Northern Data Group review
Updated September 29, 2026 · DeployCue Editorial Team · Facts checked September 29, 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
Northern Data Group suits teams wanting enterprise grade Hopper capacity through Taiga Cloud with predictable hourly and monthly billing. Entry through big.mini - 6 vCPU, 90 GB RAM at $474.55 and GPU time from $2.60 billed per minute keeps high end training budgets clear.
Key facts
- Best for
- Enterprise Hopper training on European bare metal capacity
- GPU models
- NVIDIA H100 SXM, NVIDIA H200 SXM↗
- GPU hour from
- $2.60↗
- Billing granularity
- Per minute↗
- Locations
- 10↗
- Interconnect
- 400 Gb InfiniBand↗
- Legal entity
- Northern Data AG↗
- Headquarters
- Germany↗
- Support channels
- Email, Phone↗
- Pricing model
- Hourly billing
- Official website
- northerndata.de
Pros and cons
Pros
- Focused Hopper lineup covering NVIDIA H100 SXM and NVIDIA H200 SXM linked by 400 Gb InfiniBand for large training runs
- Flexible hourly and monthly billing with per minute granularity for short tests and longer runs
- European footprint across GB, PT and SE with 10 locations and email and phone support access
Cons
- Entry remains enterprise priced, with big.mini - 6 vCPU, 90 GB RAM at $474.55 for teams with tight test budgets
- Catalog stays narrow at NVIDIA H100 SXM and NVIDIA H200 SXM, suited to Hopper training rather than lighter inference mixes
- Coverage centres on GB, PT and SE with 10 locations, limiting choice outside those markets
Who it suits
- Hopper training teams who pick Taiga Cloud over Nosana for datacentre grade cards and InfiniBand scale
- European teams who prefer Northern Data Group to Clore.ai for capacity across EU markets
- Monthly planners who favour Northern Data Group over ComputePrices.com listings for direct bare metal terms
Who should look elsewhere
- Small testers better served by Nosana where consumer GPUs lower the cost of early experiments
- Shoppers wanting many card families better served by Clore.ai with broader Hopper and Blackwell choice
- Price comparison shoppers better served by ComputePrices.com for side by side Hopper listings
How Northern Data Group compares
Northern Data Group vs RunPod8.6 vs 9.5
Northern Data Group vs ScalaHosting8.6 vs 8.3
Northern Data Group vs Contabo8.6 vs 9.2
Northern Data Group vs Kamatera8.6 vs 9.5
Northern Data Group vs Cloudways8.6 vs 9.2
Northern Data Group vs PlanetHoster8.6 vs 8.7
Northern Data Group vs Acronis8.6 vs 9.9
Northern Data Group vs Aruba8.6 vs 9.3
Northern Data Group vs Hostinger8.6 vs 9.1
Northern Data Group vs Decodo8.6 vs 9.1
Northern Data Group vs MarsProxies8.6 vs 9.0
Northern Data Group vs hide.me8.6 vs 8.7
Northern Data Group vs Proton Drive8.6 vs 8.3
Frequently asked questions
What GPUs does Northern Data Group offer+
Northern Data Group lists NVIDIA H100 SXM and NVIDIA H200 SXM for Taiga Cloud, aimed at training and large inference jobs. Nodes use 400 Gb InfiniBand to keep multi GPU runs linked. The catalog stays focused on datacentre Hopper cards rather than broad consumer options.
How does GPU pricing work+
GPU time starts from $2.60, billed per minute under hourly and monthly options. That structure suits both brief scale tests and longer reserved style runs. Actual workload cost depends on card choice, cluster size and run duration.
What is the entry plan and price+
Entry is through big.mini - 6 vCPU, 90 GB RAM, priced at $474.55. That bundle provides vCPU and RAM for Hopper driven projects rather than small trial use. Teams should size memory and CPU needs before choosing larger bare metal shapes.
Where does Northern Data Group host workloads+
Capacity runs across GB, PT and SE, with 10 locations reported. That footprint suits teams wanting European hosting for latency, data control and support hours. Availability by card and cluster size can vary across those markets.
How can teams get support+
Support runs on email and phone only, through Northern Data AG. Email suits ticket style requests while phone suits urgent cluster issues. Teams should confirm response hours and escalation paths before moving production training.
What billing choices are available+
Plans use hourly and monthly billing under a hourly model for bare metal servers. Hourly billing suits elastic training while monthly billing suits steady capacity. Teams should match billing choice to utilization to avoid idle spend.
Sources
- northerndata.de checked September 29, 2026
Backs: Locations, Support channels
“10 infrastructure locations globally”
- northerndata.de/legal/imprint checked September 29, 2026
Backs: Headquarters, Legal entity
“60322 Frankfurt am Main”
- northerndata.de/taiga-cloud-ai-pricing checked September 29, 2026
Backs: Billing models, Entry plan, Entry price, GPU hour from, GPU models, Pricing model, Server types
“Price/Hour ($) Price/Month (730h) ($)”
- northerndata.de/taiga-cloud-by-northern-data-group checked September 29, 2026
Backs: Billing granularity, Interconnect, Regions
“transparent pay-as-you-go, per-minute billing”
Compiled from 4 source pages, checked September 29, 2026. We did not open an account.
Not confirmed yet: Egress fee.