AI Model Speeds Drug Discovery 10,000 Times Faster

A groundbreaking AI model can now predict in seconds what used to take months of supercomputer calculations, bringing hope that new medicines could reach sick patients years sooner.

Researchers at Chalmers University of Technology and the University of Gothenburg in Sweden have created artificial intelligence that speeds up molecular simulations more than 10,000 times faster than traditional methods. The technology could transform how pharmaceutical companies identify promising new drugs.

Developing a new medicine typically takes over ten years and costs billions of dollars. Much of that time gets spent in early testing, where scientists must screen thousands of molecules to find the rare few worth pursuing further.

The traditional approach simulates molecules step by tiny step, calculating forces between atoms at intervals of one femtosecond (a millionth of a billionth of a second). Because drug-related processes unfold over much longer timescales, these simulations require billions of calculations and massive computing power.

The new AI model learns the underlying rules of how molecules move and change over time. After training on examples of molecular behavior, it can predict how new molecules will evolve without performing those intensive calculations.

"What sets our AI model apart is that it learns the underlying dynamics over longer time scales," says Simon Olsson, the research leader and associate professor who led the study published in Science Advances. The model doesn't just show what shapes molecules take, but also how quickly they change and through which pathways.

The researchers tested their system on over 12,500 organic molecules and more than a thousand peptides (short chains of amino acids). The AI accurately predicted molecular behavior consistent with the laws of physics, even for molecules it had never seen during training.

Think of it like jumping between scenes in a molecular movie instead of watching every single frame. The AI observes what happens over nanoseconds but can predict changes occurring over periods a thousand times longer.

The Ripple Effect

The breakthrough extends far beyond faster computers. Quicker, more accurate drug screening means pharmaceutical companies can identify promising candidates earlier and eliminate dead ends sooner, saving both money and precious time.

For patients waiting for treatments for cancer, Alzheimer's, rare diseases, and other conditions, years shaved off development timelines could mean the difference between life and death. Lower development costs could also make medicines more affordable and encourage research into diseases that affect smaller populations.

Juan Viguera Diez, an industrial doctoral student at AstraZeneca who contributed to the research, emphasizes the significance: demonstrating that AI can predict physical phenomena by understanding underlying physics represents a genuine scientific advance, not just a computational trick.

The researchers are already looking ahead to laboratory applications, where they'll measure specific molecular properties like how molecules interact with cell membranes. Though that work still lies in the future, the foundation is now in place.

This AI doesn't just make drug discovery faster—it makes hope move faster too.