capexScore 35/100Research

DeepMind's AlphaFold 2 required 32 component algorithms over 5 years with 20-person multidisciplinary team (biologists, physicists, chemists, ML engineers); scientific AI requires deep domain expertise

Demis Hassabis· Google DeepMind· AI· 2026-07-18· about DeepMind (GOOGL)
the other thing we did which was a nice innovation for for alpha fold was was have the system predict its own confidence in its own predictions... a huge research effort over sort of five years took about 20 people at its maximum and it was a truly multidisciplinary effort so we needed biologists and physicists and chemists as well as machine learners

Why it matters

AlphaFold 2's development required 5 years, 20 FTE peak, and deep multidisciplinary expertise. This sets a high bar for R&D in scientific AI and suggests significant capital and talent concentration requirements.

Investment implication

Companies attempting to compete in scientific AI (drug discovery, materials, climate) should expect similar R&D intensities (5+ years, large multidisciplinary teams). Expect consolidation around firms with sufficient capital and talent. Smaller biotech may depend on partnerships with well-capitalized AI labs.

Source

Using AI to Accelerate Scientific Discovery - by DeepMind's Demis Hassabis (YouTube)
← All signals