ITglobal.com: AI Cloud (GPU Cloud)

Main article: What is IaaS

2025

3x power increase

ITGLOBAL.COM actively develops the GPU infrastructure DPCs Russia in and, Kazakhstan placing on these facilities projects related to: artificial intelligence the deployment of large language models (), LLM infantry and other AI services. This was announced by ITglobal.com on December 22, 2025.

To implement such projects, ITGLOBAL.COM provides GPU power in two formats: in the cloud and on dedicated servers. Of greatest interest to them are Nvidia RTX Pro 6000 Blackwell Server Edition cards - the flagship server solution in the RTX class from Nvidia. The solution is economically attractive for business, and it also provides high performance for AI tasks, including data processing and machine learning.

GPU-based infrastructure solutions remain a key strategic focus for ITGLOBAL.COM, and we continue to build capacity in this segment. We reserve resources in advance and keep equipment available, ensuring the rapid launch of projects, and quickly adapt supply to demand. For example, having noticed a growing interest in NVIDIA RTX Pro 6000 Blackwell Server Edition accelerators for AI and machine learning tasks, we have multiplied the volume of purchases of this equipment, "said Evgeny Svidersky, director of the ITGLOBAL.COM cloud business, ITG Corporation.

Access to Nvidia A100 graphics accelerators

ITGLOBAL.COM expanded the line of affordable video cards in Kazakhstan and provided access to Nvidia A100 graphics accelerators. This type of video card is available on the basis of the Sairam level TIER III data center. This was announced by ITglobal.com on June 20, 2025.

This accelerator is available both as part of the GPU Cloud cloud service and in the configurations of dedicated physical servers, which provides maximum flexibility in choosing an architecture for specific tasks. NVIDIA A100 has high processing power, energy efficiency and the ability to handle large-scale loads, making it the best solution for resource-intensive tasks - from modeling and analytics to training and neural network interference in high-performance environments. In addition, thanks to the support of NVLink technology, it is possible to combine several accelerators into a single high-performance cluster with a total graphics memory of up to 320 GB, which provides scalability and flexibility in building computing solutions for tasks of varying complexity.

Our key goal is to provide the business with maximum flexibility and adaptability in choosing IT solutions at all sites of ITGLOBAL.COM presence, and that is why we are consistently expanding the line of available equipment. Kazakhstan is a strategically important region for us, where technological updates are being implemented as a priority and in an accelerated mode. The emergence of NVIDIA A100 significantly expands the capabilities of the GPU Cloud platform, providing customers with access to enterprise performance without the need to post data outside the country, "said Evgeny Svidersky, director of the ITGLOBAL.COM cloud business, ITG Corporation.

Before the update, along with the GPU Cloud service, NVIDIA A16 and A800 video cards were offered in Kazakhstan, and the configurations of dedicated servers were offered with NVIDIA L40S accelerators. H100, H200 and SOPHGO SC7 HP75 were provided on an individual request, taking into account the customer's requirements.

2024: Year-over-year revenue growth of 70%

ITGLOBAL.COM analyzed the financial results of the provision of services and services in the public cloud in 2024. The AI Cloud service (GPU Cloud) showed noticeable dynamics, revenue from its implementation almost doubled compared to 2023. The company announced this on January 29, 2025.

According to ITGLOBAL.COM, the main customers of GPU Cloud are startups, large retailers, transport companies, financial institutions and design and engineering organizations. GPU Cloud is of particular value to IT partners who deploy cloud power into complex business solutions - from AI-as-a-Service to dedicated environments for high-load projects. The presence of several types of graphics accelerators and the variability of configurations allows you to quickly solve problems related to both artificial intelligence and machine learning, as well as video rendering and processing. In addition, GPU support and framework compatibility reduce solution deployment time, and turnkey rentals save customers the cost of their own hardware, allowing them to focus on developing innovations.

We focused on investing in the development of a cloud rental service with graphic processors ones at the beginning of 2024. The equipment was available, we quickly gave power to customers. In parallel, we began to develop the provision of GPU Cloud according to the partner model. These efforts had a positive impact on revenue. The growth of this segment is another confirmation of the reliability and security of the cloud infrastructure ITGLOBAL.COM. In 2025, we expect an additional increase due to the rental service, supercomputer which provides the ability to use up to 256 GPUs. With this scale and power, customers will be able to accelerate research and development. In this regard, in 2025-2026. GPU will remain a priority area for ITGLOBAL.COM, - said the Eugene Svidersky director of the cloud business of ITGLOBAL.COM, ITG Corporation.

2023: Launch of AI Cloud service in Russia

The company ITGLOBAL.COM began providing its customers Russia AI with the Cloud service, under which the business can rent cloudy servers with powerful GPUs. The company announced this on August 28, 2023.

Five to ten years ago, graphics processors were used mainly for graphics and visualization. With the advent of general-purpose computing on GPUs, GPUs began to be used in other areas, such as AI high-performance computing and. machine learning

The expansion of the use of GPUs was due to their parallel computing capabilities. Unlike CPUs designed to handle sequential tasks, GPUs can handle many tasks in parallel.

When the business grew to complex analytics of its products and customers, building demand models and implementing AI for smart recommendations to consumers, when machine vision and IoT devices were introduced in production, CPU performance became lacking, and parallel computing on the GPU entered the game. Then the issue of purchasing expensive equipment became acute, which is not enough to simply buy, subsequently it must be maintained, repaired and updated. It is important to take into account that GPU models are released annually and are very superior to their predecessors. So that companies can effectively solve their problems, but at the same time save budgets, we launched the AI Cloud service, "said Alina Zhigalova, director of service development ITGLOBAL.COM.

ITGLOBAL.COM 's GPU Cloud Compute is a view public cloud based on VMware where virtual machines I work with graphics video cards Nvidia A800, tensor cores and 80 GB of RAM. For AI Cloud, customers pay according to the Pay as you go model, that is, only for the resources issued by the VM.

Within AI Cloud service to customers well six vGPU configurations on VM: 1vGPU.10GB (1vGPU Instance, 10 GB vGPU RAM), 1vGPU.20GB, 2vGPU.20GB, 3vGPU.40G, 4vGPU.40G, 7vGPU.80GB. Profiles have already been configured, they are easy to select and add when creating a VM. Thus, a business can use the most suitable option for its tasks, - said Alina Zhigalova, director of service development ITGLOBAL.COM.

The cloud provider for August 2023 has eight data centers in seven countries around the world on three continents. AI Cloud service has already been built and implemented in Russia, Kazakhstan and the Netherlands. In data centers, ITGLOBAL.COM installs its own high-end hardware: high-performance vStack-R servers, GPU A800 accelerator NVIDIA and storage NetApp All-Flash.