Akamai Cloud GPUs Powered by NVIDIA

Distribute AI inference and GPU-accelerated workloads on demand with NVIDIA-powered instances running on Akamai’s globally distributed cloud.

Create an account to deploy now, or join the waitlist for NVIDIA RTX PRO 6000 Blackwell Server Edition.

Why choose Akamai GPUs

• Enterprise-grade acceleration: RTX PRO VMs optimized for AI inference and graphics
• Low-latency delivery: run closer to users and data for faster responses
• Predictable costs: on-demand pricing with low egress and no long-term contracts
• Developer-friendly: deploy via UI, API, CLI, or Terraform; 24/7/365 support

NVIDIA GPU options on Akamai

• NVIDIA RTX PRO 6000 Blackwell Server Edition (96 GB GDDR7)
• Designed for advanced AI inference, agentic and physical AI, 3D, rendering, and scientific compute
• Availability via waitlist. Join the waitlist and review the data sheet
• NVIDIA RTX 4000 Ada Generation
• Balanced price/performance for ML inference, analytics, and media workloads using CUDA, Tensor, and RT cores
• View the data sheet
• NVIDIA Quadro RTX 6000
• A media workhorse with dual encode/decode engines and AV1 support; great for transcoding and visualization
• View the data sheet

Not sure which to pick? Talk with our team and match your workload to the right GPU.

Pricing and egress

• On-demand GPU pricing with no contracts; RTX 4000 Ada plans start at $0.52/hour
• Low egress: save up to 90% with pricing as low as $0.005 per GB in most regions
• See all current plans and rates on the GPU pricing page

Performance and efficiency highlights

• Up to 86% lower inference cost demonstrated with Stable Diffusion on Akamai Cloud. Download the white paper
• Benchmarks show NVIDIA RTX PRO 6000 Blackwell on Akamai Cloud delivers up to 1.63x higher inference throughput than H100. Read the benchmarking analysis
• Media pipelines benefit from dedicated encode/decode engines (including AV1) on supported GPUs

Built for your stack

• Kubernetes-ready: add GPU nodes to Akamai Kubernetes (LKE)
• Easy migration: resize Shared or Dedicated CPU instances into GPU instances
• CI/CD friendly: APIs, CLI, Terraform provider, custom images
• Data protection and durability: Backups with automated snapshots
• Manage via UI, API, or CLI with full documentation and 24/7/365 support

What you can run

• Real-time AI inference and agentic workloads
• RAG pipelines and personalization services at the edge
• Media transcoding, rendering, and interactive streaming
• Scientific computing, analytics, and visualization

Deploy in minutes

1. Create your account and choose a GPU plan
2. Provision in your preferred region and attach storage/networking
3. Install NVIDIA drivers and CUDA toolkit using the GPU setup guide
4. Launch your model or container (vLLM, TensorRT, PyTorch, or your stack) and scale on demand
5. Optional: add GPU nodes to LKE and deploy with KServe/Kubeflow

Go further with Akamai Inference Cloud

If you’re building edge-native AI or agentic applications, pair GPUs with Akamai Inference Cloud for global traffic management, AI-aware security, vector databases, and low-latency inference at the edge.

Resources

• See all GPU pricing and plans
• Read “Edge Is All You Need” on why inference belongs at the edge: Akamai blog
• NVIDIA CUDA setup and operations: GPU documentation

Ready to start? Create an account. Prefer guidance? Book an AI consultation with our team.