Quantum computing use cases in pharma

Where quantum computing could apply in pharma

The McKinsey Quantum Technology Monitor, April 2026, centers the pharma case on molecular simulation and computer-aided drug design, and TQI sorts the published examples into six areas. Each has a different level of public evidence.

Reaction and enzyme chemistry: one vendor-reported run and one resource estimate

In June 2025 IonQ announced, with AstraZeneca, Amazon Web Services and NVIDIA, a workflow that simulated a step of a Suzuki-Miyaura reaction, a class of reactions used to make small-molecule drugs. IonQ’s Forte processor handled the quantum part, orchestrated with NVIDIA CUDA-Q on Amazon Braket, and NVIDIA H200 GPUs on AWS handled the classical part. IonQ reports more than a 20 times improvement in overall time-to-solution compared with previous implementations and calls it the most complex chemical simulation run on its hardware. The release does not identify the earlier implementations, give accuracy figures or state how many qubits this run used.

The best-known resource estimate for pharma chemistry comes from Google Quantum AI and Boehringer Ingelheim. Their paper on cytochrome P450, posted in February 2022 and published in PNAS according to Google’s 2023 post, compares classical methods with quantum phase estimation for models of the enzyme. The post says a few million physical qubits would be needed to reach quantum advantage for this problem, that the quantum approach starts to win at about 20 to 30 orbitals, and that the enzyme models need more than 60. All of it assumes an error-corrected quantum computer, which the post says is still years away.

A 2023 review in Nature Communications by 18 chemistry and quantum computing researchers adds a caution. For ground-state energy estimation, the most common task in quantum chemistry, it finds that evidence for an exponential quantum advantage across chemical space has yet to be found, while saying quantum computers may still prove useful for chemistry.

Two 2026 announcements use Quantinuum hardware. On August 18, 2026 Quantinuum described a method, built with NVIDIA and Pfizer, in which transformer models propose circuits that prepare the ground state of imipramine, a benchmark molecule. The post says the approach was validated on Quantinuum’s Helios hardware, and it gives no speedup figure or qubit count. In April 2026 Qubit Pharmaceuticals and the Centre for Quantum Technologies in Singapore announced a two-year project to test quantum chemistry algorithms on Quantinuum H2 and Helios systems, and presented first hardware results on a sampling algorithm. The release reports no figures and treats speedups as goals to explore.

Protein and binding studies: larger molecules, still hybrid

In March 2026 Cleveland Clinic and IBM described a workflow that approximates the electronic structure of Trp-cage, a 303-atom miniprotein, in its folded and unfolded states on an IBM Quantum Heron r2 processor, with classical computing at Michigan State University and Cleveland Clinic. The method splits the molecule into fragments, solves the simpler ones classically and sends the hardest to the quantum processor using sample-based quantum diagonalization. IBM says the workflow is competitive with classical approaches. The preprint benchmarks the fragment results against classical methods, is a preprint, and does not claim a quantum advantage.

In September 2023 Pasqal reported, with Qubit Pharmaceuticals, a preliminary water placement algorithm run on its Fresnel 1 neutral-atom machine for the pocket of a protein called MUP-1. The aim is to place water molecules in a protein pocket, a step toward studying how a drug candidate binds. The post calls the run a proof of concept, shows agreement with theory in two-dimensional figures, and gives no error metrics, runtime or classical comparison. The full algorithm is planned for a future machine of about 1,000 qubits.

Candidate design: the one study with lab-tested molecules

Researchers at Insilico Medicine, the University of Toronto, Zapata AI and other institutions published a hybrid quantum-classical generative model for KRAS inhibitors in Nature Biotechnology on January 22, 2025. The quantum part of the model was trained on a 16-qubit IBM computer. The team synthesized 15 candidate molecules and found two promising for further development, one of which showed a binding affinity of 1.4 micromolar to KRAS-G12D. The authors present the result as evidence that quantum computing can produce experimentally validated hits that compare favorably with classical models. TQI covered the study at the time.

mRNA design and NMR: small hardware runs

Moderna and IBM researchers posted a preprint in May 2024 on predicting mRNA secondary structure, a combinatorial problem relevant to RNA therapeutics. For sequences up to 60 nucleotides, which map to problems of 10 to 80 qubits, a CVaR-based variational algorithm on IBM Eagle and Heron processors, with minimal error mitigation, produced minimum free energy structures that matched those of the classical CPLEX solver. Matching a classical solver shows the method works at that size. It does not show an advantage.

In October 2025 Google Quantum AI reported that its Willow chip ran an algorithm it calls Quantum Echoes 13,000 times faster than the best classical algorithm on one of the world’s fastest supercomputers, a comparison Google made itself. In a separate proof-of-principle experiment with UC Berkeley, it ran the algorithm on molecules of 15 and 28 atoms and said the results matched traditional NMR measurements. Google suggests quantum-enhanced NMR could eventually help with drug discovery. That is a possibility, and the 13,000 times figure refers to the algorithm, not to the molecular experiment.

Clinical trial optimization: proposals, no published pilot

McKinsey’s pharma slide lists optimization of clinical trials among the potential levers. The public literature is a 2024 perspective paper, published in Trends in Pharmacological Sciences, volume 45, issue 10, that proposes quantum algorithms for trial simulation, site selection and cohort identification; its preprint carries a different title. It reports no hardware experiments, and no pharma company has publicly reported quantum optimization on a live trial, so TQI does not rate this area.

