Listen to the AI generated audio article.
Armenia is once again facing a consequential choice about its nuclear future.
A few weeks ago, Rosatom, Russia’s state atomic energy corporation, formally offered Armenia a VVER-1000 unit, while Prime Minister Nikol Pashinyan declared that a small nuclear reactor (SMR) will most likely not be enough to meet the country’s future electricity needs. At the same time, advocates of the Russian nuclear option have become increasingly vocal.
There are some common talking points that are regularly repeated and recycled by the proponents of the “Russian path” and the associated 1000 MWt VVER-1000. While these contain some legitimate points, the proponents often engage in cherry-picking fallacies and distort some of the most salient points. Thus, it’s worth analyzing these narratives with closer scrutiny. Below, I examine several of these claims and present some counterarguments of my own.
Claim 1: The choice of the nuclear reactor must be purely technical and economic, not geopolitical
Sounds very tempting. Let’s assume that’s true.
Then why not simply replace Metsamor with more natural gas-fired capacity? At Hrazdan and Yerevan, Armenia already has about 1,400 MW of installed thermal generating capacity, most of it used as load following generators. Let us lay new pipelines from Russia, Iran or even Azerbaijan and generate an additional 400, 600, 1,000 MW. Natural gas fired power plants are faster and cheaper to build than nuclear power plants. Why is this a bad idea?
Because it would make Armenia even more dependent on those countries, correct?
Nuclear power, by contrast, lets us preserve at least some energy sovereignty: fuel can be stockpiled for years; gas cannot. Similarly, when we talk about nuclear, the potential dependency on Russia for a nuclear reactor may undo what advantages one gains from a nuclear power plant.
To be clear, the moment we talk about energy, we enter the realm of geopolitics, whether we like it or not. It would be naïve to expect otherwise.
Claim 2: An American small modular reactor (SMR) is risky
If we are talking about risks, let’s start with the biggest one: the stability of the supplier country.
A nuclear plant is an 80-year marriage: about 10 years of construction, 60 years of operation, and another 10 of decommissioning. Throughout those 80 years the supplier country is expected to provide servicing and personnel training.
The U.S. political system has functioned without interruption for 250 years. In Russia, the state system has collapsed twice in the last hundred years, in 1917 and in 1991. Who will guarantee the next 80 years?
There is also the question of legal guarantees.
If an American company breaches the contract or fails to deliver, the dispute goes to international arbitration in London, Paris or Stockholm, and the court’s decision can be enforced, because the company’s assets are within reach and the courts are independent.
This is not theory: when Westinghouse was unable to meet its contractual obligations for the construction of the nuclear plant in the U.S. state of Georgia in 2017, its parent company, Toshiba, paid $3.68 billion under its guarantee to the plant’s owners.
What will happen in Rosatom’s case?
On paper there is arbitration. But the plant is, as a rule, built with a Russian state loan; that is, Armenia would be suing the very state it owes money to. Russian law also allows Russian companies under sanctions to move a dispute to a Russian court, regardless of what the contract says.
The Finns have been in litigation with Rosatom since 2022 over the Hanhikivi plant that was never built, with mutual claims in the billions. And we need not look far for another example: Armenia paid for Russian weapons, did not receive them on time, and the prime minister publicly confirmed this in 2023. The matter was not resolved in court. Now imagine the same with a multibillion-dollar nuclear plant.
There is an even more tangible risk: sanctions. More than $2 billion earmarked for the Akkuyu Nuclear Power Plant, which Rosatom is building in Turkey, was frozen at an American bank, and construction is running late. Does Armenia want to end up in a similar situation?
Claim 3: The United States has not built a reactor in 20 years
This is a distortion of reality.
In 2023 and 2024, two Westinghouse AP1000 units came online in the U.S. state of Georgia. Yes, they were behind schedule and at far greater cost than planned, and that must be stated openly and honestly. But the experience exists, and it works.
Of the roughly 37 units under construction in China, nine are the Chinese licensed version of that same American AP1000 (the CAP1000). Poland has chosen three AP1000s for its new reactor fleet, and Bulgaria has selected two.
How about SMRs?
At the Darlington site in Canada, GE Vernova Hitachi’s BWRX-300 has been under construction since 2025. It will be the first of four units. Just weeks ago, the same reactor design received a construction permit in the United States as well. Romania has taken a final investment decision on six NuScale modules.
The critics of SMRs are right about one thing: there is not yet an operating SMR in the West, and American construction experience and supply chains are only now recovering after a long pause.
But the SMR is not a technology invented from scratch. The leading designs, such as BWRX-300 and AP300 are scaled-down versions of light water reactors (LWRs) that have been tested for decades, with additional passive safety systems. “Modular” means that the main components are assembled in a factory, not on the construction site, which is faster, more controllable, and more predictable.
Most Important of All: The Size of the Grid
This, in my view, is where the discussion becomes particularly important for Armenia.
