GPU as a Service
Secure dedicated NVIDIA H200 capacity with fixed monthly pricing. Deploy instantly via VMs, Kubernetes, Notebooks or Inference endpoints without hidden fees or extra costs.
The all-inclusive AI engine
Secure your AI/HPC compute requirements without the unpredictable overhead of traditional providers. When you reserve NVIDIA H200 capacity in our Sovereign AI Factory, you are unlocking a fully integrated stack at a single, fixed rate. Every reservation comes standard with high-performance storage, non-blocking interconnects, and the flexibility to choose your deployment environment — VM, Kubernetes, Notebooks or Inference at no extra cost.
Technical specifications
Validated by NVIDIA and powered by HPE supercomputing, our architecture delivers a globally recognized benchmark for AI performance and reliability.
Component | Specification |
|---|---|
GPU Accelerator | - NVIDIA H200 Tensor Core GPUs
- 4.8 TB/s memory bandwidth |
Storage | - Ultra Low Latency Block Storage (NVMe) and S3-Compatible Object Storage
- High-throughput performance |
Networking | - NVIDIA InfiniBand dedicated 3.2 Tbps fabric
- TELUS PureFibre® secure, ultra-low latency external connectivity
- RDMA enabled |
Operational sustainability | - Tier III design-certified facility in Rimouski
- Powered by 99% renewable hydroelectricity
- LEED Gold certification
- N+1 power redundancy and 2N cooling systems |
Dynamic deployment options
Your reserved capacity isn't locked into a single configuration. Through the TELUS AI Studio console, you can instantly toggle between deployment "flavours" to suit the current phase of your project.
Virtual Machines (VMs)
Ideal for:
- Custom training workflows
- Legacy applications
- Familiar OS environments
Capabilities:
- Persistent storage
- Pre-configured NVIDIA drivers
Managed Kubernetes
Ideal for:
- Scalable, production-grade orchestration for containerized AI workloads
Capabilities:
- Auto-scaling node pools
- Automated lifecycle management
- Integration with MLOps pipelines
AI Services
Ideal for:
- LLM inferencing at scale
- Data science and model fine-tuning
Capabilities:
- Spin up Jupyter Notebooks instantly
- Deploy any open-source model or pick from NVIDIA optimized Models directly on your reserved nodes
Pricing
No hidden fees. No egress charges. Just raw compute.1
Instance Type | Single | Small | Medium | Large |
|---|---|---|---|---|
GPU Type | NVIDIA H200 | NVIDIA H200 | NVIDIA H200 | NVIDIA H200 |
GPU Qty | 1 | 2 | 4 | 8 |
Memory | 141GB VRAM | 1.1TB VRAM | 1.1TB VRAM | 1.1TB VRAM |
Price starting from | $5.5/GPU hr | $11/GPU hr | $22/GPU hr | $44/GPU hr |
Availability | Available Now | Available Now | Available Now | Available Now |
- Reserved term options: 1 or 3-year terms.
- Billing: Flat monthly rate per GPU.
- Includes: Compute, Storage, Networking, and AI Studio Platform access.
Why TELUS
Dedicated supercomputing power with one all-inclusive price.
While other clouds add hidden costs for storage, networking, and software, TELUS provides a fully integrated AI Factory at a predictable, fixed rate—with no surprise line items.
Elite supercomputing
Access the nation’s fastest infrastructure—
ranked #1 in Canada by TOP500
—powered by NVIDIA H200 GPUs and 3.2 Tbps InfiniBand networking.All-In predictability
Your reservation is a complete package. High-performance storage, high-speed networking and NVIDIA software are all included in your single fixed price.
Total sovereign control
Your capacity is truly yours. With 100% Canadian residency and no vendor backdoors, your intellectual property remains under your exclusive jurisdiction.
Architectural flexibility
Deploy your way. Use the TELUS AI Studio console to instantly toggle between full-root VMs, managed Kubernetes, or rapid prototyping tools at no extra cost.
Talk with an AI expert
TELUS AI experts can help you determine the best solution for your organization’s unique needs.
Frequently asked questions
Expert AI delivery
Beyond infrastructure
Beyond compute, TELUS engineers work alongside you to accelerate every stage, from data curation and model fine-tuning to production-ready GenAI applications.
Fuel iX
Fuel iX is a powerful generative AI platform built for the enterprise, prioritizing control and flexibility. With its unified control plane, you can build, manage, and monitor GenAI applications from one streamlined location.
Data for AI training
Power your AI models with high-quality, expertly sourced data at scale. From multimodal collection and annotation to validation and testing, we provide solutions to improve model accuracy and safety.
Enterprise AI deployments
Transition from experimentation to full-scale production with ease. Combining data strategy with agentic AI and automation, we help you build secure, production-grade systems tailored to your needs.