A storm cloud lit pink and blue by lightning over the rocky Swedish west coast at night, with a red boathouse on the shore.
Nov 25 Meet us at Innovatum Demo Day, Trollhättan

Sovereign AI infrastructure,
in your rack or ours

A complete AI platform, hardware and software included, from a 10 inch Mini to a full 19 inch Large. Compatible with the stack you already run, accountable to the jurisdiction you actually operate in.

NVIDIA AI Sweden

Built for defense, energy, healthcare and the public sector.

Partnered with NVIDIA. Based at AI Sweden in Göteborg.

We build with

  • NVIDIA
  • AI Sweden
  • Run:ai
  • Dell Technologies
  • VAST Data
  • PEAK:AIO
  • Arista
  • Red Hat
  • VMware
  • Weights & Biases
  • Fortinet
  • NetClean
  • TD SYNNEX

Why it matters

Who else can reach your data?

Most AI capacity today is rented from providers that answer to laws outside Europe. For a defense supplier, a grid operator or a hospital, that leaves a question no audit lets you skip.

With Gridcloud Pod, the capacity is yours: a dedicated AI platform in your own rack, or in ours. Your training data, your inference and the models in between stay on your Pod, and the platform does not call home.

Gridcloud Pod

One platform, four sizes

The four Gridcloud Pod sizes side by side, from the small 10 inch Gridcloud Pod Mini to the tall 19 inch Gridcloud Pod Large, each rack lit pink and blue.
  1. Mini 10 inch rack Get started on your own site.
  2. Small 19 inch rack For a team's first production workloads.
  3. Medium 19 inch rack Room for several teams sharing GPUs.
  4. Large 19 inch rack Your organisation's own AI factory.
Illustrations. Configurations are agreed per customer.
A row of Gridcloud Pod Large racks standing side by side and receding into the dark, in our cloud.

Or just the compute

Buy just the compute and run on Gridcloud in our racks, in Europe, under European law. No Pod of your own.

See SaaS

What you get

Your own AI platform, as one product

NVIDIA hardware, the software to run it and the people who look after it, from the first install to every update after.

Exploded view of a Gridcloud Pod Mini: two gold NVIDIA DGX Spark systems, three control nodes, a network switch, a fan tray and the parts of its 10 inch rack.

NVIDIA compute in one rack

GPU systems, control nodes, storage and networking, built into one rack and installed by Stormgrid. Four sizes, from a 10 inch Gridcloud Pod Mini to a 19 inch Gridcloud Pod Large.

The Gridcloud Pod portal for a Gridcloud Pod Medium: GPU allocation at 91 percent, 31 workloads running, and the Pod's network, GPU nodes and control plane all reported healthy.

Ready for your teams on day one

Kubernetes, GPU scheduling with no license fees, a private image registry and monitoring come installed, with a portal that shows what every GPU is doing. Your containers run as they are.

A Gridcloud Pod glowing pink and blue through the window of a building on the shore, while a thunderstorm passes over at night.

Keeps running when the internet drops

In your own rack, the Pod runs on your network and does not call home, so a lost connection does not stop your workloads. Or we host it for you, in Europe, under European law.

Gridcloud Pod Mini, a 10 inch rack with two gold NVIDIA DGX Spark systems, glowing pink and blue on a wet granite rock by the sea, with a thunderstorm over the horizon.

Start with a Gridcloud Pod Mini

Lease a 10 inch Pod with two NVIDIA DGX Spark systems from SEK 30,000 a month, excluding VAT. Try it where your data lives, and keep what you build when you move to a bigger size.

How it works

From first call to running models in three steps

  1. 01

    Tell us what you will run

    Your workloads, and how sensitive the data is.

  2. 02

    Pick a size and a place

    From a Gridcloud Pod Mini to a Large, in your own rack or in ours.

  3. 03

    Your teams run their AI

    Train and serve on NVIDIA compute, with the tools your teams already use.

Every AI runs somewhere.
Yours can run
in your own rack.

Who is behind it

Built by people who have run critical systems

Our team comes from defense, automotive, energy and the public sector, and has built AI infrastructure since 2019. We work out of AI Sweden at Lindholmen in Göteborg and build on NVIDIA.

Meet the team
  • NVIDIA

    Built on NVIDIA

    From a Gridcloud Pod Mini to a full rack of GPUs, on one platform.

  • AI Sweden

    Architect lead at AI Sweden

    We led the architecture when AI Sweden redesigned its national AI infrastructure, and we work out of Lindholmen in Göteborg.

  • GPU

    Trusted with GPU fleets

    Automotive and public sector teams run their GPU scheduling and AI infrastructure with us.

  • ISO

    27001, 9001 and 14001

    Certification work is underway. Security is built into every service we ship.

Questions

What buyers ask us first

Who can be forced to hand over my data?

It depends on where your Pod stands. In your rack, it is under your own jurisdiction. If we host it, it stands in Europe, under European law. Stormgrid is a Swedish company.

Can you handle secret and top secret data?

For data at that level, we recommend running the Pod in your own rack, on your network and under your own processes. It keeps running if the internet drops and does not call home. Tell us what your rules require and we will say plainly whether it fits.

How is it priced?

A Gridcloud Pod Mini is leased from SEK 30,000 a month, excluding VAT. The bigger sizes are priced by size and by where they run: talk to us for a quote.

Who do you work with?

Organisations in defense, energy, automotive, healthcare and the public sector across EU, Nordic and NATO countries, wherever the data rules out a foreign cloud.

Start on your own site

Try it before
you commit

Lease a Gridcloud Pod Mini, run real workloads on it and keep what you build when you move to production.

Talk to us info@stormgrid.se. A person replies, usually within a working day.

Gridcloud Pod

Talk to us

We only use your details to reply to you.