AI Drug Repurposing 2026: Why Old Medications Are Suddenly Curing New Diseases

I’ve been tracking the pharmaceutical industry for years, and what I’m witnessing in 2026 is nothing short of a revolution. For decades, the process of bringing a new drug to market has been agonizingly slow and astronomically expensive, often taking over a decade and costing billions. But now, artificial intelligence is rewriting the rules, not by inventing entirely new molecules from scratch, but by uncovering hidden potential in medications we already have. I believe this shift, known as AI drug repurposing, is one of the most valuable health insights people need to know right now.

My research shows that traditional drug discovery is a daunting journey, typically spanning 10 to 17 years with an average cost of $2 billion to $3 billion per approved therapy. The success rate from Phase I trials is a dismal 11%. This financial and temporal burden means countless promising treatments never see the light of day, especially for rare or complex diseases with smaller patient populations. However, AI drug repurposing is fundamentally changing this calculus. It’s identifying new therapeutic uses for existing approved or investigational drugs, leveraging their known safety profiles to dramatically cut down development timelines and costs. I've found that repurposing can reduce development timelines to a mere 3 to 12 years, with average costs plummeting to $300 million and success rates approaching 30% from Phase I. This isn't just an incremental improvement; it's a paradigm shift that makes previously unattainable therapies suddenly viable.

The AI Advantage: Speed, Cost, and Unseen Connections

I’ve observed that the core power of AI in drug repurposing lies in its ability to process and analyze vast, complex datasets at speeds impossible for human researchers. Imagine sifting through millions of scientific papers, clinical trial results, genetic data, and patient records to find subtle connections between a known drug and an unaddressed disease. That's precisely what machine learning algorithms, deep learning, network biology, and knowledge graphs are doing. They model intricate interactions between genes, proteins, and drugs, learning high-level representations from integrated genomic, proteomic, and clinical data to identify novel drug-disease associations. This capability isn't just about efficiency; it's about seeing patterns that have been hidden in plain sight for years, or even decades.

The economic implications are staggering. The global market for AI in drug repurposing alone was valued at $1.3 billion in 2025 and is projected to skyrocket to $7.7 billion by 2033, demonstrating a remarkable compound annual growth rate (CAGR) of 24.5% from 2026 to 2033. This growth is fueled by the relentless pressure on pharmaceutical companies to find more cost-effective development strategies, coupled with the rising prevalence of rare and complex diseases. The broader AI in pharmaceuticals market tells a similar story, expected to grow from $1.97 billion in 2025 to an astounding $21.51 billion by 2035. I believe this rapid expansion underscores the profound value AI is delivering across the entire pharmaceutical pipeline.

Real-World Triumphs: From Pandemics to Rare Diseases

My research has highlighted several compelling examples of AI drug repurposing already delivering tangible results. One of the most striking occurred during the desperate early days of the COVID-19 pandemic. In January 2020, BenevolentAI, a London-based AI drug discovery company, leveraged its proprietary AI platform to scour its massive biomedical knowledge graph. Within just 48 hours of computation, their system flagged baricitinib, a rheumatoid arthritis drug, as a potential treatment for COVID-19. This incredibly rapid identification, followed by human expert validation and clinical trials, led to an FDA Emergency Use Authorization, making the drug available to millions of patients years earlier than traditional repurposing timelines would have allowed. I find this case particularly compelling because it demonstrates AI's ability to respond to urgent global health crises with unprecedented speed.

Beyond pandemics, AI is actively addressing other significant medical challenges. I've noted that BenevolentAI is also collaborating with DNDi (Drugs for Neglected Diseases initiative) since 2022 to identify drug repurposing candidates for severe dengue, a disease affecting millions worldwide. Furthermore, for the 30 million Americans suffering from rare conditions with no current therapies, the U.S. government's ARPA-H launched its MATRIX project, which received $48 million in funding to use AI to predict how existing drugs can be effective for these diseases. This initiative, along with companies like Insilico Medicine developing advanced AI frameworks such as TargetPro–TargetBench, which has identified that 46% of its predicted novel drug targets are associated with approved drugs in other indications, underscores the vast potential for repurposing to unlock treatments for underserved patient populations. My findings suggest that this focus on rare and complex diseases is an unexpected but incredibly valuable angle, as these areas have historically struggled with the high costs and low success rates of traditional drug development.

A Regulatory Embrace: The FDA's Pivotal Shift

Perhaps one of the most significant, yet understated, developments I’ve observed is the evolving stance of regulatory bodies. Historically, the pharmaceutical industry has been cautious about adopting radically new technologies. However, the U.S. Food and Drug Administration (FDA) is now actively embracing AI's role in drug repurposing. I found that in May 2026, the FDA announced a new Drug Repurposing Initiative, explicitly soliciting public input to identify approved drugs that could be repurposed for new indications, particularly for chronic or rare diseases where current treatments are inadequate. Crucially, this initiative emphasizes considering