innFactory AI Consulting based in Rosenheim, Germany supports enterprises across the DACH region with GDPR-compliant self-hosting of Meta Llama. With open weights you have full control - no data leaves your infrastructure.
Key Strengths
Open Weights (Community License)
- Full Control: Model runs in your infrastructure
- No API Costs: Only hardware/cloud costs
- Customizable: Fine-tuning on your own data possible
- GDPR-Friendly: No data leaves your company
Flexible Deployment Options
- On-Premise: Own servers or private cloud
- Edge: Local devices, smartphones
- Cloud: AWS, Azure, GCP with own instance
Hardware Requirements
| Model | VRAM | Recommended GPU |
|---|---|---|
| Llama 4 Scout | 80+ GB | H100 / A100 |
| Llama 4 Maverick | 400+ GB | Multi-H100 |
| Llama 3.3 70B | 40+ GB | A100 80GB |
| Llama 3.2 11B | 24 GB | RTX 4090 |
| Llama 3.2 3B | 8 GB | RTX 4070 |
| Llama 3.2 1B | 4 GB | Smartphone |
Integration with CompanyGPT
CompanyGPT supports Llama models and enables completely self-hosted operation without external dependencies.
Our Recommendation
Llama 4 Scout is ideal for companies with strict data protection requirements or high request volumes. Investment in own infrastructure pays off from medium usage volumes. For smaller deployments or edge applications, the compact Llama 3.2 (1B/3B) models are a cost-effective choice.