Company Deep-Dive

Isomorphic Labs watch

Alphabet/DeepMind spinout building on AlphaFold; 17 programs across oncology, immunology, and cardiovascular, with the first AI-designed cancer candidate targeting Phase I by end-2026.

Updated May 28, 2026
Stage
venture
Valuation
~$2.5B (raising)
Revenue
Milestone-based
Multiple
n/a (private)
Conviction
watch
Thesis Fit

The convex drug-discovery option — sized as a venture call, not a core position, underwritten only against de-risked clinical readouts.

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.

Sources

Independent analysis for discussion only — not investment advice. Figures are from public sources, time-sensitive, and may be outdated or inaccurate. Private-company valuations are estimates that vary by source.