Apheris is a platform for federated and privacy-preserving machine learning, built around Apheris Foundry. Federated networks let pharmaceutical companies and other data holders train AI models on distributed, sensitive data without pooling the raw data. The current offering focuses on drug discovery, with networks for structural biology, ADMET properties, antibody developability and real-world evidence. Models are delivered into the customer's own environment, where all computations take place.
The company behind the platform is Apheris AI GmbH, with an address in Berlin, Germany, so it is established in the EU. The website names no investors, does not state its headquarters explicitly and names no hosting provider for the platform itself. Because computations run inside the customer's environment and the company says raw data is never pooled, copied or transferred, the product resembles self-operated software more than a vendor cloud service. For a European organisation, the contractual terms for model access and the customer's own infrastructure matter most. The website does not state whether the software is open source.