Apheris develops a platform enabling pharmaceutical and life-sciences organizations to train artificial-intelligence models collaboratively on distributed datasets without centralizing proprietary information. Founded in 2019, the company addresses a structural constraint in drug discovery: the inability to pool data across competitors for better AI training whilst maintaining confidentiality and data sovereignty.
The core product, Apheris Gateway, allows organizations to participate in federated data networks whilst retaining complete control over sensitive datasets. The platform supports secure local applications for drug-discovery workflows, including structural-biology models and ADMET prediction tools that operate without exposing underlying data. This approach contrasts with traditional centralised data repositories, which pharmaceutical companies have largely rejected due to privacy, security and intellectual-property concerns.
Apheris powers the AI Structural Biology Network, an industry-wide federated collaboration that includes eight top-20 pharmaceutical companies - among them AbbVie, AstraZeneca and Johnson & Johnson. The network demonstrates the commercial viability of the federated-learning model in high-stakes pharmaceutical research, where data sensitivity and competitive dynamics make conventional data-sharing untenable.