Nscale Cloud (GPU compute, Inference Endpoints & Managed Slurm/Kubernetes)

Offering overview

Hydro-powered sites in Glomfjord and Narvik, Norway anchor Nscale's own data center footprint, complemented by Loughton in the UK, Texas in the US, and partner facilities in Portugal, Iceland, and elsewhere in Norway, the UK, and the US. The company designs its own modular, liquid-cooled halls and sells the capacity through Nscale Cloud as bare metal, virtual instances, a Kubernetes service, and managed Slurm for distributed training. On top of the infrastructure layer sit AI services: serverless Inference Endpoints exposing open models such as Llama 4 Scout, GPT-OSS 20B/120B, Mixtral 8x22B, GLM, Qwen3 variants, FLUX.1 schnell, and Stable Diffusion XL; a Prompt Workbench for iteration; and a fine-tuning service. Enterprise IAM, strict tenant isolation, observability, and a Control Center for fleet operations target larger AI teams and sovereign-minded buyers. GPU cluster pricing is not published; capacity is reserved through sales in blocks of thousands of GPUs. Per-token rates for the serverless models are shown only inside the console and via the models API, so no public starting price is available.

Technical specifications

Containers

Unknown

Kubernetes

Unknown

Root access

Unknown

API access

Unknown

Autoscaling

Unknown

Persistent processes

Unknown

AI capabilities

Model serving

Llama 4 Scout, GPT-OSS 120B, GPT-OSS 20B, Mixtral 8x22B, GLM 5.3, Qwen3.8-27B, Qwen3-Embedding-8B, FLUX.1 schnell, Stable Diffusion XL

AI frameworks

Not stated

Inference endpoints

Unknown

Fine-tuning

Unknown

Vector database

Unknown

Model catalog

Unknown

Monitoring

Not stated

Evaluation

Not stated

Security & compliance

Geography

Provider headquarters are separate from offering infrastructure and availability.

Available plans

Pricing can change. Promotional prices are shown separately from standard prices.

No active plans are recorded for this offering.

Questions & answers

What infrastructure services does Nscale Cloud provide?

Nscale Cloud offers bare metal, virtual instances, a Kubernetes service and managed Slurm for distributed training.

What AI services are offered beyond GPU compute?

Nscale offers serverless Inference Endpoints for open models, a Prompt Workbench for iteration and a fine-tuning service.

Where is Nscale's infrastructure located?

Its footprint includes Glomfjord and Narvik in Norway, Loughton in the UK and Texas in the US, alongside partner facilities in Portugal, Iceland, Norway, the UK and the US.

What features support enterprise operations?

Nscale provides enterprise IAM, strict tenant isolation, observability and a Control Center for fleet operations.

How can buyers obtain pricing?

GPU cluster pricing is not published, and capacity is reserved through sales in blocks of thousands of GPUs. Serverless model rates are shown in the console and through the models API.