Knowledge management: your corporate knowledge, queryable

Companies face the challenge of capturing, organising and providing knowledge efficiently. AI automates these processes, speeds up access to relevant information and prevents knowledge loss. It thus increases productivity and improves decision-making.

A cloud GPU is a graphics processing unit provided via the cloud that lets companies run compute-intensive tasks without their own physical hardware. Instead of running an expensive GPU locally, users access powerful computing capacity over the internet – at centron, a transparently communicated hourly usage fee applies. Using a cloud GPU is particularly advantageous for machine learning, 3D modelling, simulations, complex data analysis and the inference of large language models, as the computing power required scales flexibly and is used as needed.

An ingest pipeline regularly reads from sources, parses documents and updates the embeddings – new documents can be queried after the next run. Everything runs within your VPC; deleted sources are also removed from the index.

Configuration I: 8 cores, RTX A4000 (16 GB VRAM), 16 GB RAM, starting at, for example, with 500 GB storage – with inference: €146.23 per month (approx. €0.2031 per hour), without inference: €53.64 per month (approx. €0.0745 per hour). Configuration II: 16 cores, A100 (40 GB VRAM), 64 GB RAM, starting at, for example, with 500 GB storage – with inference: €602.79 per month (approx. €0.8372 per hour), without inference: €113.32 per month (approx. €0.1574 per hour). For a bespoke solution, please contact our sales team.

The starting prices are deliberately low: Cloud GPU from €92.59 per month, volumes from €0.05 per GB per month and S3 Object Storage from €5.00 per month. New accounts receive a €200 starting credit valid for 60 days – you can calculate the cost of your specific configuration transparently using the price calculator.

Yes – and that is the critical issue with AI-supported knowledge management: internal documents regularly contain personal data and trade secrets, which could fall into the wrong hands if an external AI service were used. With centron, the knowledge base, embeddings and model remain in Germany. The data centres are certified to ISO 27001 on the basis of IT-Grundschutz, with an unrestricted BSI C5:2020 Type 1 attestation for ccloud³ / Managed Cloud. Further details can be found in the Trust Centre.

AI-supported knowledge management frequently works with sensitive internal documents, processes and company data. The audit under the BSI C5:2020 Type 1 attestation confirms that our cloud has been audited against the BSI’s established criteria catalogue for secure cloud services. This gives companies a traceable basis for operating AI workloads with high requirements for security, transparency and compliance.

Object storage is ideal for AI-supported knowledge systems with large datasets, as it processes big data flexibly, scalably and cost-efficiently. Training data, documents or databases can be stored without fixed folder structures and retrieved at any time. This makes object storage perfect for workloads that grow continuously. Compared with classic file systems, object storage enables particularly high resilience and simple integration into AI frameworks and pipelines. With centron S3 Object Storage, companies benefit from a GDPR-compliant, highly available storage solution with high-performance connectivity, designed for modern AI processes.

AI-supported knowledge management systems offer new employees personalised training content, provide relevant information and speed up onboarding. Through continuous learning and automated updates, corporate knowledge always stays up to date.

centron relies on a fair, transparent pricing model with a strong price-performance ratio: no hidden costs – you pay only for the performance used – and flexibility to adjust your GPU use to your needs at any time; all terms in the GPU pricing.

Yes, you can rent various server solutions from centron that fit your requirements: virtual machines, quickly configurable and flexibly scalable, and managed servers with maintenance and operation by centron – you focus entirely on your work.