Home / Directory / AI infrastructure / Inference & model hosting / Hyperstack
Hyperstack
European self-serve GPU cloud (NexGen Cloud’s Hyperstack) for training and inference hosting with on-demand, reserved, and spot VMs.
Teams that need self-serve GPU VMs for training or inference and prefer transparent hourly pricing with spot and reservation options.
Buyers who only want a hosted model API with scale-to-zero endpoints and no VM management.
Verdict
Hyperstack fits builders who want a GPU cloud with clear hourly SKUs for H100, A100, and lighter cards, including spot and reservation discounts. Public on-demand examples include A4000 around $0.15/hour and H100 around $2.50-$3.20/hour depending on variant. It is closer to infra hosting than Replicate’s model API; capacity and region fit matter as much as the sticker price.
Score Breakdown
How Hyperstack scores in the categories that matter to its buyers.
Pricing
Hyperstack publishes on-demand, reservation, and spot GPU hourly rates on hyperstack.cloud/gpu-pricing. Sample USD rates below as of September 2026; billed per minute on prepaid on-demand.
| Plan | Price | What stands out |
|---|---|---|
| NVIDIA A4000 | $0.15 / GPU-hr | Entry workstation-class GPU |
| NVIDIA A100 (80GB) | $1.35 / GPU-hr | On-demand; reservations and spot lower |
| NVIDIA H100 | $2.50 / GPU-hr | SXM/NVLink variants higher on card |
| NVIDIA H200 SXM | $3.99 / GPU-hr | Top listed on-demand SKU |
The Field at a Glance
Where Hyperstack ranks among Inference & model hosting vendors we reviewed, by Overall Score and relative typical engagement cost.
Hyperstack scores 7.2 beside Modal (8.4), Together AI (7.3), and Replicate (7.3). Relative cost often looks favorable on raw GPU hours versus hosted model APIs.
Use-case matrix
| Use case | Fit | Notes |
|---|---|---|
| Self-serve GPU VM hosting | Strong | Core Hyperstack job. |
| Training and batch inference | Strong | Common workload. |
| Hosted model API scale-to-zero | Mixed | Wrong product shape vs Replicate. |
| Serverless Python job platform | Mixed | Closer to Modal’s lane. |
Who it’s for
Good fit
- Teams comfortable managing GPU VMs
- EU-conscious buyers checking Hyperstack regions
- Workloads that benefit from spot/reservations
Poor fit
- Buyers wanting only a model HTTP API
- Shops that refuse any capacity planning
- Users who need Modal-style serverless functions only
Review Excerpts
Below are excerpts from public reviews. Paid reviews and pay-for-play sites such as Clutch were excluded.
“We spin Hyperstack GPU VMs for training and inference jobs when we want hourly control instead of a black-box model endpoint.”
“Published GPU hourly pricing makes it easier to compare H100 and A100 options against other neo-clouds before you commit.”
“Spot and reservation paths help when batch jobs can tolerate interruption or longer commitments.”
“You still own VM and driver operational work; this is not a one-click model API for product teams.”
Methodology
This page is an independent evaluation of Hyperstack for buyers comparing options in inference & model hosting. AI Industry Reviews accepts no sponsorships, advertising, or pay-for-placement fees. Hyperstack did not pay for this review.
What we scored
The headline number is an Overall Score on a 0-10 scale. Eight criteria fall under it in two groups.
- Buyer outcomes (for inference & model hosting)
- GPU VM availability & SKUs
- Published price transparency
- Inference / training hosting fit
- Self-serve UX
- Company & commercial
- Innovation & product leadership
- Project management & communication
- Pricing
- Contract fairness
Pricing measures whether the price looks fair for the value delivered, including packaging and renewal friction that show up in real buying cycles.
Score Composition
| Input | Weight | What it covers |
|---|---|---|
| Reviews | 40% | A proprietary read of what practitioners say about likes, complaints, and day-to-day use, including public review sites, forums, and private chat rooms. Paid reviews and pay-for-play sites such as Clutch are out of scope. |
| Product | 35% | Hands-on look at screens and workflows. |
| Pricing | 15% | Whether the price looks fair for what you get. |
| Docs & training | 10% | Docs, tutorials, and training material. |
How we balanced the evidence
The Overall Score is the simple average of the eight criteria. Recommendation language follows that score and the fit pattern described above.
Scope
Hyperstack is graded here as inference & model hosting. Criteria scores can move as more review volume and product checks are added.