Expanding our Agentic Inference Cloud: Introducing GPU Droplets Powered by AMD Instinct™ MI350X GPUs

DigitalOcean has launched GPU Droplets powered by AMD Instinct MI350X GPUs as part of its Agentic Inference Cloud expansion. Built on AMD CDNA 4 architecture, these GPUs target high-performance AI inference workloads, offering optimized prefill performance, low latency, high token throughput, and support for large models and context windows. Pricing is usage-based with no hidden fees, setup is simplified to a few clicks, and the offering is HIPAA-eligible and SOC 2 compliant. The new droplets are available in the Atlanta (ATL1) datacenter, with AMD Instinct MI355X liquid-cooled GPUs planned for next quarter. Customer examples include Character.AI achieving 2x throughput and 50% cost reduction, and ACE Studio using the platform for AI-driven music creation.