Servercore Expands GPU Server Lineup in Central Asia
Servercore, an international IT infrastructure provider, has expanded its lineup of cloud GPU servers in data centers across Uzbekistan and Kazakhstan, giving local companies access to next-generation graphics processors and flagship enterprise AI accelerators. The provider’s dedicated and cloud server lineup now covers the entire AI development cycle, from lightweight inference to training massive models (up to 120 billion parameters) and complex computational tasks.
Servercore is expanding its GPU server offerings in response to the growing demand from local companies for AI computing resources. Each solution in the lineup is optimized for specific workload requirements and covers a wide range of use cases. The NVIDIA® RTX A5000 is ideal for inferencing smaller language models, running RAG, and generating images. NVIDIA® A100 graphics cards are designed for more resource-intensive scenarios, such as fine-tuning and inferencing large LLMs, processing massive data streams, and conducting research that requires intense computational power. The NVIDIA® RTX PRO 6000 Blackwell is tailored for demanding enterprise tasks, including running LLMs with 70 to 120 billion parameters on a single card, working with agentic systems, multimodal models, and video generation.
The updated server lineup allows companies to keep their infrastructure intact as their computing needs scale. Work can begin on an A5000, with a seamless transition to an A100 or RTX PRO 6000 Blackwell as the project grows and tasks become more complex. By the end of the year, the company also plans to launch servers powered by the NVIDIA® H200. Utilizing higher-performance GPUs accelerates model training and significantly shortens the time-to-market for AI products.
The new cloud GPU servers can be deployed just minutes after ordering, eliminating the need to purchase and maintain expensive in-house equipment. A flexible pay-as-you-go model allows clients to pay only for the actual time resources are used, completely avoiding capital expenditures. Configurations can be customized for specific tasks, ranging from a one-time experiment to continuous computational workloads. This makes high-performance GPUs accessible not only to large corporations but also to startups and small teams.
The servers are located in local data centers, ensuring full compliance with the data protection laws of Uzbekistan and Kazakhstan regarding personal data processing. Servercore’s infrastructure is already being used by local market leaders to develop their AI projects. For example, Cybernet AI migrated to custom cloud servers, which improved their IT system response times by 80% and drastically reduced the time it takes to launch new products. Expanding the GPU lineup will allow even more companies to access state-of-the-art computing resources to build and scale their AI solutions.