Why it expresses the thesis
The highest-quality expression of the front-end drug-discovery bet: Alphabet/DeepMind lineage, an AlphaFold-derived platform, 17 active programs across oncology, immunology, and cardiovascular disease, and a milestone-payment model that de-risks revenue versus a single-asset biotech. The first AI-designed cancer candidate is targeting Phase I by end-2026.
Why conviction is “watch,” not “high”
This is sized strictly as an option, not a core holding. 2026 is the year of binary clinical tests, and additional failures are statistically likely given historical attrition. AlphaFold-class models still have blind spots — membrane proteins, disordered regions, novel chemotypes — and experimental validation remains the ultimate arbiter. Entering human trials is the real test; getting there is not proof of efficacy.
What to diligence
- Validate the partnership / milestone revenue base independent of the internal pipeline.
- Scrutinize platform blind spots and where the model’s predictions are weakest.
- Demand clarity on what an early clinical readout actually proves versus signals.
- Entry only if access and price reflect the binary clinical risk ahead.
Pair any exposure here against recurring-revenue platform names so the sleeve survives a clinical failure.