Pathos Dependence on Tempus
Pathos
This partnership gives Pathos speed, but it also puts a core input for its drug engine in a partner’s hands. Pathos is building around multimodal oncology data and machine learning for trial design and patient selection, while Tempus already sells the data, diagnostics, trial matching, and AI tools that sit next to that workflow. As Tempus adds more of the stack, Pathos gets more capable, but also less insulated.
-
Tempus is not just a data vendor. It has moved into real time patient finding through Deep 6 AI, which added access across 750 plus provider sites and 30 million plus patients, and into longitudinal cancer monitoring through its July 20, 2026 agreement to acquire Personalis.
-
That matters because these products touch the exact points where drug developers create leverage, finding eligible patients, linking clinical and molecular records, and tracking response over time. A partner that controls those pipes can stay a tools company, or push upward into shared economics on programs built on top of them.
-
Roche already shows what a durable oncology evidence stack looks like in practice. Flatiron brings 5M plus research ready patient records, 15 plus years of longitudinal data, global regulatory usage, and a clinico genomic database built with Foundation Medicine, making it deeply embedded in biomarker work and regulatory grade evidence generation.
The next phase is a race to own the layer where data becomes drug decisions. If Tempus keeps expanding from diagnostics and software into broader oncology program economics, Pathos will need its own durable edge in model performance, asset creation, or clinical execution, because data access alone will not stay proprietary for long.