Pharmaceutical companies like Roche are embedding AI across drug development and diagnostics, shifting how medicines are discovered and produced—enabled by powerful computing hardware that has only recently made decades-old mathematical models computationally feasible.

What happened

Roche, the Swiss pharmaceutical company, has trained approximately 100,000 employees across the company in AI handling and is now using AI in many activities at its Penzberg site and beyond. The company stores data in a secure cloud area with Google and equips most employees with laptops and digital tools for AI-enabled work.

Why it matters

The mathematics underlying biological and pharmaceutical models date from the 1950s and '60s but were too computationally expensive to run until after 2010, when GPUs (graphics cards specialized for performing many simple calculations simultaneously) became widely available. This technological breakthrough has unlocked the ability to apply these models at scale, fundamentally changing how drug developers operate.

What to watch

Drug developers are now using AI at nearly every stage of development to make processes faster and more efficient and to discover new active ingredients. At the same time, technology companies are increasingly entering drug development using AI, which means the rules of the game in drug development are changing significantly.