AMD shipped thousands of Instinct MI300x GPUs to one of the world's largest privately-held cloud computing platforms.

AMD's Instinct MI300X is a high-performance AI accelerator built on the company's third-generation CDNA architecture and TSMC's advanced 5nm and 6nm processes. The processor features 19,456 stream processors, 192GB of HBM3 memory, 304 compute units, 1,216 matrix cores, and a thermal design power (TDP) of 750 watts. In early benchmark tests, it significantly outperformed Nvidia's RTX 4090. Vultr, a leading privately held cloud computing platform, has announced plans to integrate this advanced AMD hardware with its scalable cloud infrastructure to manage GPU-accelerated workloads more effectively and offer seamless performance across both data centers and edge computing.

The partnership aims to provide enterprises access to a world-class AI development environment. "Innovation thrives in an open ecosystem," said J.J. Kardwell, CEO of Vultr. "The future of enterprise AI workloads is in open environments that allow for flexibility, scalability, and security. AMD accelerators give our customers unparalleled cost-to-performance. The balance of high memory with low power requirements furthers sustainability efforts and gives our customers the capabilities to efficiently drive innovation and growth through AI." AMD's architecture, combined with Vultr's platform, will enable businesses to work with open-source, pre-trained models, streamlining AI project development.

"We are proud of our close collaboration with Vultr, as its cloud platform is designed to manage high-performance AI training and inferencing tasks and provide improved overall efficiency," said Negin Oliver, corporate vice president of business development, Data Center GPU Business Unit, AMD. With the adoption of AMD Instinct MI300X accelerators and ROCm open software, Vultr's customers will benefit from having a truly optimized system tasked to manage a wide range of AI-intensive workloads.

The partnership also optimizes Vultr's infrastructure for GPU-accelerated Kubernetes clusters, capable of handling complex workloads globally. Vultr has not revealed the exact number of Instinct MI300X units it has ordered, only that the initial cluster is in the "thousands."