Wall Street remains skeptical about AI's ability to transform drug discovery economics despite tangible lab efficiencies. The technology has accelerated protein folding prediction, target identification, and virtual molecular screening — tasks that traditionally consumed months of manual hypothesis testing.

At Roche's Genentech, computational biologist Aviv Regev has implemented a "lab in the loop" system where AI models propose targets and molecules, researchers validate them experimentally, and results feed back into the models. Regev emphasizes that AI's advantage isn't superior reasoning but its capacity to synthesize vastly more biological knowledge than any individual scientist across disease areas like Alzheimer's and cancer.

Yet investors have little incentive to revalue drugmakers the way they've rewarded AI infrastructure companies until the technology demonstrably alters the cost and failure rates that define pharmaceutical economics. The sector's high attrition and decade-long timelines remain the ultimate test.

The market's patience is finite. If AI-driven programs don't yield clinical candidates with meaningfully better success rates within the next 18-24 months, the current valuation premium for "AI-native" biotechs will compress, separating genuine platform advantages from marketing narratives.