Services
AI infrastructure, from Spain for Europe
On-demand NVIDIA GPUs to train and experiment, and dedicated inference to serve open-source models in production. All on European infrastructure.
On-demand GPUs
Whole NVIDIA GPUs in seconds, to train, fine-tune or experiment. Pay by the minute and release them when you're done.
- A100, H100, H200, L40 and B300
- Per-minute billing
- SSH and JupyterLab
- Zero egress fees
Dedicated inference
We deploy and run the most requested open-source models, or your own fine-tune, on an endpoint with GPUs just for you.
- GPUs reserved for you
- OpenAI-compatible API
- Your model or your fine-tune
- Data stays in the EU
Clodei for business
The same GPUs with terms built for teams and companies: monthly invoicing and direct debit.
- NET-30 payment terms
- SEPA B2B direct debit
- EU VAT (VIES)
- Monthly invoice
Models we serve
- Text and reasoning
- DeepSeek V4 ProDeepSeek
- DeepSeek V4 FlashDeepSeek
- Qwen3.6-35B-A3BQwen
- Qwen3.6-27BQwen
- Kimi K3Moonshot AI
- GLM-5.2Z.ai
- gpt-oss-120bOpenAI
- gpt-oss-20bOpenAI
- Llama 4 MaverickMeta
- Mistral Large 3Mistral AI
- Gemma 4 31BGoogle
- Nemotron 3 UltraNVIDIA
- Code
- Qwen3-Coder 480BQwen
- Devstral 2Mistral AI
- Vision
- Qwen3-VL 235BQwen
- Llama 4 ScoutMeta
- Embeddings and reranking
- Qwen3-Embedding 8BQwen
- Qwen3-Reranker 8BQwen
- BGE-M3BAAI
- Audio
- Whisper large-v3 turboOpenAI
- Voxtral SmallMistral AI
- Image and video
- FLUX.2 [dev]Black Forest Labs
- Qwen-ImageQwen
- Wan 2.2 A14BAlibaba
Don't see yours? We deploy any open-weights model supported by vLLM or SGLang.
Shall we talk about your project?
Tell us what you need and we'll suggest GPUs, model and deployment.
Contact us