What quantum has to beat

Classical computing already delivers large gains in the same field. The 2024 Nobel Prize in Chemistry, announced on October 9, 2024, went half to David Baker for computational protein design and half to Demis Hassabis and John Jumper for an AI model that predicts protein structures. Any quantum workflow has to be compared with these tools and with well-tuned classical chemistry methods, not with a weak baseline. McKinsey makes a related point: its pharma impact levers are footnoted as the combined impact of quantum computing and AI.

What the value estimates say, and what they do not

The McKinsey Quantum Technology Monitor, April 2026, estimates that quantum computing could be worth up to about $400 billion for pharma by 2035, equal to up to 12 percent of projected industry gross sales, excluding “black swan” breakthroughs. Its overview slide gives a range of $80 billion to $400 billion. Savings are measured against the current state and not against future technology alternatives, the report calls its value estimates approximate and not definitive projections of business value, and it says implementation depends on factors unique to each company. TQI reads it as an order of magnitude and not a forecast for any one company.

The same report includes a readiness chart built on TQI’s Quantum Index 2025, which scores the 15 most quantum-ready companies in each industry. For life sciences, European companies total 560 points and rank first, ahead of North America at 370.

How to read a quantum pharma claim

1. What is the baseline? IonQ’s 20 times is measured against previous implementations that the release does not name.
2. Hardware, emulator or estimate? The P450 figure is a resource estimate for a machine that does not exist yet. Trp-cage ran on a Heron processor inside a hybrid workflow where classical computers solve the simpler fragments.
3. How big was the problem? Moderna and IBM researchers folded sequences of up to 60 nucleotides, and the KRAS model had a 16-qubit component.
4. Did it beat a classical method or only match it? The mRNA hardware results matched CPLEX.
5. Was anything tested in the lab? Only the KRAS study synthesized and tested molecules. The others stop at computation.
6. Is it peer reviewed? The KRAS paper, the P450 paper and the 2023 review are published. The Moderna and Cleveland Clinic papers are preprints, and the IonQ and Pasqal results are company announcements.

What a pharma team can do now

• List the molecular problems where classical methods are already the bottleneck, and record the accuracy and cost today. The P450 paper uses DMRG and coupled-cluster calculations as its baseline.
• Follow the resource estimates, because they say what hardware must exist before a result matters to a drug program.
• Ask any vendor or adviser for the baseline, the problem size, and whether the run used hardware, an emulator or an estimate.
• For company, funding and patent data across quantum, see TQI Data.

Related guides

The series index lists every industry in this series. For another industry with a chemistry case see the energy guide, and for the one with the longest list of bank pilots see the finance guide.

Questions

Sources

1. McKinsey Quantum Technology Monitor 2026, April 2026 (pharma value estimate, impact levers and readiness chart; the readiness chart cites TQI’s Quantum Index 2025).
2. IonQ speeds quantum-accelerated drug development application with AstraZeneca, AWS, and NVIDIA, IonQ press release, June 9, 2025.
3. Reliably assessing the electronic structure of cytochrome P450 on today’s classical computers and tomorrow’s quantum computers, Goings et al., arXiv:2202.01244, February 2, 2022.
4. Developing industrial use cases for physical simulation on future error-corrected quantum computers, Google Research blog, October 12, 2023.
5. Is there evidence for exponential quantum advantage in quantum chemistry?, Lee et al., arXiv:2208.02199, published in Nature Communications 14, 1952 (2023).
6. Cleveland Clinic protein quantum-centric supercomputing workflow, IBM Quantum blog, March 23, 2026.
7. Molecular quantum computations on a protein, Shajan et al., arXiv:2512.17130, December 18, 2025 (v2 January 16, 2026; preprint).
8. Quantum algorithm can help drug discovery, Pasqal blog, September 27, 2023.
9. Quantum-computing-enhanced algorithm unveils potential KRAS inhibitors, Ghazi Vakili et al., Nature Biotechnology 43, 1954 to 1959, January 22, 2025.
10. Quantum computing-enhanced algorithm unveils novel inhibitors for KRAS, arXiv:2402.08210, February 13, 2024 (preprint version).
11. mRNA secondary structure prediction using utility-scale quantum computers, Alevras et al., arXiv:2405.20328, May 30, 2024 (preprint).
12. Teaching AI with quantum data, Quantinuum blog, August 18, 2026 (work with NVIDIA and Pfizer).
13. Qubit Pharmaceuticals announces strategic collaboration with Singapore’s Centre for Quantum Technologies, press release, April 30, 2026.
14. Quantum Echoes and verifiable quantum advantage, Google blog, October 22, 2025.
15. How can quantum computing be applied in clinical trial design and optimization?, Doga et al., Trends in Pharmacological Sciences 45(10), 880 to 891, October 2024 (DOI 10.1016/j.tips.2024.08.005); preprint arXiv:2404.13113, April 19, 2024.
16. Nobel Prize in Chemistry 2024, Lindau Nobel Laureate Meetings, October 2024 (announcement of October 9, 2024).
17. Quantum Computing Helps Design New Cancer Drug Candidates, The Quantum Insider, January 23, 2025 (context only).