NEW YORK: DigitalOcean Holdings, Inc. (NYSE: DOCN) announced a collaboration with AMD to provide its customers with access to AMD Instinct™ GPUs, (AI GPUs) starting with the MI300X model, to power artificial intelligence workloads.
The cloud computing company said it will later expand its offerings to include AMD’s high-performance MI325X GPUs, giving developers more affordable options for AI training and inference.
AMD’s MI300X GPUs are designed for high-performance computing and generative AI applications, featuring large memory capacity to handle complex AI models without splitting them across multiple GPUs. The chips are supported by AMD’s ROCm™ software platform, which accelerates AI and HPC development.
DigitalOcean will offer these GPUs through its GPU Droplets and Kubernetes worker nodes, available in single or eight-GPU configurations. Pricing starts at $1.99 per GPU per hour, with simplified setup compared to competitors, the company said.
“This collaboration makes AI more accessible to our customers,” said Bratin Saha, DigitalOcean’s chief product and technology officer. “AMD’s GPUs provide flexibility and performance for a range of AI workloads.”
Additionally, DigitalOcean will power AMD’s new Developer Cloud, a managed platform offering instant access to MI300X GPUs for AI development without upfront hardware costs.
“We’re proud to work with DigitalOcean to deliver cutting-edge solutions for developers,” said Negin Oliver, AMD’s corporate vice president of business development for data center GPUs. “Together, we’re supporting the growing demands of tech businesses.”
DigitalOcean, known for its developer-friendly cloud services, aims to simplify AI adoption for small and midsize businesses with this partnership.