A Strategic Collaboration with AMD, Broadcom, and Juniper Networks
Vultr has announced the launch of its AMD GPU supercomputer cluster in Chicago, leveraging cutting-edge technologies from AMD, Broadcom, and Juniper Networks. This initiative highlights Vultr’s ongoing efforts to provide high-performance, scalable solutions for AI and enterprise workloads.
The cluster is hosted at Centersquare’s Lisle, Illinois data center, strategically located in the Illinois Technology and Research Corridor. The facility spans 194,057 square feet (18,028 square meters) and offers up to 15MW of utility power—an ideal environment for the new supercomputing infrastructure.
Powered by AMD Instinct MI300X GPUs and ROCm Software
Cutting-Edge Hardware for AI and HPC Workloads
At the heart of Vultr’s supercomputer cluster are AMD’s Instinct MI300X GPUs, paired with the ROCm open software stack. This combination is designed to optimize performance for AI, machine learning, and high-performance computing (HPC) applications. Although specific details on the number of GPUs and overall compute power have not been disclosed, Vultr’s integration of AMD’s latest GPUs marks a significant step forward in GPU-driven cloud computing.
According to Negin Oliver, corporate vice president of AMD’s data center GPU business unit, “AMD is proud to collaborate with Broadcom, Juniper, and Vultr to bring state-of-the-art AMD Instinct MI300X GPUs and ROCm software stack to Vultr’s composable cloud infrastructure to power enterprises’ AI development and deployment.”
Networking Backbone: Broadcom and Juniper Solutions
The cluster’s networking capabilities are built on Broadcom’s Ethernet switchers and Juniper Networks’ AI-optimized Ethernet solutions. This robust combination ensures low-latency connectivity, high bandwidth, and enhanced scalability, enabling seamless operation of the supercomputing infrastructure.
Vultr’s Expanding AI Cloud Ecosystem
Diversified GPU Offerings for Maximum Flexibility
The launch of the AMD GPU supercomputer cluster complements Vultr’s existing GPU cloud offerings, which include Nvidia H100 GPUs and GH200 chips. This diversification allows enterprises and developers to select the most suitable hardware for their specific needs, from training AI models to deploying sophisticated applications.
Vultr began offering AMD Instinct MI300X GPUs in September 2024, underscoring the company’s commitment to supporting next-generation AI workloads. With its presence in 32 data centers across six continents, Vultr is well-positioned to meet the growing demand for high-performance GPU computing.
Sovereign and Private Cloud Solutions
In addition to its public cloud offerings, Vultr launched a sovereign and private cloud solution in April 2024. This move aligns with global trends emphasizing data sovereignty, security, and compliance. By providing tailored cloud environments, Vultr addresses the unique requirements of government agencies, enterprises, and industries with strict regulatory needs.
Driving Innovation Through Open Ecosystems
Collaborative Efforts to Accelerate AI Adoption
Vultr’s collaboration with AMD, Broadcom, and Juniper Networks exemplifies the power of open ecosystems in driving innovation. “Our collaboration empowers enterprises and AI innovators to harness the full potential of accelerated computing with the highest levels of flexibility, scalability, interoperability, and security,” said J.J. Kardwell, CEO of Vultr.
The deployment of the AMD GPU supercomputer cluster in Chicago is a testament to Vultr’s vision of democratizing access to advanced computing resources. By combining state-of-the-art hardware with innovative software and networking solutions, Vultr enables enterprises to accelerate their AI and HPC initiatives.
Centersquare’s Lisle Data Center: A Strategic Location
Key Features of the Facility
The Lisle data center, operated by Centersquare, serves as the hosting site for Vultr’s new cluster. Located within the Illinois Technology and Research Corridor, this facility offers:
- 15MW of utility power
- 194,057 square feet of operational space
- Proximity to key tech hubs and enterprises in the region
The data center’s robust infrastructure provides the necessary foundation for hosting high-performance supercomputing workloads.
FAQ: What You Need to Know
What is the significance of the AMD Instinct MI300X GPUs?
The AMD Instinct MI300X GPUs are designed for AI and HPC workloads, offering high performance, energy efficiency, and compatibility with AMD’s ROCm open software stack.
Where is the Vultr AMD GPU supercomputer cluster located?
The cluster is hosted at Centersquare’s Lisle, Illinois data center, situated in the Illinois Technology and Research Corridor.
What other GPU options does Vultr offer?
In addition to AMD’s GPUs, Vultr’s GPU cloud features Nvidia H100 GPUs and GH200 chips, providing flexibility for a range of applications.
How does the collaboration with Broadcom and Juniper Networks enhance the cluster?
Broadcom’s Ethernet switchers and Juniper’s AI-optimized Ethernet solutions enable high-speed, low-latency connectivity, ensuring optimal performance for AI and HPC workloads.
Does Vultr offer private cloud solutions?
Yes, Vultr launched sovereign and private cloud solutions in April 2024, catering to organizations with specific compliance and security requirements.
