Kazakhstan expands AI infrastructure with thousands of new GPUs

Scaling the infrastructure layer

Kazakhstan is moving from the planning phase of its artificial intelligence strategy into active deployment. This shift is happening across the entire AI stack, which includes the physical hardware, the underlying models, and the final applications used by the public. The country is currently increasing its supply of accelerated computing to meet growing demand from various sectors.

Kazakhtelecom, which serves as the nation’s largest telecommunications company, is leading the hardware expansion. The company has already installed 400 NVIDIA H200 and 480 NVIDIA B300 GPUs in the cities of Astana and Almaty. These high-performance chips provide the necessary power for complex data processing and machine learning tasks.

The expansion will continue into next year with even larger deployments. Kazakhtelecom plans to install more than 1,000 GPUs in the Data Center Valley. Additionally, the company intends to place more than 1,000 NVIDIA Vera Rubin GPUs in the Ekibastuz region to serve diverse workloads for institutions and enterprises.

Empowering local AI development

The increase in local compute capacity is designed to give domestic developers more room to work. Having these resources nearby allows companies to train, fine-tune, and serve increasingly capable models without relying solely on external infrastructure. This creates a foundation where local startups and universities can experiment with AI more freely.

By providing access to these resources at scale, the nation aims to turn raw computing power into useful applications. The goal is to connect the hardware with the people who build the software, including researchers and industrial players. This connection is what ultimately transforms accelerated computing into economic value for the country.

As successful applications emerge, they naturally create a feedback loop of demand for more compute. The current announcements represent a strategic effort to build the AI foundation at a new scale. This infrastructure is intended to support a wide range of needs, from basic data processing to advanced model development.

Testing autonomous transport

One of the most visible applications of this new infrastructure is the exploration of robotaxi services. The companies inDrive and Pony.ai are collaborating to test these autonomous vehicles in Kazakhstan. They have identified Astana as the first city for this deployment, marking a significant step for the nation's transport technology.

The project will assess how feasible it is to integrate Pony.ai’s seventh-generation robotaxis into the existing inDrive app. This integration relies on specific hardware and software systems provided by NVIDIA. The setup includes the Hyperion sensor and compute platform as well as the Halos safety system to ensure the vehicles can navigate safely.

These partnerships demonstrate how AI-native companies are using the new compute capacity to build real-world products. By testing robotaxis, the companies can evaluate the technology in a live environment. This move highlights the transition from theoretical AI capabilities to practical, everyday services for the public.

A multi-layered AI strategy

The developments in Kazakhstan are not defined by the number of GPUs alone. Instead, the strategy focuses on the different layers of the AI stack. This includes the infrastructure layer, the access to compute for developers, and the actual development of AI-native products that people can use.

By addressing all these layers simultaneously, the nation is attempting to build a comprehensive ecosystem. This approach ensures that the hardware being deployed has a clear path toward practical use cases. It connects the physical chips in the data centers to the software being written by local engineers.

The ongoing expansion suggests a long-term commitment to becoming a hub for AI activity. As the infrastructure reaches new scales, the focus will likely shift toward how these tools can be integrated into various industries. These announcements serve as a roadmap for the nation's technological future.