Dedicated GB10 · Serverless API · Managed GPU

European AI compute, from API to dedicated GB10.

Rent dedicated NVIDIA GB10 compute by the hour, week, month or year, or run leading open models through one OpenAI-compatible API — all in Europe.

Built by founders and engineers with 15+ years in highly regulated industries. Every new account includes 3M free serverless tokens every 30 days.

Three products

Run AI your way.

Same account, same EU regions, three amounts of control.

Serverless · Managed · Rental

  • Serverless Models

    Choose a hosted model and call the API. We operate everything underneath.

    Available

    €0.32/M input tokens, from

    Explore serverless models

  • Managed GPU

    Tell us what to run. We deploy and operate the environment for you.

    Available

    €664.30/month, dedicated GB10

    Explore managed GPU

  • GPU rental

    Rent a dedicated machine with full SSH access — by the hour, week, month or year. You control the stack.

    Available

    €1.19/hour, DGX Spark (GB10)

    Rent a GB10

The fleet

Infrastructure behind AxForge.

Named hardware, current status. Rent it directly, or have us run it for you.

EUR · ex VAT

The AxForge hardware fleet with memory, status and list price
System Memory Status Price
NVIDIA DGX SparkGB10 Grace Blackwell · dedicated machine 128 GB unified Available from €0.85/hour
RTX 6000 ProBlackwell workstation 96 GB GDDR7 Request capacity €1.58/GPU-hr
4× RTX 5090One system 32 GB GDDR7 each Request capacity €0.89/GPU-hr
4× RTX 3090Budget tier 24 GB GDDR6X each Request capacity €0.62/GPU-hr
RTX 3060 ×2 (dual-GPU machine)Starter tier — two 12 GB GPUs per machine, or one card 12 GB GDDR6 each Available €0.41/hour
NVIDIA H100 / H200Capacity on request 80 / 141 GB HBM Request capacity —
NVIDIA B200Capacity on request 192 GB HBM3e Request capacity —

€1.19/hour on demand · €0.95/hour by the week · €0.91/hour by the month · €0.85/hour by the year, no VAT charged for a dedicated DGX Spark. Models are €0.02 – €2.00 per million tokens on the same account (every list price). Request-capacity systems carry their launch price; the deployment is confirmed with your request. Dual RTX 3060 machines rent like the GB10 — Check availability.

Check availability Reserve capacity

  • eu-es-1 · Málaga — dedicated GPUs
  • eu-se-1 · Stockholm — serverless

What we took out

Six things you expect. None of them are here.

Every one of them is normal somewhere else. You have probably paid for a few.

One line each

  • No sales call to learn the price. Every list price is on the pricing page.
  • No egress fees. What you pull out is not metered against you.
  • No subscription. Rent by the hour, week, month or year — end it when you're done.
  • No invented SKU names. DGX Spark is the part, and the invoice line.
  • No training on your data. In the contract and the DPA.
  • No lock-in to one level. Move up or down on the same account.

The proof

The usage is the receipt.

One request. The response carries exact token usage — multiply by the published prices.

Returned on every call

api.axforge.ai/v1TLS 1.3 · terminated in the EU

# one variable changes. your client does not.
$ export OPENAI_BASE_URL=https://api.axforge.ai/v1
$ curl -sS "$OPENAI_BASE_URL/chat/completions" \
    -H "authorization: Bearer $AXFORGE_KEY" \
    -d '{"model":"qwen3.8-27b-nvfp4", ... }'

< HTTP/2 200
{ "model": "qwen3.8-27b-nvfp4",
  "usage": { "prompt_tokens": 118, "completion_tokens": 412 },
  "choices": [ ... ] }
Where it executed
eu-se-1 · StockholmPinned on the key — inference runs in-region.
In region
What happened to the data
Retention none · training disabledNothing written after the response — in the contract and the DPA.
Contract
The machine class
NVIDIA DGX Spark · 128 GBThe same part you can rent — dedicated, from €0.85/hour.
Named
The price
€0.00086 for this call118 in + 412 out, at the published per-token prices.
Metered
Read it yourself
Your logs, not our slidesUsage on every response × the published prices — an auditor just multiplies.
Every call

Sovereignty

Where your data lives, as facts.

The six questions a procurement team asks, answered before they ask.

Enforced per request

Sovereignty facts: regions, data at rest, training, operator, subprocessors and egress
Fact Answer Detail
Regions eu-se-1 · eu-es-1 eu-se-1 · Stockholm for serverless inference and RTX 3060 rentals, eu-es-1 · Málaga for dedicated DGX Spark workloads.
Data at rest EU only Prompts, outputs, weights and logs, keys held in region
Training Never Customer data is not used to train models.
Operator AxForge AxForge — operated in the EU, with infrastructure in Sweden and Spain. Contracting entity details are provided in the Legal Notice.
Subprocessors Named in the DPA All EU-established, 30 days' notice before a change
Egress None outside the EU No egress fee inside it; metadata kept 30 days, in region

Residency

Residency is enforced at the edge — a request pinned to a region fails rather than leaves it.

Audit position

We make no certification claims on this page; ask for the current audit position and you will get a straight answer.

Figure — the path of one request Application → EU ingress → named accelerators
Your application Any OpenAI-compatible client, on your own infrastructure.

EU region boundaryeu-se-1

EU ingress · api.axforge.ai Auth and metering. The region is pinned on the key.
Inference on named accelerators Your dedicated systems, in that region only.

Training corpus

No path leads into it. Your prompts, outputs and weights never train a model.

Request path
EU region boundary
No path — training corpus unreachable

Start here

Start now, or tell us the shape.

Create an account and your API key is in the console. For a machine, check availability and reserve.

Every new account includes 3M free serverless tokens every 30 days. A key works with any OpenAI-compatible client the moment it is issued.

What we need from you

For models
Model, tokens per day, latency target, region.
For GPUs
Accelerator, count, fabric, region, term.
You get
A written configuration and a price.
© 2026 AxForge · EU-hosted AI infrastructure Pricing Docs Trust Privacy Terms

Engineering

Tell us the shape. An engineer replies.

Pick what is closest — rough is fine.

Reply from a person.

Sent — an engineer will email you at the address above.