Local AI, GDPR and the EU AI Act: what people ask us most

It is artificial intelligence that runs on your own server, inside your company, without sending data to the cloud or to third parties. Your documents and queries never leave your infrastructure, which gives you full control over your information and makes GDPR compliance easier.

With ChatGPT and similar services, your data travels to external servers, often outside the European Union. With an on-premise solution, open-source models run locally: you decide who has access, you do not depend on per-use fees, and your confidential information never leaves the company.

Because the data is processed entirely on your premises, you avoid transfers to third parties and to countries outside the EU, one of the most sensitive points of the GDPR. You keep traceability and control over the processing. We help you integrate it into your data protection measures.

Yes. Since 2 February 2025, every organisation that uses AI systems must ensure a sufficient level of competence among its staff to operate them (a duty of means, not of result). Authorities can enforce compliance from 2 August 2026. Art. 4 has no specific penalty attached, but it is part of the obligations of the regulation. We offer certified training for your team.

It is especially useful for organisations that handle sensitive or confidential data: clinics and pharmaceutical companies, professional firms, industry, and any SME that wants to use AI without exposing its information. We adapt it to your use cases and your workload.

There are two routes. A monthly package with no minimum term from 2,500 € per month, with hardware on loan, tools and support included. Or a direct purchase of the infrastructure from 18,000 € as a one-off payment, with maintenance contracted separately. In the Madrid region it can be combined with public grants such as Kit Digital, Kit Consulting and the Cheque Innovación.

The system is up and running in about 10 days: installation of hardware and models in the first days, definition of the team and the highest-impact use cases, certified staff training and production launch. All without interrupting daily work.

The platform is hardware-independent. The open-source models run on any infrastructure: from a Mac Studio for a pilot team to servers with NVIDIA GPUs or several nodes for hundreds of concurrent users. To start quickly in a pilot or small team we recommend the Mac Studio for its price-performance ratio and immediate installation. As the number of users or the workload grows, we scale the hardware up without changing the way you work. It is a question of budget, not a limit of the platform.