AI Is Rewriting Drug Discovery. The First Results Are In.
In October 2025, a drug designed entirely by artificial intelligence entered Phase II clinical trials for the first time. The molecule, developed by Insilico Medicine for idiopathic pulmonary fibrosis, had been identified, designed, and optimized by AI systems in 18 months: a process that traditionally takes four to five years. If it works, it won't just be a new treatment. It will be proof that AI can fundamentally accelerate how we develop medicine.
The statistics are brutal. It takes an average of 12 to 15 years and $2.6 billion to bring a single new drug to market. Over 90% of drug candidates that enter clinical trials fail. The process involves searching a chemical space of approximately 10^60 possible drug-like molecules: a number so large it makes the number of atoms in the universe look manageable.
Traditional drug discovery is essentially educated guessing: identify a biological target, screen thousands of compounds against it, optimize the best candidates through years of iterative chemistry, and hope that what works in a test tube also works in humans. It's expensive, slow, and has a staggering failure rate.
AI attacks drug discovery from multiple angles simultaneously:
Target identification: AI models analyze genomic data, protein interactions, and disease pathways to identify which biological targets are most likely to be relevant. DeepMind's AlphaFold provides the 3D structures these targets need for drug design.
Molecule generation: Generative AI can design novel molecules with specific properties: the right shape to bind a target, the right solubility to be absorbed, the right stability to survive the digestive system. Instead of screening existing libraries, AI creates bespoke molecules from scratch.
Please enable JavaScript to read the full article.