Tempus and Daiichi Sankyo Push AI Upstream Into ADC Design

Tempus’ collaboration with Daiichi Sankyo is notable less for the fact of an AI partnership than for where it is aimed: antibody-drug conjugates, a category that has become strategically central in oncology. ADCs are attractive because they promise targeted delivery of potent payloads, but they are also notoriously difficult to optimize across biology, chemistry, tolerability, and patient selection. That makes them a natural proving ground for multimodal AI.

The real opportunity is not simply predicting which targets look interesting. In ADC development, value is created by integrating heterogeneous data: tumor expression, internalization behavior, linker stability, payload potency, toxicity signals, and clinical phenotype. If Tempus can bring its data and modeling capabilities to those intertwined decisions, AI could become part of the architecture of ADC programs rather than a downstream analytics layer.

For Daiichi Sankyo, the partnership reinforces a broader industry pattern: large drugmakers are no longer treating AI as a general innovation talking point. They are applying it to high-value platforms where cycle time, trial design, and portfolio prioritization have direct commercial implications. Oncology remains the easiest domain in which to justify that investment because the datasets are comparatively rich and the clinical economics are compelling.

The strategic question is whether these collaborations generate differentiated compounds or simply faster filtering. Even if AI mainly improves attrition management rather than inventing radically new ADCs, that would still be meaningful in a field where small design missteps can derail entire programs. The bar for success is no longer whether AI can analyze ADC data, but whether it can measurably improve development decisions before the clinic reveals the answer the hard way.