Hyperstack review
Updated September 28, 2026 · DeployCue Editorial Team · Facts checked September 26, 2026
We earn a commission when you sign up through links on this site. Partner status can affect where a brand appears in our rankings. Partners with a live offer in your country come first, then every other brand by score.
Our verdict
Hyperstack suits developers and small teams who want on-demand NVIDIA GPUs without a control plane bill. The Kubernetes control plane is free, GPU hours start at $0.50, and free egress means a trained model leaves the platform for nothing at $0.
Key facts
- Best for
- On-demand GPU rentals with a free Kubernetes control plane and free egress.
- GPU models
- NVIDIA H100 SXM, NVIDIA H100 PCIe, NVIDIA A100, NVIDIA L40, NVIDIA RTX A6000, NVIDIA H200 SXM, NVIDIA HGX B200, NVIDIA GB200 NVL72↗
- GPU hour from
- $0.50↗
- Locations
- 3↗
- Legal entity
- NexGen Cloud↗
- Headquarters
- United Kingdom↗
- Founded
- 2023↗
- Support channels
- Email, Phone, Help-centre↗
- Pricing model
- Hourly billing
- Official website
- hyperstack.cloud
Kubernetes
Pros and cons
Pros
- Egress and ingress are free, with egress at $0 per GB, so moving checkpoints out adds nothing
- The control plane is free at $0, and the high availability control plane is listed at $0 as well
- GPU rentals start at $0.50 an hour, with NVIDIA H100 SXM, NVIDIA H100 PCIe and NVIDIA A100 on the list
- Object storage at $0.014 per GB per month, S3-compatible, with object and block storage available
Cons
- Kubernetes clusters cap at 20 nodes
- Object storage spans 1 region while the GPU fleet runs in 3
- Billing is hourly only, per hourly
Who it suits
- Developers who need a specific accelerator from NVIDIA H100 SXM, NVIDIA H100 PCIe and NVIDIA A100 and want to start with a small job
- Kubernetes users who want a free control plane at $0 and a managed interface
- Teams moving large checkpoints and datasets, since egress and ingress are free
Who should look elsewhere
- Buyers who need older data center capacity, since SberCloud lists parts including V100 and A40 alongside H100 and A100
- Teams chasing the broadest accelerator catalogue, which at ComputePrices.com runs from RTX consumer boards up to HGX B300
- Cheap consumer-card capacity seekers: Salad's list, NVIDIA RTX 3060, NVIDIA RTX 3070 and NVIDIA RTX 3070 Ti, is built around RTX consumer parts rather than data center silicon
Who runs it
Hyperstack operates as NexGen Cloud, headquartered in the United Kingdom and founded in 2023. That makes it one of the younger names in on-demand GPU rental. Support runs on email, phone and help centre, so a phone line and a help center sit alongside email, which matters when a training job stalls overnight.
How Hyperstack compares
Hyperstack vs RunPod9.0 vs 9.5
Hyperstack vs ScalaHosting9.0 vs 8.3
Hyperstack vs Contabo9.0 vs 9.2
Hyperstack vs Kamatera9.0 vs 9.5
Hyperstack vs Cloudways9.0 vs 9.2
Hyperstack vs PlanetHoster9.0 vs 8.7
Hyperstack vs Acronis9.0 vs 9.9
Hyperstack vs Aruba9.0 vs 9.3
Hyperstack vs Decodo9.0 vs 9.1
Hyperstack vs Hostinger9.0 vs 9.1
Hyperstack vs MarsProxies9.0 vs 9.0
Hyperstack vs hide.me9.0 vs 8.7
Hyperstack vs Proton Drive9.0 vs 8.3
Frequently asked questions
How much does an hour of GPU time cost on Hyperstack?+
GPU rentals start at $0.50 per hour on hourly billing, listed as hourly. Stronger accelerator classes are priced above that entry point, and the site median for GPU hour entry prices is €0.63, which puts the starting rate below the middle of the market.
Does the Kubernetes control plane cost anything?+
No. The control plane price is $0, and the high availability option is listed at $0 as well, so you can run a managed control plane without a platform fee. Node capacity is the part that costs, and a cluster tops out at 20 nodes.
