 |
zCLOUD delivers Model-as-a-Service (MaaS), hosting open-weight and open-source models such as Kimi, GLM, Llama, GPT-OSS, Gemma, and Qwen on Zettabyte's GPUs and making them available through a single API.
- Offers open-weight and open-source models through an API, running on Zettabyte's GPUs;
- Provides access to NVIDIA H100, H200, B200 and B300 GPUs;
- Offers three fully managed ways to access capacity, backed by a 99.5 percent uptime SLA target;
TAIPEI CITY, Taiwan, Aug. 13, 2026 /PRNewswire/ -- Zettabyte, a global AI computing company, today announced the general availability of Model-as-a-Service on zCLOUD, its GPU cloud. Teams can efficiently call open-source models through API on Zettabyte's NVIDIA H100, H200, B200 and B300 GPUs.
Model-as-a-Service
With Model-as-a-Service, zCLOUD hosts popular open weight and open-source models, ready to call through an API and run on Zettabyte's GPUs. Teams can build on them directly, without sourcing, installing, and tuning a model themselves. Teams that want raw capacity can still rent GPUs on zCLOUD and run their own models.
GPU Availability
NVIDIA H100 is available today, with pricing from US$1.99 per GPU-hour depending on needs and scale. zCLOUD is backed by a 99.5 percent uptime SLA target, and reserved clusters are built to deploy in under six hours from confirmed requirements.
Ways to Access GPU Capacity
Zettabyte offers three ways to access capacity:
- Short-Term Capacity - GPU instances that scale to hundreds of GPUs, with no long-term commitment.
- Reserved Clusters - Predictable, fully managed capacity for sustained training and inference under a term commitment.
- Private Cloud - Custom deployments for teams with procurement, security, or region requirements.
"AI computing will be everywhere, from hyperscale data centers to the edge. The supporting AI infrastructure and software must be compatible, resilient, and agile. zCLOUD was built to deliver Model-as-a-Service, bringing more than 40 hardware providers together and managing the underlying AI infrastructure through our zSUITE full stack," said Dr. David Ku, Technology Leader at Zettabyte, former Corporate Vice President and CTO of AI at Microsoft.
Model-as-a-Service on zCLOUD is generally available today at www.zettabytecloud.com. （Related： Bakery China-Powered Bakery ASEAN to Debut in August, Showcasing Innovation in the Bakery and Confectionery Industry ｜ Latest ）
About Zettabyte
Zettabyte is a global AI computing company focused on delivering and measuring high-quality compute. Through its zSUITE platform, the company enables organizations to improve reliability, readiness, observability, cost transparency, and energy efficiency across AI infrastructure environments. Learn more at www.zettabyte.space.
