Kazakhstan is moving from AI infrastructure planning to deployment as the country increases computing capacity for developers, enterprises and institutions. The expansion spans data center buildouts, planned next-year GPU installations and new AI-native applications, including a proposed robotaxi rollout in Astana.
Highlights
- Kazakhtelecom deployed 400 NVIDIA H200 and 480 NVIDIA B300 GPUs in Astana and Almaty, signaling accelerated AI compute expansion.
- In 2025, Kazakhtelecom plans to add over 1,000 GPUs in Data Center Valley and more than 1,000 NVIDIA Vera Rubin GPUs in Ekibastuz.
- inDrive and Pony.ai partner to assess robotaxi services in Astana using NVIDIA AI infrastructure, boosting AI-native industry applications in Kazakhstan.
Compute deployment plan gains scale
As reported by NVIDIA, citing Kazakhtelecom, the company says it has already deployed 400 NVIDIA H200 and 480 NVIDIA B300 GPUs in Astana and Almaty.Next year, the telecommunications group plans to add more than 1,000 GPUs in the Data Center Valley, alongside more than 1,000 NVIDIA Vera Rubin GPUs in the Ekibastuz region. The infrastructure is intended to support a range of workloads that enterprises, developers and institutions can use at scale.
The broader shift is visible across the AI stack, from infrastructure to models and applications. The latest announcements point to both a rising supply of accelerated computing in Kazakhstan and growing demand for additional capacity as the ecosystem develops.
AI-native demand broadens industry impact
Local AI-native companies are also expanding their computing capacity and the applications built on top of it, reinforcing the link between available infrastructure and commercial use cases.inDrive and Pony.ai are collaborating to explore robotaxi services in Kazakhstan, with Astana set as the first deployment city. The companies are assessing the feasibility of bringing Pony.ai's seventh-generation robotaxis to the inDrive app, supported by NVIDIA AI infrastructure and built on the NVIDIA Hyperion sensor and compute platform with the NVIDIA Halos safety system.
More domestic capacity gives developers and companies room to train, fine-tune and serve more capable models. Connecting that infrastructure with universities, startups and established industries could help Kazakhstan turn higher GPU availability into broader economic and technological applications.
Our earlier article on the U.S. AI data center boom highlighted how rapid buildouts are colliding with higher financing costs, construction delays, and tightening power availability. We noted that a significant share of planned projects has been pushed back or questioned as credible, with weaker investor confidence adding pressure on developers to deliver on time and within budget.