Armenia has officially committed to replacing its Metsamor nuclear facility with Small Modular Reactor (SMR) technology. The discussion on how sound this direction of travel is has revolved around the maturity of the technology, with the assumption that Yerevan is making a geopolitical decision, opting for an American rather than Russian technological path. There is a broader picture, and to understand the context of Armenia’s choice, Caucasus Watch reached out to an industry expert. Henry Tillman is US-born and the founder of both Asia Investment Research Global (Asia Nuclear) and Ayana Research, a London-based advisory specialising in energy transition in emerging markets, often incubating renewable energy projects with a specific focus on Small Modular Reactor (SMR) technology. He points towards a third technological path that is not yet factored into the calculation of Yerevan’s options, namely Chinese SMRs and Thorium reactors.   

The strongest criticism of U.S. SMRs is that they are still experimental, as there is not a single onshore U.S. SMR plant. We are essentially talking about designs with the prospect of being deployed as operational nuclear plants. You have been discussing the entry into the market of Chinese Thorium SMRs. Are we talking about competitive experiments or actual mature solutions?

Thorium technology has a long history, starting initially in the United States. Responding to criticism of the threat of nuclear waste derived from uranium-enriched spent fuel, U.S. President Eisenhower launched a program called Atoms for Peace in 1953. By 1959, Edward Teller, the Hungarian-American regarded as “the father of the hydrogen bomb,” proposed and subsequently led a project that resulted in a molten salt reactor (MSR) powered by thorium. Thorium is found across about five countries (India, Brazil, Australia, the United States, and China), making it much easier to locate, and is 98% consumed during the generation process, leaving only 2% waste. Unlike conventional reactors, thorium reactors carry no threat of explosion. Furthermore, thorium waste decays within 200 rather than 10,000 years. Thorium was first used in a commercial reactor in 1962. From 1962 to 1972, the NS Savannah operated as the U.S. nuclear-powered cargo and passenger ship, and the programme ran perfectly. From 1965 to 1969, Oak Ridge National Laboratory in Tennessee also operated a molten salt reactor. In 1972, the U.S. Government elected to prioritize plutonium over thorium. Consequently, from 1972 to 2011, this technology was effectively shut down. While discussions on TMSR technology preceded the 2011 disaster, they heavily framed political debate. In 2008, Republican Senator Orrin Hatch and Democratic Senator Harry Reid introduced the Thorium Energy Independence and Security Act, which failed to gain support in both 2008 and 2010, reopening the technological debate. In 2011, the Fukushima disaster occurred. This led to the technology becoming accessible across the globe, and specifically to China, which had been aware of it since the 1950s through Canada’s CANDU thorium reactor programme. China immediately began putting this technology into action, as did other countries, while the USA’s initial public moves were made in 2016. China expects to release its first commercial SMR during 2026. Initial recipients are anticipated to include the BRI countries of Malaysia and Thailand. In April, a Chinese group and the UAE also agreed that an SMR would power the local grid. Meanwhile, Russia’s Rosatom continues to lead in floating SMR units, operating them since 2019, and in June broke ground on Central Asia’s first onshore SMR units.

This is useful context. So, we have two mature SMR technologies currently on the market rather than in a piloting phase. Now, Armenia is the only country that has a nuclear plant in the South Caucasus, reliant on a standard Russian VVER (Water-Water Energetic Reactor) reactor. The Armenian government recently committed to an SMR technological path. How plausible is that for a country with a $9bn budget? Can you spend billions on untested technology?

I'm not a politician, but I focus on data and practical implementation. I just outlined that China and Russia are years ahead. China will be the first country to export commercial land-based SMRs in 2026. Rosatom has been operating floating nuclear power plants and SMRs since 2019, and its geographic reach now extends across ASEAN and many African countries (with a Rwanda agreement announced last week). By 2029–2030, we should expect Thorium-based reactors (TMSRs) to come to market. For the U.S., we are expecting conventional SMRs to enter the market by 2033–2035, although TerraPower claims it will do so by 2031. Kazakhstan is pursuing both Chinese and Russian technologies. In June 2025, Kazakhstan, the world’s largest uranium producer, awarded its first three nuclear power plants to Rosatom and China National Nuclear Corporation (CNNC)—the U.S. never even submitted a bid.

Who pays for the development of these reactors?

Generally, Russia and China try to bring turnkey packages including education/training and financing, but they can't always do that. Because of sanctions, Rosatom has turned to Chinese capital markets through so-called “Panda bonds” (RMB-denominated debt issued in China by foreign governments and corporations). It is reasonable to question Russian financing. In March, Vietnam and Rosatom agreed to build a nuclear power plant. In April, Vietnam signed MoUs with a number of Korean financial institutions regarding nuclear financing.

