Can Quantum Computing Reduce the Cost of Failed Drug Candidates?

By Abdul Ahad NazakatReviewed by Louis CastelFeb 13 2026

Pharmaceutical R&D is expensive and inefficient. Bringing a single drug to market costs between $2.6 billion and $3.5 billion, and roughly 90% of candidates that enter clinical trials ultimately fail.^{1, 2} The economic burden is staggering: the industry spends an estimated $50–60 billion annually on failed oncology trials alone.³ Most failures trace back to lack of efficacy, unexpected toxicity, or poor pharmacokinetic profiles, problems that often reflect limitations in the computational and experimental models used during early, stage discovery.^{4, 5} As pharma companies search for ways to improve hit rates and reduce waste, quantum computing might be the solution to many of those problems.

Image Credit: Gorodenkoff/Shutterstock.com

Why Do Drug Candidates Fail?

Drug development attrition occurs at every stage of the pipeline. Among the various causes, lack of efficacy, particularly the inability to reproduce promising preclinical results in human trials, remains the most frequently cited reason for clinical termination. Safety and toxicity signals account for another large share of failures.⁴ Pharmacokinetic and pharmacodynamic (PK/PD) problems, including poor absorption, rapid metabolism, or inadequate target engagement, also drive significant attrition.⁵

Beneath these technical failures lies a more fundamental challenge: the computational and experimental models used in early discovery are imperfect predictors of in vivo performance. Classical molecular docking and QSAR models, while valuable, can overlook subtle electronic effects, struggle to account for protein flexibility, and often fail to capture the quantum-mechanical nature of bond formation.⁶ As a result, compounds advance into expensive testing without sufficient confidence in their true binding affinity, selectivity, or metabolic stability.

The financial consequences are severe. A case study of insulin-like growth factor-1 receptor (IGF-1R) inhibitors illustrates the scale: sixteen different programs collectively ran 183 clinical trials at an aggregate cost of $1.6–2.3 billion, yet none secured approval.³ Improving the fidelity of early-stage predictions, so that fewer poor candidates enter the pipeline, is therefore a key lever for reducing overall R&D costs.

No time to read now? Download the PDF of this article

What Quantum Computing Offers

Quantum computing promises to address the limitations of classical simulation by exploiting the same quantum-mechanical principles that govern molecular interactions. Electrons in a protein active site exist in superposition and exhibit entanglement; accurately modeling these states on a classical computer requires resources that scale exponentially with system size. Quantum computers can represent and manipulate superpositions natively, enabling tractable simulations of electronic structure at a level of fidelity that classical methods cannot match.⁷

First, quantum algorithms can enable high-accuracy quantum chemistry calculations for enzyme active spaces. Classical methods applied to active spaces of approximately 50 orbitals can demand more than 1,000 years of compute time. In contrast, advanced quantum algorithms, such as sparse qubitization, are estimated to reduce those runtimes to days on an error-corrected quantum device, making previously impractical calculations realistically achievable.⁸

Second, quantum machine learning (QML) can replace selected modules within classical computer-aided drug design pipelines. A hybrid workflow known as HypaCADD substituted a QML module to predict mutation impacts on binding and was executed on commercial quantum hardware. In a SARS-CoV-2 protease case study, the QML component matched or outperformed classical baselines, demonstrating that targeted quantum enhancements can integrate effectively into existing discovery workflows.⁹

Third, quantum optimization can explore combinatorial chemical space more efficiently, potentially identifying higher-quality lead compounds with better multi-objective profiles.^{6, 9} A variational quantum-classical approach applied to a serine-protease fragment achieved chemical accuracy and produced a mean absolute error of 0.25 kcal/mol in binding energy prediction - precision that classical force fields rarely achieve.¹⁰

Current Industry Efforts and Collaborations

Several companies and academic groups are actively developing quantum tools for drug discovery, though most efforts remain at the proof-of-concept stage. IBM has demonstrated variational quantum eigensolver (VQE) workflows for molecular simulation in drug-discovery case studies. ¹¹ The HypaCADD workflow executed QML modules on commercial quantum devices and validated the approach on a SARS-CoV-2 protease target.⁹

Reviews note active collaboration between pharmaceutical companies and quantum vendors, with quantum annealers mentioned as industrially relevant platforms. ^{6, 11} However, detailed public disclosures of specific partnerships involving companies such as Qubit Pharmaceuticals, Aqemia, ProteinQure, D-Wave, Google Quantum AI, Boehringer Ingelheim, or Roche remain sparse in the peer-reviewed literature. Many industry collaborations appear to be in early stages or conducted under confidentiality agreements.

Challenges to Overcome

Current quantum hardware is noisy and limited in qubit count and coherence time. Near-term noisy intermediate-scale quantum (NISQ) devices can run small hybrid algorithms, but many pharmaceutically relevant targets will require error-corrected architectures.^{6, 8} Error correction imposes substantial qubit overhead, thousands of physical qubits per logical qubit, which means that large-scale, fault-tolerant quantum computers are still years away.

