HPE Ships First NVIDIA Grace Blackwell System: A Game Changer in AI Infrastructure
1. Introduction
On February 13, 2025, Hewlett Packard Enterprise (NYSE:HPE) announced a significant milestone in the realm of high-performance computing (HPC) and artificial intelligence (AI) with the shipment of its first NVIDIA Blackwell family-based solution, the NVIDIA GB200 NVL72. This development marks a pivotal advancement in HPE’s commitment to delivering cutting-edge technology designed for service providers and large enterprises facing the increasing demands of AI workloads.
2. The NVIDIA GB200 NVL72: A Technological Marvel
The NVIDIA GB200 NVL72 system stands out due to its innovative rack-scale architecture tailored to facilitate the rapid deployment of extensive AI clusters. With a focus on efficiency and performance, the system incorporates advanced direct liquid cooling solutions that are crucial for modern data centers facing escalating power requirements.
Trish Damkroger, senior vice president and general manager of HPC & AI Infrastructure Solutions at HPE, emphasized the urgency for scalability and performance in today’s market. “AI service providers and large enterprise model builders are under tremendous pressure to offer scalability, extreme performance, and fast time-to-deployment,” she stated. HPE’s strategic advantage stems from its experience in developing some of the fastest supercomputers globally, equipped with direct liquid cooling, allowing for lower costs per token in AI model training.
Key Features of the NVIDIA GB200 NVL72
- Architecture and Performance: The system features a shared-memory, low-latency architecture, capable of managing extremely large AI models exceeding one trillion parameters within a single memory space. It integrates NVIDIA CPUs and GPUs, compute and switch trays, as well as networking and software, to deliver exceptional performance tailored for heavily parallelizable workloads like generative AI model training and inference.
- Liquid Cooling Technology: Drawing on five decades of expertise, HPE's direct liquid cooling technology is a cornerstone of the GB200 NVL72, ensuring optimal performance even as data center densities and power requirements increase.
- Memory and Bandwidth: The system boasts an impressive configuration of 72 NVIDIA Blackwell GPUs and 36 NVIDIA Grace CPUs, interconnected via high-speed NVIDIA NVLink, and offers up to 13.5 TB of total HBM3e memory with a staggering 576 TB/sec bandwidth.
3. Collaboration with NVIDIA: A Strategic Partnership
The release of the NVIDIA GB200 NVL72 is the result of a collaborative effort between HPE and NVIDIA, aimed at addressing the evolving needs of engineers, scientists, and researchers. Bob Pette, vice president of enterprise platforms at NVIDIA, highlighted the importance of cutting-edge liquid cooling technology in meeting the increasing power and compute demands of modern applications. This collaboration underscores both companies' commitment to providing scalable solutions for large AI clusters.
4. Industry-Leading Services and Support
In addition to its state-of-the-art hardware, HPE offers comprehensive support services that enhance the deployment and management of AI solutions. The HPC & AI Custom Support Services cater to a wide range of customer needs, ensuring rapid installation and faster time-to-value. Key service offerings include:
- Onsite Engineering Resources: HPE provides dedicated resident engineers who work closely with customer IT teams to optimize system performance and availability.
- Performance and Benchmarking Engagements: HPE's expert team fine-tunes solutions throughout the lifecycle of the system, ensuring peak performance.
- Sustainability Services: HPE prioritizes environmental impact reduction, offering energy and emissions reporting, sustainability workshops, and resource monitoring.
5. Conclusion
Hewlett Packard Enterprise’s shipment of the NVIDIA GB200 NVL72 marks a significant leap forward in AI infrastructure capabilities. As businesses increasingly rely on sophisticated AI solutions, HPE is well-positioned to meet these demands with its innovative technology and extensive support services. The partnership with NVIDIA further solidifies HPE's role as a leader in the HPC and AI landscape, promising customers a future of enhanced performance and efficiency in their AI initiatives.
For more information about HPE’s innovative solutions, visit www.hpe.com.
