Large Language Models (LLM) are increasingly being used and are becoming a decisive competitive factor. Companies are particularly concerned with how they can deploy AI models quickly and, above all, securely. In response to this development, T-Systems now offers AI Foundation Services. These services provide companies with a private and secure development, testing, and production environment that meets the strictest data security and privacy requirements. AI models and services are pre-configured, ready for immediate use, and form the basis for enterprise-wide scaling of AI solutions across different business units and regions. The platform meets regulatory requirements and security standards, enabling safe use even in sensitive areas such as healthcare, defense, or public administration.
In addition to a continuously growing portfolio of currently over 15 Large Language Models – including the latest open-source models like Meta Llama 3.3, Mistral Small 3, or DeepSeek R1 – pre-configured services for Retrieval-Augmented Generation (RAG) and model fine-tuning are included. This allows AI solutions to be expanded with company-specific data sources to generate more accurate, context-based answers and precisely adapt to internal requirements. Additionally, machine learning operations, data management, and various deployment options (public, private cloud, on-premises) are available. The broad portfolio of models and services and the associated flexibility also help customers avoid dependency on a single technology provider or company.
“We make AI easy and secure to use: With the AI Foundation Services on our secure and sovereign cloud platforms, we offer our customers an effective use of LLMs – even in sensitive corporate environments,” says Dr. Ferri Abolhassan, CEO T-Systems and Member of the Board of Management Deutsche Telekom AG. “In this way, we offer companies from all industries the opportunity to use AI innovations without jeopardizing their independence and sovereignty. It would be hard to achieve greater sovereignty in the use of AI.”
The AI Foundation Services are provided on the Open Telekom Cloud and are thus subject to the sovereign, secure, and legally compliant regulations of the European Union. GDPR compliance and data sovereignty are ensured. For regulated industries such as healthcare and the public sector, the provision of Large Language Models from the Foundation Services ensures industry-specific compliance requirements. All models and services are operated in Telekom data centers in Germany, Switzerland, and the Netherlands, and can alternatively be deployed on customer-owned infrastructure. Popular AI models from OpenAI and Anthropic are also optionally available through T-Systems partners like Microsoft, AWS, or Google Cloud – in the same interface as the open-source models. Before commissioning, all services undergo a thorough security scan and a product security screening by Telekom according to ISO/IEC 27001. The AI models can be accessed directly via the Open Telekom Cloud Marketplace and tested immediately.