Bree Foster is a science writer at Drug Discovery News with over 2 years of experience at Technology Networks, Drug Discovery News, and other scientific marketing agencies.
Drug discovery is advancing at a rapid pace, driven by new technologies, deeper biological insights, and a growing emphasis on patient-centric design. Despite this, many opportunities remain underexplored — opportunities that could redefine how therapies are developed and delivered, making them faster, more effective, and more ethical.
To explore this further, DDN asked a range of industry leaders and academic experts where they see the greatest untapped potential in their fields. From reducing patient burden in clinical research to harnessing the promise of artificial intelligence (AI) and quantum computing, their perspectives shed light on the transformative possibilities that could define the next era of drug discovery.
The greatest untapped opportunity lies in the potential to reduce patient burden in clinical research, either by decreasing the number of patients required in a given trial or by reducing the total number of trials needed. While process automation and faster timelines are valuable, the real impact will come from fundamentally changing how clinical studies are designed.
By using AI to develop smarter, more informed protocols that reflect a deeper understanding of patient populations, we can reduce screen failure rates, accelerate recruitment, and ensure studies better represent the diverse populations they aim to serve. This approach not only improves operational efficiency but also enhances the ethical foundation of clinical research by minimizing unnecessary patient exposure.
See also: What could clinical trials look like in the future?
From a diagnostics perspective, one of the greatest opportunities lies in multiomic approaches to support drug discovery. Right now, we don’t look at sample analysis holistically. The standard paradigm is to analyze each analyte in isolation and then combine the results. But in doing so, we miss the biological connections between them, such as the interplay of DNA methylation and genetic mutation rates.
There are many diseases that present early, progress relentlessly, and currently have limited or no disease-modifying therapies. For example, monogenic kidney disorders such as Alport syndrome, steroid-resistant nephrotic syndrome, and autosomal dominant polycystic kidney disease.
These conditions highlight a broader opportunity in drug discovery: the targeting of monogenic diseases. What makes this space particularly promising is the increasing accessibility of genetic diagnoses, improved natural history data, and growing use of clinicogenomic platforms to stratify patients and identify actionable targets. Furthermore, the single-gene nature of these diseases makes them well-suited to precision approaches such as gene therapies and RNA-based therapeutics, targeted small molecules that correct downstream pathway dysfunction, and repurposing existing drugs based on molecular mechanism alignment. However, patient identification remains a key barrier. Addressing this gap through broader genetic screening and real-world data integration could unlock a new wave of targeted therapies.
One of the biggest untapped opportunities in drug discovery is the integration of AI with quantum computing. Quantum systems have the potential to solve complex quantum mechanical problems that classical computers struggle with — such as calculating binding free energies or mapping reaction pathways. By combining quantum computing with machine learning, we could simulate large biomolecules or materials in chemical Hilbert space, unlocking insights well beyond the reach of classical methods.
Emerging technologies like quantum-enhanced generative models, such as quantum generative adversarial networks, could improve the diversity and quality of molecular proposals, while quantum algorithms may redefine retrosynthesis by generating entirely new reaction pathways. Although still in its early stages, quantum computing holds the promise to address some of the most challenging drug targets, representing a transformative step for the future of discovery.
The most exciting untapped frontier is causal, multimodal reasoning at scale. Imagine models that can learn cause-and-effect across perturbational omics, single-cell readouts, CRISPR screens, and high-content imaging and then connect those biological responses back to a compound’s chemistry and target context using high quality, large datasets.
The opportunity is to train foundation models that reason over these modalities together and plan interventions, linking chemistry to cell-state changes and safety risks before animal studies even begin. From there, we can establish an agentic AI framework where an AI co-scientist asks, “Which perturbation most reduces uncertainty about mechanism?” and then runs that assay under real-world constraints — assay capacity, timelines, budget — updating its priors as new data arrive. The payoff is months of trial-and-error condensed into days of focused learning: fewer but smarter experiments, earlier visibility into pathway compensation and off-targets, tighter dose forecasts, and faster, more confident go/no-go decisions.
The future of drug discovery isn’t fewer experiments; it’s better ones, guided by models that explain what a molecule will do and why. The best time to uncover failure is in silico; the best time to uncover success is sooner. AI and computer-aided drug design is going to change the way we design drugs.
See also: Explore how expansive datasets are advancing AI models for faster and more accurate predictions.
Some responses have been edited for length and clarity.
Bree Foster is a science writer at Drug Discovery News with over 2 years of experience at Technology Networks, Drug Discovery News, and other scientific marketing agencies. She holds a PhD in comparative and functional genomics from the University of Liverpool and enjoys crafting compelling stories for science.
Explore how a needs-driven approach to electronic laboratory notebook selection can improve data integrity, reproducibility, and scientific continuity.
