Where can AI have the largest impact in scientific discovery?

In 2015, three MIT postdoctoral physicists met over pizza for a late-night brainstorming session. They saw untapped potential in quantum physics for molecular discovery—and began with a clear ambition: to reinvent crystal structure prediction (CSP) and transform how science finds new solid forms.

That foundation proved decisive. XtalPi’s CSP platform demonstrated industry-leading accuracy in a Pfizer blind test, and later supported a pivotal collaboration that helped bring Paxlovid to market during the COVID-19 pandemic. From solid-state research, we expanded into AI-powered drug discovery—first across small molecules, then into antibodies, peptides, and other biologics—building the Scientific AI capabilities to design, predict, and optimize candidates end to end.