Armenia consumes about 1,000 MW of electricity on average. Now imagine a single 1,000 MW unit on this grid. In effect, the entire country would depend on a single machine. And that machine stops for weeks every 12–18 months for refueling, to say nothing of unplanned outages.
Who will keep the lights on during those days? The thermal power plants?
Then Armenia would have to keep another 1,000 MW of gas-fired capacity that would stand idle most of the year (specifically, 95%) but would still have to be maintained. That is very expensive and inefficient.
The example is right in front of us.
This year Metsamor was shut down for months, and gas-fired thermal plants filled the gap. That was a matter of 400 MW. What if it had been 1,000?
The benchmark adopted in IAEA guidance is that the largest unit should not exceed about 10% of the system. With Metsamor, Armenia is already beyond that limit—30%.
Belarus, in order to connect large Russian units to its much larger grid, had to plan for 800 MW of backup gas turbines.
“But Armenia is building AI factories, so consumption is going to grow.”
Let’s do the math together. The Firebird High-Performance Computing facility is expected to consume 300 MW after completion.
The question is: how many Firebirds does it take for a 1,000 MW unit to carry no more than the same weight on the grid that Metsamor carries today—30%?
The answer is seven.
For a 1,000 MW unit to constitute 30%, total consumption must be about 3,000 MW
Today Armenia consumes about 1,000 MW.
Firebird’s target is 300 MW (only about 15 MW is operating today).
So: n = (3,000 – 1,000) / 300 ≈ 7
Seven Firebirds, at their full planned capacity.
And by the IAEA’s 10% rule, 30.
That is the number of AI factories that would justify a single 1,000 MW unit. Armenia is very far from that.
Recently Prime Minister Pashinyan also stated that Amazon would be ready to invest in AI factories in Armenia if there were additional 1,000 MW of electricity available. Now suppose the prime minister is right, and Armenia really will need another 1,000 MW.
That is an argument for more capacity, not for one giant unit.
This is exactly what Canada is doing at Darlington: four 300 MW units, 1,200 MW in total. If one stops, you lose 300 MW, not 1,000. And you build them one after another, in step with real rather than hypothetical demand.
The critics are right: the choice of a nuclear plant must be economically justified.
That is precisely why a single 1,000 MW unit is the wrong answer for a grid the size of Armenia’s, no matter who the manufacturer is.
Opinion
Toward a Polycentric Transnation?
What if the Armenian Diaspora stopped asking what it can do for Armenia and began asking what it needs to become in its own right? Drawing on the history of New Julfa, Tigran Yegavian argues for a polycentric Armenian world in which Armenia and the Diaspora remain connected without one being subsumed by the other.
Read more
Tlön, Uqbar, Gandzasar: Notes on the Imaginary Land Called “Caucasian Albania”
Using Jorge Luis Borges’ imagined world of Tlön as a lens, Sebouh David Aslanian examines Azerbaijan’s decades-long reinvention of Caucasian Albania, arguing that a fabricated historical narrative has been deployed to appropriate Armenian heritage, erase the region’s Armenian past, and reshape the South Caucasus in the present.
Read more
From Survival to Statehood: Reflections on a Sovereign Armenia
Drawing on Armenia’s history and recent security challenges, K.M. Greg Sarkissian reflects on what sovereignty means for a small state navigating competing powers and argues that resilience depends not on choosing between geopolitical camps, but on strengthening institutions, diversifying partnerships, and building a country where Armenians can confidently build their future.
Read more
When “Diplomacy” Becomes a Vehicle for Information Manipulation
The European Union has invested heavily in Armenia’s resilience against foreign information manipulation. But when its own Special Representative reduces Azerbaijan’s state-sponsored historical revisionism to competing “narratives,” diplomacy risks becoming a vehicle for the very manipulation the EU seeks to counter.
Read more
Anchored in the Past: History as Ontological Resource in Armenian Collective Memory
Why do nations cling to familiar stories, even when reality changes? Drawing on theories of ontological security, collective memory and political psychology, Andronik Papyan examines how the loss of Artsakh unsettled Armenia's national narrative and the search for a new sense of collective identity.
Read morePodcast
EAEU Warns Armenia Over Leaving the Bloc
In EVN Report's news roundup for the week of October 23: Eurasian Economic Union members—Russia, Belarus, Kazakhstan and Kyrgyzstan—warn Armenia of the consequences of leaving the bloc; parliament begins ratification process of the U.S. framework agreement on TRIPP; pressure on opposition candidates and alleged vote-buying raise concerns in municipal elections and more.
Read more
Three Days of Journalism, Ideas and Exchange
From investigative journalism and storytelling to digital security, information integrity and democracy, the fourth EVN Report Media Festival brings journalists, writers, researchers, artists and media practitioners from Armenia and abroad to Yerevan.
Over three days, the festival will feature conversations with leading voices, live podcast recordings, screenings, readings and opportunities for audiences to engage directly with the people reporting on and shaping the media landscape today.
More guests and collaborations will be announced soon. For now, expect three days of ideas, conversation, curiosity and a few unexpected encounters.
