AI Models Take on Medicine’s Rarest Challenges

AI is speeding up rare disease research and reshaping drug discovery worldwide

8 Dec 2025

Artificial intelligence is beginning to reshape rare disease drug discovery, offering new momentum in a field long constrained by high costs, small patient populations and slow development cycles.

More than 7,000 rare diseases still lack an approved treatment, making the sector a testing ground for new research approaches. Companies such as Excelsior Sciences and Healx are applying AI to drug discovery and development, aiming to shorten timelines and reduce risk in areas that have struggled to attract sustained investment.

Excelsior Sciences said this year that AI now sits at the centre of its discovery workflow, particularly in early-stage chemistry. The company has pointed to significantly faster internal timelines, although it has not publicly confirmed that a rare disease therapy has been discovered solely through its platform. Even so, its claims have prompted wider interest across the industry.

By allowing computer models to assess large numbers of potential drug candidates, AI systems can help more ideas move into early development. Analysts say this could encourage other companies to reassess traditional research models that rely heavily on slower, laboratory-based screening.

Healx has reported similar early traction. The company focuses on using AI to repurpose existing medicines for rare diseases, an approach designed to lower development costs and safety risks. It has seen rising interest from international research groups looking to compress development cycles and improve the odds of success.

Academic institutions are also exploring AI’s role. Researchers at Vanderbilt University Medical Center have shown that machine learning tools can identify subtle patterns in clinical records, helping clinicians spot certain rare diseases earlier and refer patients more quickly for testing and treatment.

Despite the growing interest, experts caution that the technology is still at an early stage. Limited datasets, small trial populations and evolving regulatory expectations remain significant challenges. Regulators and clinicians have also raised concerns about transparency and explainability in algorithmic decision-making.

Even so, investors expect increased partnerships and acquisitions as AI tools mature. Many analysts see the coming decade as a period of gradual but meaningful progress, with technology expanding the range of rare diseases that companies are willing to pursue.