Industry · Early Pilots
Quantum Computing for Materials Science
Quantum computers are uniquely suited to simulating quantum mechanical systems — making materials science one of the most promising application areas. Simulating high-temperature superconductors and designing novel catalysts are near-term research targets.
Maturity
Early pilots; the strongest scientific case for quantum computing
Most credible quantum workload
Simulating strongly correlated materials and catalysts
Experiments so far
Spin models, Hubbard-model physics, and dynamics on 50 to 150 qubit devices and annealers
Hardware needed for industrial accuracy
Thousands of logical qubits; published estimates for key catalysts reach millions of physical qubits
Strongest classical competition
DFT, DMRG, quantum Monte Carlo, tensor networks, and AI force fields
Realistic horizon
Scientific insights now; industrial design impact in the 2030s
Overview
Materials science is the field where quantum computers have the most natural advantage, and it is also where the earliest scientific results are appearing. The properties of a material, whether it conducts electricity without resistance, stores energy, absorbs sunlight, or resists corrosion, come from the quantum mechanics of its electrons. In many materials those electrons interact strongly, and the number of states needed to describe them grows exponentially with the size of the sample. This is the problem Richard Feynman had in mind in 1982 when he proposed quantum computers.
Today's classical methods, such as density functional theory, coupled cluster, and tensor networks, handle a huge range of materials well. The cases that resist them involve strong correlation: high-temperature superconductors, certain magnets, transition-metal catalysts, and some battery materials. Quantum simulation aims at exactly those cases, and experiments on analog and digital quantum devices have begun to explore simplified models of them.
Why this industry is a candidate for quantum computing
The cost of simulating a quantum system on a classical computer typically grows exponentially with its size, while a quantum computer can represent the state with a number of qubits that grows only linearly. For the simplest models, such as lattices of interacting spins and electrons, this has already allowed experiments that probe regimes where classical methods are strained.
There are two routes. Analog and digital simulation with near-term devices studies model Hamiltonians directly, and has produced published experiments on magnetism, phase transitions, and dynamics. Fault-tolerant quantum phase estimation promises accurate energies for realistic materials, but needs thousands of logical qubits. Both routes are active research, and the field is candid about the distance between them.
Key facts and numbers
| Item | Detail |
|---|---|
| Feynman's proposal | 1982 paper suggesting quantum computers to simulate quantum physics |
| IBM utility experiment | 127-qubit kicked Ising dynamics published in Nature in 2023, later matched by classical simulations |
| Google Quantum Echoes | Reported 13,000 times faster than the best classical method tested (Nature, October 2025) |
| Chemical accuracy | About 1.6 millihartree, the target for useful energy predictions |
| Catalyst resource estimates | Thousands of logical qubits for smaller systems, millions of physical qubits for FeMoco and P450 |
| D-Wave materials claim | 2025 Science paper on beyond-classical magnetic simulation, challenged by tensor network studies |
Where quantum computing could help
Strongly correlated materials and superconductors
The problem: High-temperature superconductors and related quantum materials are described by models such as the Hubbard model, where electrons interact strongly. Exact solutions are known only in special cases, and approximate methods disagree about key properties.
The quantum approach: Digital and analog quantum simulators implement the Hubbard model directly, using trapped ions, neutral atoms, and superconducting qubits, and measure properties such as pairing and magnetism.
Where things stand: Experiments on devices from Google, Quantinuum, IBM, and cold-atom platforms have simulated spin and Hubbard dynamics at scales that stress classical methods. A 2023 IBM experiment on 127 qubits and a 2025 Google result called Quantum Echoes are examples, and several claims were later matched by improved classical simulations. Genuine, accepted advantage for a useful material is not yet shown.
Related algorithms: Trotter-Suzuki Decomposition, Quantum Error Mitigation Techniques, Quantum Annealing
Catalysts and reaction mechanisms
The problem: Catalysts that make ammonia, hydrogen, and fuels, and enzymes such as nitrogenase and cytochrome P450, involve transition metals with many nearly degenerate electronic states that approximate methods handle poorly.
The quantum approach: Quantum phase estimation with compressed Hamiltonians computes ground-state energies of the active space with controlled accuracy.
Where things stand: Resource estimates range from thousands of logical qubits for smaller catalysts to a few million physical qubits and days of runtime for FeMoco and P450. Work by von Burg, Reiher, and others on ruthenium catalysts and by Google and Boehringer Ingelheim on P450 is typical. Hardware to run them does not yet exist.
Related algorithms: Quantum Phase Estimation (QPE), Linear Combination of Unitaries (LCU), Quantum Signal Processing (QSP)
Batteries and energy materials
The problem: Predicting electrolyte stability, ion transport, and electrode behavior needs accurate energies and dynamics for molecules and interfaces.
