ASTANA – Kazakhstan can turn its strengths in energy, mining, transport and industrial operations into an AI competitive advantage by investing in infrastructure, encouraging local talent and supporting entrepreneurs. In an interview with The AT, NVIDIA Vice President of Physical AI Simulation Rev Lebaredian said access to intelligence will influence where companies are built, where research happens and how productive an economy can become.
President Tokayev holds meeting with NVIDIA VP Lebaredian in Astana on October 1. Photo credit: Akorda.
During his visit to Kazakhstan this October, Lebaredian held talks with President Kassym-Jomart Tokayev about the company’s plans for Kazakhstan’s AI infrastructure development, as the nation has already launched two supercomputer clusters based on NVIDIA technologies.
The country has emerged as the second market for US-based AI infrastructure company Firebird and NVIDIA, which are developing a computing facility in Ekibastuz. The project envisages developing computing infrastructure with a capacity of approximately 300 megawatt, with the potential to expand to 1 gigawatt.
Rev Lebaredian.
In a Q&A interview below, Lebaredian discussed the infrastructure needed to support the transition, the opportunities AI agents and physical AI offer local businesses and why access to computing, practical education and opportunities for young entrepreneurs to solve real-world problems will be critical to building a sustainable AI ecosystem.
Kazakhstan is developing its AI ecosystem, from public-sector digitalization to private-sector innovation. From NVIDIA’s perspective, what are the country’s key strengths today, and what capabilities could be strengthened further as Kazakhstan develops its AI ecosystem?
Kazakhstan is thinking about AI at the level of national development. President Tokayev’s vision recognizes that this technology will shape the economy and the opportunities available to the next generation.
The country has energy resources, substantial industries and people who understand difficult physical problems. Mining, energy, manufacturing and logistics are all places where AI can create tremendous value.
The next step is to connect those strengths. Computing infrastructure should be accessible to entrepreneurs, researchers and students. Industrial companies should become customers and development partners for local teams. Universities should give people opportunities to build with technology.
I would put particular emphasis on entrepreneurship. A country captures more of the opportunity when its people create the products and companies that others use. Kazakhstan has an opportunity to develop that capability and serve markets far beyond its borders.
AI infrastructure is becoming a critical factor in a country’s ability to develop and deploy AI. What infrastructure does Kazakhstan need to build or strengthen — including computing capacity, data centers and networking — to support the next stage of its AI development?
AI is becoming essential infrastructure. Access to intelligence will influence where companies are built, where research happens and how productive an economy can become.
You have to think about the whole system: energy, chips, infrastructure, models and applications. Power and cooling enable the computers to operate. Networking connects them into an AI factory. Software and models turn that capacity into something people can use.
An AI factory produces intelligence. That intelligence can power agents doing useful work across an economy—helping design products, develop software, serve customers and operate industrial systems. Infrastructure investment creates the foundation for that productive activity.
Kazakhstan also needs people who can design, operate and continually improve these systems. And it needs access models that put the capacity within reach of small teams. The value of the infrastructure grows with what people build on top of it.
Beyond infrastructure, what opportunities do you see for Kazakhstan’s local technology companies, startups and researchers to develop AI solutions of their own? Are there particular sectors where Kazakhstan could build distinctive capabilities?
Start with problems Kazakhstan understands deeply. The country’s experience in energy, mining, transport and industrial operations is a source of valuable knowledge.
A team that understands how a mine works could develop intelligence for autonomous equipment. Researchers working on materials or energy systems could build AI tools that help engineers explore designs much faster. A logistics company could turn its operational knowledge into software used across many markets.
AI agents expand who can participate. People with deep knowledge of an industry will increasingly be able to direct agents to build software, construct simulations and test ideas. Their expertise becomes much more powerful when they can translate it into a working system.
That opens the door to a new generation of entrepreneurs. The ambition should be to solve important problems in Kazakhstan and build companies that serve the world.
Your work at NVIDIA focuses heavily on Omniverse and simulation technologies. How could digital twins, simulation and physical AI be applied in Kazakhstan — for example, across industry, energy, mining, logistics, urban planning or other areas of the economy?
Simulators are time machines. They let us explore possible futures before committing to one in the physical world.
You can test how a factory will operate before building it. You can explore changes to a mine without interrupting production. You can train an autonomous machine through situations it might take years to encounter in reality.
Physical AI depends on that capability. Robots and autonomous systems need a place to learn, practice and be tested. We need to understand how they will behave before putting them around people and valuable equipment.
Until now, building these simulated worlds has required specialized skills and a great deal of manual work. AI agents are beginning to remove that barrier. With simulation libraries and the right tools, agents can help write the software, construct environments and run experiments.
The direction is clear: everyone should be able to build and use a simulation. Someone who understands a physical problem should be able to describe what they want to investigate and work with an agent to test it. The results still need to be validated against reality. But the ability to ask and explore those questions can reach far more people.
That is a profound opportunity for a country with Kazakhstan’s industrial base.
As Kazakhstan looks to strengthen its role in the regional AI ecosystem, what capabilities would help it become a place where developers and companies build AI products and services, rather than simply consume AI technologies?
Build a culture of creation, and give people the means to act on it.
Computing access, strong education, open models, software tools and capital all matter. So do customers willing to work with a young company and institutions willing to let people experiment.
Education should give people the confidence and ability to create. Put powerful tools in students’ hands. Have them build things, test ideas and solve real problems. Connect them with engineers and companies so they can see where their knowledge can take them.
At NVIDIA, we build platforms that others can build on. AI agents are extending that opportunity to far more people. Someone with deep knowledge of an industry will increasingly be able to turn an idea into software or a simulation without having to master every underlying technology.
Kazakhstan can develop a generation of entrepreneurs around that capability. The ambition should be that a young person in Kazakhstan can imagine something important, build it from home and bring it to the world.
