Frequently Asked Questions
Answers to common questions about the Holobrain Private AI Stack—from deployment and data sovereignty to costs, models, and timelines.
Are the Private AI stack LLMs installed on servers provided by Holobrain?
No, the system is installed exclusively on your own servers to guarantee complete data sovereignty. You retain full ownership of the hardware, the data, and the deployment. You can choose to host the infrastructure in your own data center, a colocation facility, or with a trusted dedicated server provider. Regardless of your choice, we will guide you through the end-to-end hardware onboarding process.
What specific LLMs are installed?
We select, configure, and install top-tier open-weights foundation models tailored to your specific use cases and available hardware. We utilize all major open-weights model families, including Llama, Qwen, Mistral, Gemma, and Falcon. Depending on your business requirements, we deploy various kinds of LLMs:
- Large-scale text generation models for complex reasoning, coding, and writing.
- Embedding models for highly accurate RAG (Retrieval-Augmented Generation) and enterprise search.
- Multi-modal and vision models (image-to-text) for processing and analyzing visual data.
These serve as the core reasoning engines of your system, keeping all data processing entirely in-house while providing capabilities comparable to major cloud AI platforms.
What are open-weights LLM models and why use them?
Open-weights Large Language Models (LLMs) are highly advanced AI models whose core architecture and computational weights are publicly accessible. Major cloud AI platforms use proprietary models that require you to send your sensitive prompts and company data over the internet, creating critical security risks. Open-weights models, by contrast, can be deployed directly on your own infrastructure. This eliminates the risk of data leakage and vendor lock-in, ensuring 100% data privacy and operational control.
Can the system be integrated with our existing authentication systems?
Yes. The Private AI Stack is built on flexible, enterprise-grade infrastructure and can integrate seamlessly with your organization’s existing identity management and authentication systems, such as Active Directory, LDAP, OAuth, or SAML.
How long does the project implementation take?
Our systems are typically deployed in weeks, not months. The overall timeline depends primarily on the initial phases: needs assessment, hardware procurement, and integration with your existing data sources. For an organization of about 100 employees, these preparatory steps generally take about one month. Once the hardware is provisioned, the actual infrastructure deployment is highly automated and completed within hours, followed by rigorous testing, AI evaluations, and team training.
Still have questions?