The rapid acceleration of artificial intelligence threatens—or promises—to transform high-performance data centers into one of the largest new electricity consumers in Kazakhstan. Speaking at the AI & Digital Bridge 2026 forum in Astana, Kazakhtelecom CEO Bagdat Mussin outlined a strategic scenario where more than 5 GW of additional generation capacity could be directed toward compute infrastructure. Concurrently, Freedom Holding Corp. founder Timur Turlov stated that enterprise demand for Phase 1 capacity of Freedom’s GPU data center already exceeds available supply by 200%, according to reporting by Digital Business.
These pronouncements illustrate the grand scale of Kazakhstan’s digital ambitions. Yet between client demand, server room construction, and the delivery of firm, reliable megawatts lies a formidable engineering and infrastructural challenge for the national power grid.
Demystifying the 5 GW Projection
According to Digital Business, Mussin framed the expansion of AI data centers as an integral driver of Kazakhstan’s future grid expansion. He spoke of more than 5 GW of potential capacity that Kazakhstan could eventually channel toward computational clusters and high-density data centers. Separately, the head of Kazakhtelecom expressed optimism that operational compute consumption could reach 1 GW within several years.
Consequently, the 5 GW figure should not be interpreted as formal grid connection applications or an approved consumption forecast. Mussin’s remarks did not specify which generation sources will supply this power, what capital investments are required, or what timeline will govern implementation.
Nevertheless, the magnitude illustrates the transformative scale under discussion. A gigawatt is a unit of instantaneous power. To understand the annual energy footprint, one must assess operating hours, load factors, and continuous duty cycles.
The Scale: Equivalent to Over a Third of National Power Consumption
Official data from national grid operator KEGOC indicates that in 2025, Kazakhstan consumed 124.6 billion kWh of electricity while generating 123.1 billion kWh. The national power grid reached a historical peak winter load of 17,724 MW.
By comparison, 5 GW equals 5,000 MW—equivalent to roughly 28% of the country’s entire historic peak electricity demand.
If hypothetical compute facilities were to draw 5 GW continuously across all 8,760 hours of a calendar year, total electricity consumption would equal 43.8 billion kWh. That represents approximately 35% of Kazakhstan’s total electricity consumption in 2025.
This calculation illustrates structural scale rather than a short-term forecast. It assumes constant 100% capacity utilization; real-world consumption will depend on phased facility commissioning, actual compute workload profiles, and whether stated figures include cooling systems and auxiliary power usage effectiveness (PUE) overhead.
Nonetheless, the figures demonstrate that AI computing represents an entirely new industrial sector with a systemic impact on Kazakhstan’s national energy balance.
Interpreting Turlov’s 200% Demand Figure
As reported by Digital Business, Timur Turlov estimated current market demand for Phase 1 of Freedom’s NVIDIA-powered data center at 200% of deployed capacity. He also disclosed that major international enterprise partners have expressed readiness to contract an upcoming 100 MW facility for five years at premium pricing.
These metrics describe distinct operational realities. The statement does not imply that firm take-or-pay agreements totaling 200 MW have already been executed for a 100 MW facility. Specific commercial commitments, escrow structures, and Phase 1 capacity volumes remain confidential.
Notably, Turlov acknowledged that the entire 100 MW cluster—initially conceived for international compute export—might ultimately be fully consumed by domestic demand within Kazakhstan. This assessment highlights strong local enterprise and GovTech appetite for accelerated computing.
For power grid planners, the crucial next step is converting commercial interest into a binding, geographically mapped connection schedule: determining precisely which facilities, in which regions, and on what dates will energize high-voltage connections.
Compute Export Requires New Energy Infrastructure
Mussin outlines an economic model where Kazakhstan leverages its energy resources to manufacture digital compute services for export across international markets. A vital prerequisite he highlighted is the aggressive expansion of cross-border fiber-optic connectivity.
The macroeconomic logic is compelling: hosting high-value digital infrastructure domestically to perform advanced computational tasks for global clients. Yet evaluating this model requires rigorous accounting of capital expenditure on hardware, baseload generation, high-voltage transmission, and cooling efficiency.
For Kazakhstan’s power system, the source of generation is paramount. New load must be supported by dedicated new generation, available spinning reserves, and reinforced transmission corridors. Balancing coal baseload, renewable energy, and gas turbine peaking plants will be critical.
A severe systemic risk arises if data center facilities are energized before requisite power infrastructure is commissioned. Long-term commitments to computational capacity must therefore be paired with equally binding energy supply agreements and grid reinforcement projects.
Global Context: How AI Is Reshaping Global Electricity Demand
Kazakhstan’s ambitions unfold against an unprecedented global surge in compute power. In its landmark Energy and AI report published in April 2025, the International Energy Agency (IEA) projected that global electricity consumption by data centers would more than double by 2030, reaching approximately 945 TWh.
Artificial intelligence was identified as the primary catalyst of this expansion. The IEA underscored significant forecasting uncertainties, noting that final consumption will depend heavily on semiconductor energy efficiency, model optimization, and grid transmission bottlenecks worldwide.
This global trend does not guarantee that capital will automatically flow to Kazakhstan. Rather, it highlights intense global competition among sovereign states, energy utilities, and tech giants to secure stable, low-cost megawatts for compute infrastructure.
Measuring Kazakhstan’s Return on Every Megawatt
Evaluating future AI compute projects requires assessing several interconnected metrics: verified energized capacity, contractual take-or-pay load, generation sources, and tangible economic return for the domestic economy.
Critical policy questions must be resolved: Who finances substation upgrades and grid reserves? What portion of compute capacity is guaranteed at preferential rates for Kazakhstani AI startups, university researchers, and GovTech models?
If multi-gigawatt declarations materialize into operational facilities, data centers will emerge as a defining pillar of national industrial and energy policy. Kazakhstan’s competitiveness in the global AI landscape will depend not merely on silicon chips and fiber cables, but on its capacity to supply them with reliable, sustainable electric power.
