$Alphabet(GOOG)$ AI drug discovery seems to have moved past the hype phase. The real shift now is generative foundation models and de novo molecular design going from theoretical concepts into actual clinical trials.

The first wave of fully AI-discovered therapeutics is moving through pivotal Phase II and III human trials, and the timelines tell the story — preclinical pipelines that used to take years are being compressed to months.

On the de novo side, platforms are building entirely new chemical compounds and molecular glues from scratch, not just tweaking old drug templates. Multi-agent software frameworks and automated robotic labs are running continuous design-make-test-analyze cycles with minimal human input.

What stands out is the early Phase I and II success rates for AI-designed molecules. They're showing historical improvements, scaling past traditional trial benchmarks.