Thaki Cloud, an enterprise AI infrastructure and platform company, has announced the launch of Thaki NeoCloud OS, a platform designed to convert physical AI resources into public cloud services. The solution targets organizations that already own GPU servers, storage, networks, and data center facilities but lack the operational layer needed for commercial cloud delivery. By addressing this gap, Thaki Cloud aims to help infrastructure operators turn their assets into self-service, on-demand cloud offerings.
Closing the Gap Between AI Infrastructure and Cloud Services
Many enterprises and infrastructure providers continue to build large GPU fleets and AI factories, yet operating these assets as commercial services requires a separate set of capabilities. Thaki NeoCloud OS provides customer and tenant management, resource provisioning, GPU scheduling, network and storage configuration, usage metering, billing, SLA management, monitoring, and resource recovery. These functions allow operators to isolate customer environments, allocate resources automatically, and manage service levels and the full resource lifecycle.
Service Catalog and API-Driven Delivery
The operating system includes a service catalog, user and organization management, quota and policy controls, capacity management, and API-based integration with external systems. These features enable providers to publish GPU resources as standardized products that customers can order through web consoles and APIs. The platform also supports automatic resource reclamation and observability, allowing operators to maintain visibility and control across tenant environments.
Converting Complex Infrastructure into Usable Services
Thaki NeoCloud OS also focuses on the end user experience by abstracting the complexity of server allocation, network setup, storage configuration, and operating system preparation. Users can select the required GPU and compute resources and begin working quickly without understanding the underlying infrastructure. This approach helps operators move beyond simple GPU rental toward packaged, on-demand cloud services that support AI developers and enterprise customers.
Platform for Multiple Infrastructure Operators
The platform is designed for telecom companies, data center operators, AI factory builders, and GPU infrastructure providers that want to launch their own branded NeoCloud services. These organizations can use their existing assets to create AI cloud offerings, expand GPU hosting into on-demand GPU cloud businesses, or turn a completed AI factory into a multi-tenant service. Operators retain control over their brand, service policies, pricing, and infrastructure configuration.
AI Factories Require a Service Layer
Building an AI factory is only the first step, and Thaki Cloud frames cloud service conversion as a three-stage process. The stages include the physical infrastructure, the Thaki NeoCloud OS software layer, and the final customer-facing NeoCloud service with web consoles and APIs. By sitting between infrastructure and commercial operations, the platform enables providers to measure usage, enforce policies, and bill customers continuously.
Executive Perspective
Thaki Cloud CEO Kang So-ryun emphasized that acquiring GPUs and building AI factories are different challenges from creating cloud services. He noted that customers need resources to be productized, delivered automatically, measured, and billed through a cloud platform. According to Kang, Thaki NeoCloud OS supports telecom companies, data center operators, and AI factory businesses in building and operating their own NeoCloud offerings.
Thaki Cloud was established in February 2025 and is headquartered in Seoul's Gangnam district, with a focus on NeoCloud construction and private AI cloud environments. Through Thaki NeoCloud OS, Velox, and Telox, the company supports the operation and service delivery of GPU and AI factory infrastructure. The launch positions Thaki Cloud as a software provider for organizations seeking to move from AI infrastructure ownership to sustained cloud-based service operations.