How much does it cost to run AI yourself?
Hardware, monthly maintenance and power. Total costs depend on workload and team size.
Frequently Asked Questions
Answers to common questions about our hardware cost calculator and on-premise AI solutions
Is On-Premise AI worthwhile for SMEs?
Yes, especially with compliance requirements, Swiss data sovereignty and strict data protection.
Which data should never go to the cloud?
Sensitive business data, personal and compliance-relevant data. On-Premise keeps everything under your control.
For which company sizes is On-Premise AI worthwhile?
From small teams to large enterprises. Often economical from approx. 5 users.
Can ERP, CRM and document data be integrated?
Yes – interface-compliant and with minimal integration effort.
How can I get a quote?
Via the purchase page or contact form. We're happy to help with demos and decision support.
Can I evaluate On-Premise AI before purchase?
Yes – fast, realistic and interface-compliant.
How do maintenance and updates work?
Planned updates, maintenance windows, optional zero-downtime. Security patches promptly.
How performant is On-Premise AI?
Depends on the model – in practice up to 4× faster.
What's included in the monthly maintenance?
Operations, monitoring, support – and always the latest AI models. No token fees.
Which operational model is used?
Managed Service, Co-Management or Self-Service. Depending on your needs.
How does the purchase process work?
We build after purchase – Time-to-Market: approx. 3 weeks.
Can installation and maintenance be done remotely or must it be on-site?
Usually remote. On-site for installation or when needed.
How does the On-Premise AI solution scale?
Horizontally via additional units, vertically via more powerful hardware. Flexibly expandable.
Is your team based in Switzerland?
Yes. Our team is fully based in Switzerland.
Can I start small and scale later?
Yes. Start with SMALL and scale with MEDIUM or LARGE – our Kubernetes solution makes it easy.
What are typical use cases for SMEs?
From documents to code to internal assistants – data-protection compliant and high-volume.
Why only AMD and NVIDIA hardware options?
AMD Ryzen AI Max+ and NVIDIA RTX Pro 6000 offer the necessary memory bandwidth and capacity for professional LLM inference.