What Aleph Alpha does
Aleph Alpha (Aleph Alpha GmbH) builds sovereign, enterprise-grade specialized large language models (SLLMs) for regulated and data-sensitive use cases, positioned around “sovereignty, transparency, and responsibility” and deployed on European infrastructure. The company designs and deploys domain-trained models with custom evaluation frameworks intended to prove performance before models go into production, and it emphasizes integration with customer workflows rather than standalone chat experiences.
Aleph Alpha’s current product direction is organized around its PhariaAI stack (including the PhariaAssistant and supporting platform components). The stack is presented as end-to-end “from inference to infrastructure” and is used by customers to run confidential, production workloads with their own data and intellectual property.
In addition to deployable model offerings, Aleph Alpha publishes parts of its engineering approach and research methods (for example, releases around model training and scaling codebases under an Open Aleph License for non-commercial research, and engineering blogs describing its “Model Training as Code” approach via Savanna).
Strategically, Aleph Alpha’s “sovereign Europe” positioning has intersected with major industry and cloud-adjacent partners; it publicly lists customers such as SAP, Schwarz Group, Bosch, Deutsche Bank, Infineon, Hella, and others, and highlights government-sector deployments such as an AI assistant rolled out in a government agency.
Ownership has changed recently: in April 2026, Cohere and Aleph Alpha announced that they would “join forces,” with Cohere acquiring Aleph Alpha, and with a commitment from Schwarz Group for structured financing tied to Cohere’s upcoming Series E.