Algorithmic maturity is another constraint. The literature consistently recommends directing quantum resources toward specific sub-modules where they offer clear advantages, for example, selected electronic-structure fragments or machine-learning components that replace computational bottlenecks. It also emphasizes validating these targeted applications through small-scale wet-lab campaigns before expanding their use more broadly. ⁹

Regulatory and data-validation hurdles also remain significant. Most quantum demonstrations to date have been conducted at the fragment scale, and broad empirical evidence linking quantum-derived outputs to reduced wet-lab failure rates, or to measurable cost savings, has yet to appear in the peer-reviewed literature.⁶

Download the PDF of this article

Cost and Impact Potential

If technical and validation challenges can be overcome, quantum computing has the potential to reduce preclinical R&D costs by improving early-stage predictions. Resource-estimate studies suggest that once suitable error correction exists, some hard quantum-chemical calculations could be completed in days, rather than centuries with current computing methods.⁸ This would enable more rigorous in-silico screening, allowing teams to eliminate poor candidates before they consume expensive resources.

The most realistic near-term impact will come from hybrid workflows combining quantum and classical methods. A hybrid pipeline might use quantum algorithms to compute high-accuracy binding energies for candidates generated by classical docking, or employ QML modules to predict ADMET properties with greater precision. These incremental improvements could reduce attrition rates by a few percentage points - a modest gain that translates into hundreds of millions of dollars in avoided costs, when multiplied across an industry that spends tens of billions annually on failed trials.³

Looking five to ten years ahead, if error-corrected quantum computers become available and hybrid workflows are validated in prospective studies, quantum methods could become standard tools in hit-to-lead and ADMET prediction stages. However, there is currently no proper estimate of the direct dollar-per-candidate savings or validated end-to-end cost-reduction figures.⁶

Future Directions for the Industry

Several steps are needed to move quantum computing from proof-of-concept to routine use. The industry needs standards for benchmarking quantum algorithms against classical methods on drug-relevant tasks. More cross-sector partnerships between pharma companies, quantum hardware vendors, and academic groups are essential. Hardware-software co-development is critical: algorithm designers need input from chemists to target the right problems.

The field must address quantum advantage with intellectual honesty. Not every drug-discovery task will benefit from quantum computing. The goal should be to identify specific bottlenecks, such as accurate simulation of transition-metal active sites or exploration of vast combinatorial libraries, where quantum methods offer a clear advantage over classical alternatives.

Could quantum computing be the catalyst for the next wave of precision drug development? The answer depends on whether the field can deliver on its technical promise and validate its methods in real-world R&D campaigns. The early results are encouraging: quantum algorithms have achieved chemical accuracy on small fragments, hybrid workflows have matched classical baselines on commercial hardware, and resource estimates suggest dramatic speedups for hard problems. If these capabilities can be scaled and validated, quantum computing may indeed help the industry reduce the staggering cost of failed drug candidates.

How do we know we can trust quantum computers? The answer is here

Further Reading and References

1. DiMasi JA, Grabowski HG, Hansen RW. Innovation in the pharmaceutical industry: New estimates of R&D costs. J Health Econ. 2016;47:20-33. https://doi.org/10.1016/j.jhealeco.2016.01.012
2. Fernald KDS, Förster PC, Claassen E, van de Burgwal LHM. The pharmaceutical productivity gap – Incremental decline in R&D efficiency despite transient improvements. Drug Discov Today. 2024;29(11):104160. https://doi.org/10.1016/j.drudis.2024.104160
3. Jentzsch V, Osipenko L, Scannell JW, Hickman JA. Costs and causes of oncology drug attrition with the example of insulin-like growth factor-1 receptor inhibitors. JAMA Netw Open. 2023;6(7):e2324977. https://doi.org/10.1001/jamanetworkopen.2023.24977
4. Hughes JP, Rees S, Kalindjian SB, Philpott KL. Principles of early drug discovery. Br J Pharmacol. 2011;162(6):1239-1249. https://doi.org/10.1111/j.1476-5381.2010.01127.x
5. Walker I, Newell H. Do molecularly targeted agents in oncology have reduced attrition rates? Nat Rev Drug Discov. 2009;8(1):15-16. https://doi.org/10.1038/nrd2758
6. Pasupuleti MK. Hybrid quantum-classical algorithms for drug discovery and molecular simulation. 2025. https://doi.org/10.62311/nesx/rphcrcscrqc5
7. Priya D, Shaik N, Chitralingappa P. A next generation innovation in drug discovery using quantum generative AI. IOSR J Pharm Biol Sci. 2025;10(4):48-54. https://doi.org/10.35629/4494-1004448454
8. Blunt NS, Camps J, Crawford O, et al. Perspective on the current state-of-the-art of quantum computing for drug discovery applications. J Chem Theory Comput. 2022;18(12):7001-7023. https://doi.org/10.1021/acs.jctc.2c00574
9. Banchi L, Fingerhuth M, Babej T, Ing C, Arrazola JM. Molecular docking with Gaussian boson sampling. Sci Adv. 2020;6(15):eaay4058. https://doi.org/10.1126/sciadv.aay4058
10. Ayuba A, Henry J, Tera T. Quantum computing in drug discovery: A paradigm shift. OSF Preprints. 2023. https://doi.org/10.31219/osf.io/5yfs3
11. Wang PH, Chen JH, Yang YY, et al. Recent advances in quantum computing for drug discovery and development. IEEE Nanotechnol Mag. 2023;17(2):20-29. https://doi.org/10.1109/MNANO.2023.3249499

Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.