AI Drug Repurposing 2026: Finding New Treatments Faster with AI

Drug development has always been expensive and slow. The traditional path from identifying a promising molecule to an approved treatment takes twelve to fifteen years and costs upward of a billion dollars per successful drug—with the cost of failures factored in, the economics are even starker. AI is helping to change this equation, and one of the most immediately practical applications is drug repurposing: using AI to identify new therapeutic uses for drugs that are already approved.

In 2026, AI drug repurposing has moved from experimental to operational, with several repurposed drugs reaching clinical trials and initial approvals based on AI-generated hypotheses. The approach represents one of the clearest examples of AI delivering measurable value in healthcare.

What Drug Repurposing Is and Why It Works

Drug repurposing (also called drug repositioning) is the process of identifying new clinical uses for drugs that are already approved for other indications—or drugs that reached late-stage development before failing for reasons unrelated to their mechanism of action.

The economic case for repurposing is compelling:

• Already proven safe: Approved drugs have existing safety profiles from human use. Phase I safety trials, which test whether a drug is safe in humans, are much faster or unnecessary for repurposed drugs.
• Faster to market: Repurposed drugs can reach market in four to six years rather than the full twelve to fifteen year development timeline, because substantial clinical work has already been done.
• Lower failure rates: Safety concerns are the leading cause of late-stage drug failures. Repurposed candidates have already cleared the most dangerous failure mode.
• Better economics: Lower development costs mean that drugs for smaller patient populations—conditions that wouldn't justify the economics of traditional development—become viable.

The limiting factor has always been identifying which drugs might work for which new conditions. This is where AI has made an enormous difference.

How AI Identifies Repurposing Candidates

AI drug repurposing works by analyzing multiple types of data to identify unexpected connections between existing drugs and disease targets:

Molecular interaction modeling: AI systems model how drug molecules interact with protein targets throughout the body. A drug approved for cardiovascular conditions might have strong predicted binding affinity with a protein implicated in a neurological disorder—a connection that wouldn't emerge from traditional drug-by-drug clinical observation.

Gene expression analysis: Diseases alter the pattern of gene expression in affected tissues. AI systems compare these disease-specific gene expression signatures with the gene expression changes induced by various drugs, looking for drugs whose molecular effects might counteract disease mechanisms.

Electronic health record mining: Large-scale analysis of patient medical records can reveal unexpected associations—patients taking drug A for condition X show lower rates of condition Y than comparable patients not taking the drug. These population-level signals can generate repurposing hypotheses that are then investigated mechanistically.

Knowledge graph reasoning: AI systems build large structured databases connecting drugs, proteins, disease mechanisms, clinical observations, and published research. Reasoning across these knowledge graphs can identify multi-step mechanistic pathways linking drugs to new disease targets that no human researcher would encounter browsing the literature.

Literature analysis: The published biomedical literature contains millions of papers with findings that haven't been systematically connected. AI systems trained to extract structured knowledge from scientific text can surface connections across papers that were never meant to be read together.

Notable Successes and Clinical Advances

Several AI-generated repurposing hypotheses have now made it through clinical validation and into practice:

Rare disease applications: The economics of rare disease drug development are particularly difficult because small patient populations can't justify traditional development costs. AI repurposing has become an important tool for rare disease research, identifying drugs approved for common conditions that may be effective for rare genetic disorders sharing molecular mechanisms.

Mental health: AI analysis of electronic health records and molecular interaction data has generated multiple repurposing candidates in psychiatry, a therapeutic area where mechanism-based drug development has been historically difficult. Several candidates identified by AI are in advanced clinical trials.

Infectious disease: The COVID-19 pandemic demonstrated the practical value of rapid repurposing, as researchers urgently searched approved drug databases for candidates that might work against a new pathogen. AI systems accelerated this process, and the pandemic created lasting institutional investment in AI repurposing infrastructure.

Oncology: Cancer treatment has long used approved drugs for cancer types beyond their original approval (off-label use), and AI is systematizing this practice by predicting which cancer subtypes might respond to existing agents based on molecular profiles.

For context on how AI drug repurposing fits into the broader landscape of AI applications in pharmaceutical development, see our coverage of AI in drug discovery.

The Leading Organizations

Several types of organizations are active in AI drug repurposing in 2026:

Dedicated AI pharma companies: A cohort of companies have built AI drug repurposing as their core business model, partnering with traditional pharmaceutical companies to generate and validate hypotheses in exchange for licensing fees or development partnerships.

Academic medical centers: Major research hospitals and universities are running AI repurposing programs, often focused on diseases where commercial development interest is limited but patient need is high.

Big pharma AI units: Traditional pharmaceutical companies have built or acquired AI capabilities, including repurposing, as part of broader AI transformation programs. These internal programs benefit from access to proprietary clinical and molecular data that increases the power of AI analysis.

National health databases: Several countries have created national health data infrastructure specifically to enable AI-driven repurposing research, aggregating de-identified patient data at population scale under research governance frameworks.

Challenges the Field Is Working Through

AI drug repurposing isn't without significant limitations and open questions:

Hypothesis quality varies. Not all AI-generated repurposing candidates survive clinical validation. The false positive rate depends heavily on the quality and breadth of the training data, the rigor of the prediction methods, and the specificity of the mechanistic hypothesis being generated. Some approaches produce many low-quality leads; others generate fewer but more validated hypotheses.

Data quality and access. The richest repurposing signals come from large, well-curated datasets of patient outcomes. Access to this data is uneven, and the data quality of real-world health records is often poor enough to limit what can be learned from it.

Regulatory pathways. Repurposed drugs still require clinical trials demonstrating safety and efficacy for the new indication, even when safety has been established for the original use. Regulatory agencies are developing frameworks for accelerating these trials when the safety profile is well-established, but the pathway remains longer than proponents would like.

IP complexity: Approved drugs are often off-patent, which creates challenges for companies trying to develop and commercialize repurposed versions—if anyone can manufacture the drug, the economics of funding trials can be difficult to structure. New formulations, dosing regimens, or combination approaches can provide IP protection, but navigating this landscape adds complexity.

What This Means for Patients

For patients with conditions where treatment options are limited, AI drug repurposing represents a meaningful near-term source of new options. The conditions that benefit most are those where:

• The disease mechanism is well-characterized at a molecular level
• Existing drugs with relevant mechanisms of action have been approved for other conditions
• The patient population is large enough to motivate clinical trial investment, or small enough that rare disease incentives apply
• There's sufficient electronic health record data to generate real-world repurposing signals

Patient advocacy groups in several disease areas have become active participants in AI repurposing research, helping to fund trials for hypotheses identified by AI that commercial interests might not otherwise pursue. This collaborative model is one of the more promising developments in how AI research translates to patient benefit.

Conclusion

AI drug repurposing is one of the most practical near-term applications of AI in healthcare—not because it's the most transformative possibility, but because it works with what already exists: approved drugs, established safety profiles, and decades of published research that AI can now analyze at scale.

The National Center for Advancing Translational Sciences at NIH has made drug repurposing a strategic priority, investing in AI infrastructure to accelerate the identification and validation of repurposing opportunities across disease areas.

For patients with conditions where current treatments are inadequate, the question isn't whether AI will help identify new options—it already is. The question is how quickly regulatory frameworks, clinical trial infrastructure, and the economics of development can translate AI-identified candidates into treatments that reach people who need them. That gap is closing, and 2026 has seen some of the clearest evidence yet that AI drug repurposing delivers on its promise.