ZT Systems, a leading provider of AI and cloud computing solutions, has unveiled its ACX200 platform featuring the NVIDIA GB200 Grace Blackwell Superchip. The ACX200, powered by NVIDIA Blackwell Tensor Core GPUs and Grace CPUs connected via high-speed NVIDIA NVLink technology, is a liquid-cooled, rack-integrated server designed for accelerated computing. This platform, combining NVIDIA Blackwell GPUs with ZT’s expertise in hyperscale design and deployment, supports exascale AI training and inference workloads.
“The ZT Systems ACX200 enhances our customers’ ability to deliver AI at unprecedented scale with superior performance and energy efficiency,” said Tom Lattin, VP of Platform Engineering at ZT Systems. “This system is tailored to support advanced service providers in deploying next-generation AI applications. With optimized rack- and cluster-level configurations and ZT’s global deployment expertise, customers can quickly implement this powerful technology in their datacenters.”
Kaustubh Sanghani, VP of GPU Product Management at NVIDIA, added, “ZT has developed innovative server solutions featuring the new NVIDIA Blackwell architecture, providing hyperscale customers with groundbreaking performance to tackle complex challenges.”
The ACX200 platform is designed to meet the intensive compute demands of generative AI applications. It includes compute trays with Blackwell GPUs and Grace CPUs, fifth-generation NVIDIA NVLink switch trays, and specialized cable cartridges for high-bandwidth, low-latency connections, enabling the system to function as a unified massive GPU. The platform’s direct-to-chip liquid cooling supports various datacenter cooling configurations.
ZT Systems, in partnership with NVIDIA, brings cutting-edge GPU technology to market swiftly through close customer collaboration, leveraging extensive hyperscale workload experience and a robust supply chain. ZT ensures reliable, high-quality production through controlled manufacturing and unique validation processes, with global delivery and support for rapid deployment and dependable operation throughout the product lifecycle.
