The global demand for AI is growing rapidly, driving a market trend toward subscription services and GPU rental solutions. Nini Wu, CEO of GWS CLOUD, identified several critical factors behind this shift. First, organizations developing AI require significant GPU computing power to support their initiatives. Second, the accelerating global adoption of AI has intensified competition for computing resources, making GPUs a scarce commodity. Many organizations face challenges acquiring GPUs or endure lengthy procurement processes, leading to higher installation costs. These obstacles make rental services an attractive alternative to outright purchases. Additionally, GPU servers have stringent requirements for data center infrastructure, cooling systems, power supply, floor load capacity, and network transmission, adding to the complexity.

To address these challenges, many businesses have adopted GPU rental services to reduce upfront investments. In response to this demand, GWS CLOUD has launched its GWS AI Cloud virtual GPU subscription platform, designed to meet the needs of small and medium-sized enterprises (SMEs).

Expertise in Cloud and Data Centers: Laying the Groundwork for Efficient GPU Services

Nini Wu highlighted that eASPNet, the parent company of GWS CLOUD, has been operating data centers and providing cloud services in Taiwan for over 25 years. Leveraging this extensive experience, the company ensures efficient and stable data center and cloud platform operations. GPU computing services enhance the company’s capabilities by building on these core strengths. Furthermore, eASPNet’s data center serves as one of the largest data exchange hubs in the Asia-Pacific region, providing abundant network resources and bandwidth for reliable GPU computing and AI application services.

The company has also developed a range of cloud services to meet the diverse needs of Taiwan’s education sector and various industries. As a result, GWS AI Cloud not only delivers GPU computing power but also addresses requirements such as robust information security protection.

Scalable GPU Rentals: From 0.25 to Multiple GPUs

GWS AI Cloud leverages NVIDIA H100 GPUs to meet the demanding requirements of large language model (LLM) and machine learning (ML) training. The platform features an intuitive graphical management interface, replacing the traditionally complex command-line operations often associated with AI/ML tools. Preconfigured training and inference environments further simplify setup and reduce usage barriers. For specialized needs, the platform also offers customizable tools and environments.

A key feature of the platform is its flexible GPU allocation. Nini Wu noted that most market offerings rent GPUs by the full card, which can be cost-prohibitive. In contrast, GWS AI Cloud allows rentals starting from 0.25 GPUs, with scalable options including 0.5, 1, 2 GPUs, or more, depending on project requirements. For example, users renting 0.5 GPUs for six months can adjust their usage as needed during the term. If additional computing power is required, users can scale up at any time.