Substantial technological infrastructure is being developed in Armenia – NVIDIA Vice President on Armenia’s AI capacity
6 minute read
Armenia has everything it needs to take advantage of the limitless opportunities offered by the new technological revolution and to be represented at every layer of it.
NVIDIA Vice President Rev Lebaredian made the remarks online during a press conference dedicated to the Digitec 2026 technology exhibition, which will take place in November.
In his assessment, Armenia now has an opportunity that may come once in a lifetime, or even once in a century.
“We are now seeing unprecedented demand for the production of artificial intelligence. That is precisely why we are seeing AI factories being built everywhere, and even that is still not enough. For many people today, artificial intelligence may simply seem like a chatbot on a web page. But artificial intelligence is much more than that. To create artificial intelligence, you need a strong foundation,” Lebaredian said, broadly dividing the AI ecosystem into five main layers: energy, chips, computing infrastructure, models and applications.
The first layer is energy. Training and running AI systems requires large-scale computing power, which ultimately depends on electricity. According to Lebaredian, expanding energy production and the corresponding infrastructure is therefore an important task for Armenia, as the amount of available energy capacity determines the scale of computing that can be carried out.
The second layer is chips and semiconductor technologies. In Lebaredian’s assessment, Armenia occupies a distinctive position in this area relative to its size, given the presence of international technology companies operating in the country and its professional expertise.
Lebaredian pointed to NVIDIA, Synopsys, Siemens, Cadence, AMD, Broadcom and Cisco as major companies connected with Armenia’s technology ecosystem. In this context, the focus is primarily on semiconductor design, engineering, software and research capabilities.
The third layer is computing infrastructure. This includes AI factories and high-performance data centres, where large numbers of GPUs, along with networking, storage, power and cooling systems, are brought together into a single integrated infrastructure. Their purpose is to provide the computing power needed to train, refine and deploy AI models at scale.
Large-scale infrastructure is already being developed in Armenia in this area. The Eleveight AI factory, operating in Gagarin since June, runs on 512 NVIDIA Blackwell B300 GPUs. According to the company’s current figures, the first-phase investment programme amounts to up to $150 million, while the centre has a capacity of 5 MW. The next expansion is expected to increase this to 40 MW.
The first phase of Firebird’s AI factory, which opened in Hrazdan in August, includes 6,144 NVIDIA B200 GPUs and has a capacity of 15 MW. Under plans published by the company and NVIDIA, the infrastructure in Armenia is expected to be expanded significantly, reaching more than 70,000 NVIDIA Blackwell and Rubin GPUs and a capacity of 300 MW by the end of 2027.
The fourth layer is AI models – what people generally think of first when they hear artificial intelligence – systems such as GPT, Claude and other leading models.
According to Lebaredian’s reasoning, such models can be created and operated only when the preceding layers are in place: sufficient energy, the appropriate chips and large-scale computing infrastructure. The AI factories being developed in Armenia create opportunities to train, refine and run AI models locally, provided the necessary computing infrastructure is available.
The fifth and, in Lebaredian’s assessment, most important layer is applications. This is where AI’s technological capabilities are transformed into specific products and services for business, science, education, public administration, industry and other sectors.
According to Lebaredian, Armenia has a more established foundation in this area because software and technology solutions have been developed in the country for decades. In the context of this development, he also highlighted Digitec, which is being held for the 21st time this year.
At the same time, the NVIDIA Vice President said the opportunity was so significant that Armenia could not realise it in isolation. According to him, the next step is to build regional connections and partnerships that would make it possible to jointly create a technology ecosystem larger than Armenia’s own capabilities.
In this context, he attributed an important role to Digitec. This year, the event is being transformed into a conference dedicated to artificial intelligence, designed to encourage discussion and cooperation across all the layers of AI outlined above.
Lebaredian also takes a broader view of the term “region” than the South Caucasus alone, encompassing Europe, the Middle East and Central Asia. In his assessment, with such broad cooperation, Armenia could become part of a larger AI ecosystem and connect its technological capabilities with international markets.