Gradient™ AI Bare Metal GPUs for AI/ML Workloads

Now Available: Bare Metal AMD Instinct MI300X GPUs

AMD’s Instinct MI300X unleashes cutting-edge compute and memory technologies for best-in-class performance for AI model training, inference, and high-performance computing. Built to handle the most data-intensive AI and high-performance workloads, it ensures seamless data processing, reduces bottlenecks, and enables faster results.

Learn more

Supermaven's AI code completion tool boosts coding speed and efficiency with high-quality suggestions, in-editor AI chat, and a large context window of up to 1 million tokens. Powered by their advanced LLM, Babble, it's 2.5x larger than their previous model.

Gradient AI Bare Metal GPUs

Compare the key specs of Bare Metal GPUs to understand how each can power your AI applications

Talk to an expert to reserve capacity today

GPU Model	GPU Count	GPU Ram	CPU (x2)	System RAM	NVMe Storage
NVIDIA HGX H100	8	640 GB	Intel® Xeon® Platinum 8468	2,048 GiB	61.44 TiB
NVIDIA HGX H200	8	1,128 GB	Intel® Xeon® Platinum 8592+	2,048 GiB	61.44 TiB
AMD MI300XNew	8	1,536 GB	Intel® Xeon® Platinum 8568Y	2,048 GiB	61.44 TiB

NVIDIA HGX H100

GPU Count8

GPU Ram640 GB

CPU (x2)Intel® Xeon® Platinum 8468

System RAM2,048 GiB

NVMe Storage61.44 TiB

NVIDIA HGX H200

GPU Count8

GPU Ram1,128 GB

CPU (x2)Intel® Xeon® Platinum 8592+

System RAM2,048 GiB

NVMe Storage61.44 TiB

AMD MI300X

New

GPU Count8

GPU Ram1,536 GB

CPU (x2)Intel® Xeon® Platinum 8568Y

System RAM2,048 GiB

NVMe Storage61.44 TiB

High-Performance Bare Metal GPUs for AI/ML Workloads

For serious AI builders who need more control, power, and less noise in their cloud infrastructure, Gradient AI Bare Metal GPUs with accelerated computing are ready to deploy. Purpose-built for the most demanding AI/ML workloads and featuring powerful hardware, these servers offer high-performance compute capabilities and customizable options designed for your most intense processing needs.

DigitalOcean Bare Metal GPUs provide full access to all GPUs, offering dedicated, single-tenant infrastructure with no neighbors that makes them ideal for large-scale model training, real-time inference, and complex orchestration. Take direct control over Bare Metal GPU hardware to ensure highest performance and privacy. With Bare Metal GPUs, you can fully optimize model training, fine-tune pre-trained models, achieve high-speed inference, and build scalable custom architectures for specialized use cases.

Power your projects with the flexibility of Bare Metal GPUs

Platform Reliability

Get more reliable, high-performance compute for demanding workloads.

Support

Receive high-touch engineering support, helping to ensure smooth operations and peak server performance.

Pricing

Enjoy cost-efficient pricing that includes storage, and favorably compares to other providers.

Availability

Bare Metal GPUs are available in New York, USA and Amsterdam, Netherlands, with more data centers coming soon.

Bare Metal GPU Resources

What are Bare Metal GPUs?

GPU Memory bandwidth

Monitoring GPU utilization for Deep Learning

Multi-GPU on raw PyTorch with Hugging Face's Accelerate library

How to maximize GPU utilization by finding the right batch size

The Hidden Bottleneck: How GPU Memory Hierarchy Affects Your Computing Experience

Frequently asked questions about DigitalOcean Gradient™ Bare Metal GPUs

What are Gradient Bare Metal GPUs?

Gradient Bare Metal GPUs are dedicated, single-tenant servers equipped with eight high-performance GPUs (such as NVIDIA HGX H100, H200, or AMD MI300X) designed to support intensive AI/ML workloads on bare-metal infrastructure.

Which GPU configurations are available for bare-metal servers?

Available options include:

NVIDIA HGX H100: 8 GPUs, 640 GB GPU RAM
NVIDIA HGX H200: 8 GPUs, 1,128 GB GPU RAM
AMD MI300X: 8 GPUs, 1,536 GB GPU RAM

All configurations include dual Intel Xeon Platinum CPUs, 2,048 GiB system RAM, and 61.44 TiB NVMe storage.

How do Bare Metal GPUs compare to GPU Droplets?

GPU Droplets are virtualized and great for flexible, on-demand GPU workloads but share infrastructure. Bare Metal GPUs provide isolated hardware without virtualization—offering consistent, high-performance, and customizable environments for demanding AI training, inference, or HPC tasks.

What types of workloads are best suited for Bare Metal GPUs?

Ideal use cases include:

Large-scale model training and distributed deep learning

High-throughput, low-latency inference serving

Custom orchestration setups (e.g., Kubernetes clusters)

High-performance computing and scientific simulation workloads

Applications needing strict data isolation or compliance guarantees

Where are Bare Metal GPUs available globally?

Currently, Bare Metal GPUs are available in New York (USA) and Amsterdam (Netherlands). Additional regions are expected to launch in the future.

What is the provisioning and billing process for Bare Metal GPUs?

Bare Metal GPUs are provisioned via contract with DigitalOcean sales (not on-demand). Billing is handled through Paperspace, requiring a Paperspace account to manage payments.

What networking and connectivity features are included?

Each bare-metal server includes:

One public IP (via NAT) and a private VLAN IP

Public bandwidth up to 40 Gbps (North/South traffic)

Private VPC bandwidth up to 400 Gbps (East/West traffic)

GPU interconnect with IPv6 support and RDMA speeds up to 3.2 Tbps

Can I install custom software and drivers on Bare Metal GPUs?

Yes—you have root-level control. Servers come preconfigured with recommended OS (Ubuntu 22.04 or 24.04), GPU drivers, CUDA, or ROCm—but you can modify or install any software, driver, or toolkit you need.

What security considerations should I be aware of?

Bare Metal GPUs have no default firewall. It's recommended to use UFW or your preferred firewall to secure the public interface. IPMI access is not provided.

Who benefits most from using Bare Metal GPUs?

This product is especially well-suited to:

Advanced developers needing high throughput, full hardware control, and multi-node orchestration

Enterprises and AI-native businesses running consistent, high-utilization AI workloads, requiring performance predictability, compliance, and isolation

Bare Metal GPUs deliver the high-performance, scalable, and secure infrastructure demanded by AI-first, digital-native enterprises.