nordic hydro. If Italy delivers even 3-5 GW of new nuclear capacity by the mid-2030s with dedicated industrial PPAs, the country moves from "never" to "plausible" as a training-cluster location.
For carbon-accounting teams — and if your company has an SBTi target, that's you — Italy's grid intensity is currently around 250 gCO2/kWh, roughly triple France's. Any nuclear buildout drags that number down mechanically, which changes the math on Scope 2 emissions for every EU workload you run there. The Renewable Energy Guarantee of Origin market gets messier in interesting ways when a major consumer country adds a chunk of low-carbon baseload.
For anyone selling into European enterprise, expect the sovereignty argument to keep gaining ground. "Data stays in the EU" was the compliance story of the 2020s. "Compute runs on European firm power" is shaping up to be the compliance story of the 2030s. The regulatory scaffolding for that story is being written in Rome this month.
The near-term risk is a second referendum. The Five Star Movement is competent at mobilizing exactly this fight, and Italian public opinion on nuclear is roughly evenly split — a very different picture from 1987's landslide against. If the referendum comes and fails to overturn the law, the political ceiling on Italian nuclear rises dramatically and the SMR vendors will start filing site applications within the year. If it succeeds, Italy re-enters the "never" column and the EU's data-center gravity keeps pulling toward France, Spain, and the Nordics. Either way, the interesting move for practitioners is to stop treating European power policy as background noise. It is now, unambiguously, part of the stack.
China is building nuclear reactors faster than any other country: https://news.ycombinator.com/item?id=38485137Belgium stops decommissioning nuclear power plants: https://news.ycombinator.com/item?id=47961319Canada plans 'nuclear renaissance' with up to 10 rea
I'm Italian and I'm super happy with this.The referendum that made nuclear energy illegal in Italy was held right after the accident of Chernobyl; people voted with their gut, not with their mind.Curiously enough, some Italian town signs still have a small sub-sign reading "denucleari
People usually do not include the costs of actual insurance into those calculations. If you do you end up with much higher prices per kWh for nuclear.Additional cost estimate in numbers per kWhNumbers:Germany: + 0.139€-2.36 €/kWhJapan: + $0.22–5.78/kWhI could not find numbers for the US, b
It is sad that this has become a topic of the culture wars. Nothing good comes out of this, certainly not a rational weighing of the pros and cons of nuclear power.I’m particularly interested in how they plan to finance reactors in a grid dominated by solar. I wouldn’t be surprised if they legally a
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.
I've seen quite a few SMR proposals over the last couple of years and not a single one that dared to address the whole cycle from deployment to eventual decommission as a function of the total revenues - costs to operate. Most of these projects seem to be made to bilk investors or governments r