Central Asia is positioning itself as a critical frontier in global AI infrastructure development, with two major data-center initiatives now underway that could reshape how companies access high-performance computing resources in Asia.

According to AI Weekly, Saudi-backed DataVolt is constructing TAS-1, a facility located within Tashkent's IT Park in Uzbekistan, with the first phase expected to reach operational status by the end of the year. The initial 6-megawatt deployment of this Tier-3 data center is substantially pre-leased, demonstrating significant market demand from regional operators and international firms seeking alternatives to congested markets in Western Asia and Europe.

In parallel, a separate 125-megawatt complex backed by Nvidia is under development in Kazakhstan with a completion target of 2027. The scale difference between these projects underscores two distinct market strategies: DataVolt's near-term, modular approach versus a longer-term, enterprise-scale build by the semiconductor giant.

Why Geography Matters for AI

The emergence of these facilities addresses a critical infrastructure gap in the AI economy. Data centers serving machine learning training, inference, and deployment have become geographically concentrated in North America, Western Europe, and select East Asian hubs. This concentration creates latency challenges, regulatory friction, and cost pressures for enterprises operating across Central Asia, the Middle East, and South Asia.

Uzbekistan and Kazakhstan offer distinct advantages for compute infrastructure: abundant hydroelectric power resources, strategic positioning on cross-border connectivity routes, and government incentives designed to attract technology investment. The presence of established IT parks in Tashkent signals institutional readiness to support data-intensive operations.

Capacity and Specifications

  • TAS-1 Phase One: 6 megawatts, Tier-3 certification, pre-leased capacity
  • TAS-1 Full Build: 12 megawatts planned
  • Kazakhstan Nvidia Project: 125 megawatts, 2027 target date

The Tier-3 designation for TAS-1 indicates infrastructure designed for mission-critical applications with redundancy systems and high availability targets. This standard is appropriate for AI workloads requiring consistent uptime for model serving and continuous training pipelines.

Market Implications

These projects reflect broader trends in AI infrastructure decentralization. Cloud providers and chip manufacturers are recognizing that concentrating GPU capacity in traditional tech hubs creates supply bottlenecks and pricing power for incumbent operators. Regional facilities reduce dependency on oversubscribed data-center markets while offering lower latency for localized AI services.

The involvement of both private capital (DataVolt with Saudi backing) and major hardware vendors (Nvidia) suggests confidence in Central Asian growth trajectories. Companies developing large language models, computer vision systems, and predictive analytics tools will benefit from distributed infrastructure that brings compute closer to end users and data sources across Asia.

For regional governments, these investments represent validation of their technology infrastructure strategies and potential catalysts for broader digital economy development. Uzbekistan and Kazakhstan are positioning themselves as not merely consumers of AI technology but as participants in the foundational infrastructure layer supporting global AI deployment.