The fact that the United States is absent from this discussion is a matter of capital or technology. 

Particularly with Thorium, it is because those who moved first were not the initial founders of the technology {the U.S}. Now they {China} have a multi-year technological lead as I noted previously. 

There is a broader context here. Armenia is one of the few countries in Eurasia to have a direct Nvidia investment, building an AI hub the size of Microsoft's, which requires an energy-intensive data center, in part financed by Armenian systemic banks. This is a landlocked country. Can a country with a $9bn government budget invest in the prospect of U.S. SMRs at the moment? Is it a sound choice?

I can't answer the second question. The first part of the question is linked with the fact that the U.S. government under the Trump administration is providing copious amounts of financing to advance these partnerships. So, it's possible to finance it. But there's always a timeline. There's a seven-to-ten-year timeline before a U.S. SMR reaches the market. That's the gap. So, it's a risk for Armenia. If there's a 10-year gap, then I think I would look at something more available in the near term, but I don't run the country. Still, it is definitely possible for the U.S. to finance it and get it all done in seven to ten years.

Financing? Are you talking about a grant? Because we're talking about a pilot SMR, with cost overruns that could inflate the energy cost (the price per kilowatt-hour) of a middle-income economy. If Armenia takes out a loan for that, it could end up paying a fortune.

They can formulate a grant. There are all kinds of grants. China has all kinds of financial instruments in place. The DFC (U.S. International Development Finance Corporation) is now more aggressive and just invested $25 million in a port in Georgia (Poti). The Trump administration has also poured significant capital into the Export-Import Bank (EXIM). So, the U.S. can mobilize capital, both public and private.

So, Armenia could perhaps seek U.S. financing, but would the timeline be dependable? If the Armenian energy mix depends on nuclear power for 35% of its baseload capacity, based on what we know today, can Yerevan decide to commit to an American technological path?

China and Russia are the world leaders and have been world leaders for quite a long time. if I'm running a company or a public agency, I have to look at the reality of the situation. 

Amazon, Google, and the big digital platforms are the main investors in U.S. SMRs. They need the energy for data centers. How was the energy ecosystem allowed to fall so far behind?

The penny dropped a couple of years ago. There's a lot of money chasing energy generation now. That doesn't mean it's going to happen in the near term. Original Equipment Manufacturers (OEMs) are taking matters into their own hands and buying equipment directly wherever they can find it. For the moment, there is no U.S. player in SMRs that can commercially come to market. For the next five to seven years, Russia, China, and/or India will lead the sector, although South Korea and its 2025 partnership with the UAE is a very credible player. It also has trained manpower and mass production capabilities. Based on current technology, the data shows one would select either a Russian or Chinese producer, as Kazakhstan recently did.

And if you are Nvidia, would you trust working with Chinese or Russian energy producers in the same cluster?

Jensen Huang is a very clever person and politically astute. After the Trump-Xi summit in May, he remained in China for several days, mixing with the public. So, I'm a huge supporter of what Jensen has built, and I think he knows exactly what he's doing. But I think he's been put in a difficult position. Nvidia chips were a significant part of the May negotiations—the U.S. offering chips to China which were soon to be out of date, which China rejected. In early June, the U.S. then banned Nvidia chips from being exported to China. By mid-June, Chinese authorities told AI companies in China not to buy them, citing security concerns, and Nvidia stopped manufacturing them—after they represented up to 20% of Nvidia’s revenues. China remains a few years behind the U.S., Taiwan, and South Korea in this space, but is closing the gap. China continues to attract capital in the AI chip space, with over £3 billion raised in H1 by mainland Chinese tech companies. The U.S. sanctions on certain technologies have forced a number of Chinese companies to accelerate innovation, such as Huawei and Ant AI. The Chinese continue to develop this faster. In my opinion, the Chinese continue to leap ahead on energy, and that gives them an advantage in powering future AI.

Would you say that in Armenia, Nvidia has a chance to try a different strategy and work in more integrated clusters, using Chinese and Russian technology for power generation?

I do see commercial logic for partnerships in building energy platforms. China, Russia, and India have been coordinating on nuclear since at least the SCO summit in September 2025. Take the China-UAE multi-partner system announced in April, or the $30 billion BlackRock fund in May. One of my U.S. partners has been advising on U.S. data centers for the past 25 years. There was nothing in this sector before it was launched outside Arlington, Virginia. Now that area represents roughly 80–85% of the U.S. data center ecosystem. The entire country is capturing a segment of the growth in this value chain, in one component or another. So, the answer is this: as the U.S. data center market becomes fully developed, companies want to grow internationally. Our firm has a special unit designed to facilitate this expansion into Asia, Africa, and Latin America.

Interview conducted by Ilya Roubanis for Caucasus Watch