From Faster Discovery to Better Evidence: A Staged Translational Research Programme for Kenya

Artificial intelligence can increase the number of candidate molecules and hypotheses that a drug programme must consider. If it does, the constraint may move from generating candidates to validating them experimentally and producing evidence that sponsors and regulators accept. This paper asks whether a translational laboratory in Kenya makes sense and, if so, in what order it should be built. Primate study demand is not proportional to candidate volume. We decompose it as candidates × progression × residual need × award share × studies per candidate, so a rise in candidates can leave demand unchanged or lower it. We review the regulatory signals on reducing animal testing in the United States and Europe, and the scientific limits of primate models by modality. The proposed programme is staged and puts non-animal methods first. It commissions primate work through qualified partners only where a written scientific justification exists, and it treats a breeding colony as an option that must compete with purchased services. The programme states four falsifiable hypotheses (H1 to H4), including a liver microphysiological screening service evaluated on a locked external holdout. Its gates are scientific and ethical decision points. The paper reports no original experimental results, no regulatory authorisation and no executed partnerships. Its regulatory and legal statements are feasibility statements, not legal advice.