The platform hosts a variety of prominent models, including Kimi, GLM, Llama, GPT-OSS, Gemma, and Qwen, all powered by Zettabyte’s NVIDIA H100, H200, B200, and B300 hardware. Pricing for H100 units starts at US$1.99 per hour, supported by a 99.5 percent uptime service level agreement. For organizations requiring dedicated resources, the company guarantees deployment of reserved clusters within six hours of confirmation.
Zettabyte Expands zCLOUD With Managed Model-as-a-Service
From its Taipei headquarters, Zettabyte has launched a Model-as-a-Service platform on its zCLOUD infrastructure, allowing developers to deploy open-weight and open-source models via a unified API. The service leverages a high-performance fleet of NVIDIA GPUs to simplify AI development for teams seeking to bypass manual model installation and tuning.
Dr. David Ku, Zettabyte’s Technology Leader, emphasized the need for agile, resilient infrastructure that bridges the gap between hyperscale data centers and edge computing. The service offers three procurement paths: short-term instances without long-term commitments, reserved clusters for sustained training workloads, and private cloud deployments tailored to specific security or regional compliance needs. By integrating these models into its zSUITE full-stack environment, Zettabyte aims to streamline the management of complex AI infrastructure.
Comments (0)
No comments yet. Be the first!