AI infrastructure provider for enterprises

Live — NVIDIA Blackwell GPU available now

The open GPU
cloud for AI teams
and enterprises

Rent NVIDIA Blackwell GPU GPU instances from £0.30/hr. List your idle GPU and earn 77%. Deploy dedicated clusters for universities and research institutions across the UK.

Start computing → Enterprise clusters

UK data residency

99.9% SLA

Live Stripe payments

SSH in 60 seconds

Live demo — real output from our NVIDIA Blackwell GPU 310 tok/s

Run a real LLM on
NVIDIA Blackwell GPU
in 60 seconds

SSH into your GPU container, install any framework, run your model at 310 tokens/second. Download the weights when done. Your GPU, your data, your model.

1 Launch instance — container ready in 30 seconds

2 SSH in — full root access to 130.6GB GPU

3 Run LLM at 310 tok/s — PyTorch, HuggingFace, vLLM

4 Download your model — destroy instance, billing stops

Try it now — £0.52/hr Read the docs

gb10-blackwell — bash ● LIVE

          NVIDIA Blackwell GPU · 130.6GB VRAM · CUDA 13.0
          ● Live hardware
        

NVIDIA Blackwell GPU · 128GB VRAM · £0.30/hr Host earnings · 77% of every rental · Stripe direct payout Enterprise clusters · 99.9% SLA · UK data residency Free AI chat · Powered by on-site Blackwell GPU CUDA 13.0 · Driver 580.142 · Docker GPU passthrough NVIDIA Blackwell GPU · 128GB VRAM · £0.30/hr Host earnings · 77% of every rental · Stripe direct payout Enterprise clusters · 99.9% SLA · UK data residency Free AI chat · Powered by on-site Blackwell GPU CUDA 13.0 · Driver 580.142 · Docker GPU passthrough

Three ways to use the platform

Rent GPU compute

Launch a Docker container with full NVIDIA GPU access in under 60 seconds. SSH in directly, bring any framework.

£0.30/hr on-demand

Launch an instance →

Host your GPU and earn

List any NVIDIA GPU with 8GB+ VRAM on our marketplace. Earn passively while your GPU would otherwise sit idle.

77% of every rental

Start hosting →

Enterprise GPU clusters

Dedicated federated clusters for UK universities and research institutions. 99.9% SLA with automatic failover.

£500/month from

Request a cluster →

Platform numbers

128 GB

VRAM on NVIDIA Blackwell GPU — the most memory-rich GPU available on our platform

£0.30

Per hour for on-demand compute — typically 60–70% cheaper than equivalent AWS or Azure GPU instances

77%

Of every rental goes directly to the GPU host via Stripe — one of the highest host shares in the market

Hardware specs — NVIDIA Blackwell GPU

128

GB unified memory

13.0

CUDA version

580

Driver version

Blackwell

NVIDIA GPU

99.9%

SLA uptime

Data residency

$ ssh [email protected]

Connected to Open Source AI Cloud · Instance #12 · NVIDIA Blackwell GPU

$ nvidia-smi

+-------------------------------------------------------------------------+

| NVIDIA Blackwell GPU Driver: 580.142 |

| CUDA 13.0 VRAM: 128GB unified |

+-------------------------------------------------------------------------+

$ python3 -c "import torch; print(torch.cuda.get_device_name(0))"

NVIDIA Blackwell GPU

$ _

How it works — rent GPU in 4 steps

Create your account

Add a payment method

Add your card securely via Stripe at /billing. You only pay for what you use.

Launch your instance

Choose on-demand or interruptible. Your container starts in under 60 seconds.

SSH in and compute

SSH directly into your container. Bring PyTorch, TensorFlow, JAX — anything.

Pricing

On-demand

£0.30/hr

Pay per second. No commitment. Full GPU access. Cancel anytime.

→ NVIDIA Blackwell GPU 128GB
→ SSH container access
→ Docker with full GPU passthrough
→ Billed per second

Get started

Popular

Interruptible

£0.21/hr

30% cheaper. Best for batch jobs, training runs, and experiments.

→ NVIDIA Blackwell GPU 128GB
→ SSH container access
→ Docker with full GPU passthrough
→ Billed per second

Get started

Host your GPU

77% earnings

List your idle GPU and earn from every rental automatically via Stripe.

→ Any NVIDIA GPU 8GB+ VRAM
→ Earn 77% of every rental
→ Stripe direct bank payout
→ No upfront cost

Start hosting

The open GPUcloud for AI teamsand enterprises

Run a real LLM onNVIDIA Blackwell GPUin 60 seconds