Introducing AMD Instinct MI300X GPUs to the DigitalOcean Bare Metal fleet

We’re excited to announce that DigitalOcean customers now have access to AMD Instinct MI300X GPUs with ROCm Software to power their AI workloads.

At DigitalOcean, we’re committed to bringing you even more options to power your projects. One of the highest-bandwith GPUs (5.3 TB/s of HBM3 memory), AMD Instinct MI300X offers significant benefits like faster deployments and much more for AI, machine learning, and high-performance computing (HPC) workloads.

Along with expanding access to GPUs, we continue to improve on our existing offerings for developers like you. Whether you’re deploying popular AI models with our zero-configuration setup that reduces deployment time from weeks to minutes or managing your AI budget with our simple pricing structure, DigitalOcean can help make your AI journey simple, scalable, and cost-effective.

What are AMD Instinct MI300X GPUs?

AMD Instinct MI300X GPUs are designed to boost AI and HPC capabilities in a new compact, efficient package. These next-generation GPUs are now available on DigitalOcean in single-tenant Bare Metal configurations for customers seeking increased control and computing power.

MI300X offers raw acceleration power with eight GPUs per node, leading compute unit counts, and HBM3 capacity. With MI300X, you can accelerate your deployment efforts with ROCm developer resources that support multiple AI and HPC frameworks, leading software platforms and models, and networking infrastructure within an open, proven software ecosystem.

By enhancing computational throughput and simplifying programming and deployment, MI300X and DigitalOcean give you the power to improve data center efficiencies, tackle budget and sustainability concerns, and unleash a highly programmable GPU software platform to quickly build AI projects.

What’s next?

Later this year, AMD Instinct MI300X will also be available as GPU Droplets, flexible on-demand GPUs which enable you to scale AI/ML workloads on-demand without overprovisioning resources or incurring unnecessary costs.

By integrating AMD’s leading GPUs with open ROCm software and our simple, flexible, and scalable platform, we’re committed to providing the critical, innovative technologies to support the evolving needs of your growing tech business.

Ready to accelerate your next AI project? With fast and simple deployment for our bare metal systems - 1 to 2 days average provisioning time - you can start using these powerful GPUs from AMD with ease. Reserve capacity today.