VAST Data, the AI Operating System company, announced a partnership with SK Telecom (SKT), South Korea’s leading telecommunications provider that’s transforming itself into an AI company. To deploy its advanced AI platform on SKT’s new sovereign AI infrastructure powered by the latest NVIDIA Blackwell GPUs.
Named the Haein Cluster, this AI infrastructure, designed initially for GPUaaS has been selected for the Ministry of Science and ICT’s “AI Computing Resource Utilization Enhancement (GPU Rental Support) Program” in Korea, and will be actively used for the development of national AI foundation models.
By combining SKT’s advanced network and data center infrastructure with the VAST AI OS as a key component of its sovereign AI cloud, SKT is creating a fully virtualized GPU-as-a-Service (GPUaaS) environment that meets the growing national demand for AI model development and deployment, all within South Korea’s borders.
Traditional AI infrastructure built on bare-metal environments often requires days or even weeks to provision new workloads, a major bottleneck for developers and enterprises. SKT has eliminated this delay by deploying a virtualized AI infrastructure stack – Petasus AI Cloud - now integrated with VAST Data which can spin up GPU environments in as little as 10 minutes, all while preserving near bare-metal performance.
The Petasus AI Cloud architecture combines VAST’s disaggregated, shared-everything (DASE) platform on NVIDIA HGX server architecture built by Supermicro, enabling a high-throughput, secure, and multi-tenant environment that supports both AI training and inference at scale.
By integrating directly with SKT’s Petasus AI Cloud, the VAST AI Operating System provides:
Sovereign-Grade Infrastructure Control: Enables AI services entirely within national borders.
Virtualization Without Compromise: GPU environments provision in under 10 minutes, matching bare-metal performance.
Secure, Multi-Tenant AI Services: Isolates workloads with data privacy and performance guarantees.
Unified Pipeline for Training and Inference: Simplifies the full lifecycle of AI development and deployment.
Protocol Flexibility, Zero Friction: Native multi-protocol access with no need for client-side gateways or proprietary bloat.
Enterprise-Grade Uptime with Lean Ops: Supports carrier-grade resilience with minimal overhead.
Elastic GPUaaS at National Scale: Designed to meet growing demand from public and private sector AI innovators.
Dynamic Resource Allocation: Automatically provisions and isolates GPU and storage resources, including their networking fabrics, to match each tenant’s specific requirements.
By combining NVIDIA accelerated computing platforms with VAST’s AI Operating System, SKT has built its Petasus AI Cloud as an elastic, sovereign infrastructure powered by next-generation superchips such as NVIDIA Blackwell. This deployment demonstrates how telecom providers are building sovereign AI infrastructure for nations that balance speed, compliance, control, and cost, with VAST’s AI OS at the heart of this transformation.