What does it cost to send data out of Hyperstack?+
Outbound transfer is free, with egress at $0 per GB and free egress and ingress stated alongside it, so no bandwidth or data transfer charge is added. That matters most for teams pulling checkpoints or datasets back out of a training cluster.
Which GPUs can I rent?+
The catalogue covers NVIDIA H100 SXM, NVIDIA H100 PCIe and NVIDIA A100, spanning data center SXM parts, PCIe cards and workstation boards. Everything is rented on the same hourly model, hourly, so switching classes means changing the instance rather than renegotiating a contract.
How much is storage on Hyperstack?+
Object storage runs at $0.014 per GB per month and is S3-compatible, so existing tooling works without changes. Both object and block storage are offered, object and block, and egress from that storage is free. The footprint is a single region, 1, so plan for replication elsewhere if you need it.
Where is Hyperstack based and where does it run?+
The company trades as NexGen Cloud and is headquartered in the United Kingdom. GPU capacity sits in 3 locations, so pick the region closest to your data before committing a long job. Founded in 2023, it is one of the newer names in the market.
How do I reach Hyperstack support?+
Support runs on email, phone and help centre, so email, phone and a help center are the listed routes. A phone line matters when a training run is holding up a team, and the help center sits alongside the Kubernetes and storage documentation.
What do Kubernetes nodes cost?+
Nodes start at $0.35 on top of a control plane priced at $0, so a small cluster is cheap to run and the cost scales with the nodes you add. Capacity per cluster is limited to 20 nodes, which pushes very large deployments into several clusters.
Sources
- docs.hyperstack.cloud checked September 26, 2026
Backs: Storage types
“Deploy GPU/CPU virtual machines and Kubernetes clusters across three regions (CANADA-1, NORWAY-1, US-1), with block storage volumes, S3-compatible object storage, and high-speed networking.”
- docs.hyperstack.cloud/docs/kubernetes/kubernetes-ui checked September 26, 2026
Backs: Free control plane, HA control plane price, Maximum nodes
“Master, Load Balancer, and Bastion nodes - Not billed; these nodes incur no resource-based charges.”
- docs.hyperstack.cloud/docs/resource-management/regions checked September 26, 2026
Backs: Locations
“Hyperstack provides multiple regions (e.g., `CANADA-1`, `NORWAY-1`, `US-1`), each corresponding to a dedicated data center located in the country reflected in its name.”
- docs.hyperstack.cloud/docs/storage/object-storage checked September 26, 2026
Backs: Price per GB per month, Regions
“Object storage is priced at $0.000019397 per GB per hour, which is approximately $14.29 per TB per month.”
- hyperstack.cloud checked September 26, 2026
Backs: GPU models, Headquarters, Legal entity, Support channels
“NVIDIA H100 SXM, NVIDIA H100 PCIe, NVIDIA A100, NVIDIA L40, NVIDIA RTX A6000”
- hyperstack.cloud/about-hyperstack checked September 26, 2026
Backs: Founded
“In 2023, NexGen Cloud officially launched Hyperstack, a European GPU-as-a-Service platform”
- hyperstack.cloud/blog/comparison/top-cloud-gpu-providers checked September 26, 2026
Backs: GPU hour from
“starting for NVIDIA RTX A6000 at $0.50/hour”
- hyperstack.cloud/gpu-pricing checked September 26, 2026
Backs: Egress price per GB, Free egress
“Egress/Ingress traffic Free”
- hyperstack.cloud/object-storage checked September 26, 2026
Backs: S3-compatible
“Built on Amazon S3-compatible technology, it delivers a secure, flexible and API-ready way to manage everything from AI/ML datasets to backups and media files.”
- hyperstack.cloud/pricing checked September 26, 2026
Backs: Control plane price, Node price from
“On-Demand Kubernetes Free master node”
- poolcompute.com/compare/datacrunch-vs-hyperstack checked September 26, 2026
Backs: Pricing model
“Hyperstack also supports both on-demand and reserved instances”
Compiled from 11 source pages, checked September 26, 2026. We did not open an account.
Not confirmed yet: Billing granularity, Egress fee.
Related by the facts
Hyperstack
On-demand GPU rentals with a free Kubernetes control plane and free egress.