More Cloud Means More Physical Infrastructure
The rapid adoption of cloud services and artificial intelligence workloads has triggered a sharp increase in demand for specialists who handle the physical infrastructure of data centers—including servers, cooling systems, and power supplies. Global hyperscalers are pouring tens of billions of dollars each year into constructing new data centers, yet the market cannot produce enough highly skilled professionals to keep pace. Training a single large language model requires thousands of GPUs, and cooling systems must dissipate hundreds of kilowatts of heat from a single server row. However, a standard data center is not built to handle the thermal load of GPU clusters used for AI.
An hour of unplanned downtime at a data center can cost a large company hundreds of thousands of dollars, and for financial platforms, the figure can reach millions. A notable example in this context is the NCOC Tier III data center project in Atyrau, Kazakhstan. This facility became one of the first large-scale Tier III projects in Western Kazakhstan and underwent the full international certification process from the Uptime Institute—covering design, construction, and operational resilience.
The NCOC project was overseen by the Uptime Institute and Bureau Veritas, with engineering coordination led by Denis Slyusarchik, who brings over 15 years of international experience. He has served as CEO of international divisions for leading system integration companies in Ukraine, Kazakhstan, and Poland. Slyusarchik began his career at Lviv Design Institute (LPI-14) as a low-current network engineer. In 2014, he took the helm at Incom Kazakhstan LLP and also managed Compass Engineering in Ukraine and Compass Engineering Sp. z o.o. in Poland.
“The cloud is someone else’s data center. And a data center is an engineering system where a mistake in power or cooling design can bring thousands of clients to a halt all at once. The technological maturity of the market does not reduce the need for engineering expertise—it raises the bar for its level.”
Denis Slyusarchik
At the time the NCOC project was being built, only a handful of professionals in Central Asia had comparable experience coordinating high-availability infrastructure.
The complexity of a modern infrastructure project goes far beyond the technical side. “You have to simultaneously account for the client’s business requirements, the regulatory environment, equipment availability, timelines, and budget. A specialist who only sees one of these layers will not be able to deliver a workable solution,” Slyusarchik notes. He adds that Central Asia at that time had very few completed projects with comparable fault tolerance and resilient architecture.
Building infrastructure that met the standards commonly applied in developed markets like Europe and North America required adapting to Kazakhstan’s specific conditions. “One of the toughest challenges was tailoring high-availability infrastructure requirements to the operational and logistical realities of industrial Kazakhstan. We had to develop a coordination model that ensured consistency across all critical systems—from power supply to network architecture—while satisfying both international inspectors,” he explains.
The experience gained from the NCOC project underscores that “in high-availability infrastructure, there are no secondary systems. A failure in any component can cascade across the entire facility. That’s why the key was to ensure both the individual functionality of each system and their full integration as a single whole.”
Thanks to initiatives like the NCOC project, Central Asia has an opportunity to set a new benchmark for data center infrastructure—one that aligns with the current demands and challenges of an ever-evolving market.
The development of data center infrastructure in Central Asia, particularly through the NCOC project, reflects the growing need for high-quality data processing solutions. The introduction of new technologies and standards in this market not only improves company efficiency but also creates fresh opportunities for professionals in the field who can meet the rising demand for engineering expertise. This, in turn, will further fuel the growth of cloud services and artificial intelligence, both of which rely on reliable and efficient infrastructure.