The quantum approach: Variational methods probe small fragments now, and fault-tolerant simulation could handle larger ones later.
Where things stand: Studies of lithium electrolyte molecules estimate thousands of logical qubits. Meanwhile, AI-driven screening, such as the 2024 Microsoft and Pacific Northwest National Laboratory electrolyte search, finds candidates classically and sets a high bar.
Related algorithms: Variational Quantum Eigensolver (VQE), Quantum Phase Estimation (QPE)
Photovoltaics, semiconductors, and magnets
The problem: Solar absorbers, light-emitting materials, and magnetic materials depend on excited states, defects, and magnetic interactions that are hard for approximate methods.
The quantum approach: Algorithms for excited states, dynamics, and spectral functions are being developed, along with quantum-centric supercomputing workflows that combine quantum and classical resources.
Where things stand: Collaborations such as Phasecraft with Oxford PV and Johnson Matthey aim to build algorithms for these tasks, and IBM and others promote hybrid workflows. Demonstrations are on model systems.
Related algorithms: Quantum Imaginary Time Evolution (QITE), Quantum Gibbs Sampling
Who is working on it
- Google Quantum AI, IBM, Quantinuum, and QuEra, which run simulation experiments on their hardware, often with academic partners.
- Phasecraft, which builds algorithms for materials simulation with partners such as Oxford PV, Johnson Matthey, and hardware makers including Google, IBM, Quantinuum, QuEra, and Atom Computing.
- Chemical and materials companies, including BASF, Dow, Merck KGaA, JSR, and Mitsubishi Chemical, in collaborations with hardware and software providers.
- National laboratories such as Oak Ridge, Pacific Northwest, Lawrence Berkeley, and Sandia, and university groups at Caltech, Harvard, and ETH Zurich.
- D-Wave, which published a 2025 paper claiming a beyond-classical simulation of magnetic materials that other groups have challenged with classical methods.
Realistic timeline
| Horizon | What to expect |
|---|---|
| Now to 2028 | Analog and digital simulations of model systems on 100-qubit class devices, hybrid workflows, and an ongoing cycle of advantage claims and classical rebuttals. |
| 2028 to 2032 | Early fault-tolerant machines with logical qubits could compute small but chemically meaningful active spaces beyond exact classical methods. |
| 2032 and later | If fault-tolerant machines reach thousands of logical qubits, simulation of catalysts and correlated materials could start to inform industrial design. |
How to judge quantum claims in this sector
Materials science has a mature culture of benchmarking, which makes it a good model for judging quantum claims. A convincing result names the physical model, the observable measured, the system size, and the best classical method compared, ideally with an independent group trying to reproduce it. Several recent advantage claims were narrowed within months by improved tensor network or Monte Carlo methods, which is a sign of a healthy field.
The distinction between model systems and real materials matters most. A simulation of a lattice model on 100 qubits is a scientific achievement but does not predict a specific superconductor. The step from model to material needs accurate parameters and often additional physics. The most useful announcements are explicit about this gap.
What stands in the way
- Hardware scale and error rates: industrial accuracy needs error correction and many logical qubits.
- Moving classical baseline: tensor networks and machine learning keep pushing the boundary of what classical methods can do.
- State preparation: phase estimation needs a good starting state, which is itself hard for correlated systems.
- Gap to practice: accurate energies are only one input to materials design, which also needs synthesis and testing.
What organizations in this sector should do now
- Identify the specific properties where your current methods fail, and quantify the accuracy needed.
- Collaborate with academic or vendor groups on active-space selection and classical baselines.
- Follow independent verification of advantage claims, which have often been revised.
- Use classical AI and HPC for screening now, and keep quantum for the hardest cases.
Myths versus reality
- Myth: quantum computers have already designed a new material. Reality: experiments have simulated model systems, not delivered industrial materials.
- Myth: all materials problems need quantum computers. Reality: classical methods handle most, and quantum targets the strongly correlated minority.
- Myth: every advantage claim is final. Reality: several were narrowed by better classical algorithms.
The bottom line
Materials science is the field most likely to see real, early scientific value from quantum computers, and it is already producing papers. Industrial impact depends on fault-tolerant hardware and on how far classical methods keep advancing, so organizations should engage through collaborations and watch independent verification closely.
Frequently asked questions
Why is materials science a good fit for quantum computers?
Because material properties come from quantum mechanics of interacting electrons, whose complexity grows exponentially on classical machines for strongly correlated systems.
Has a quantum computer simulated a material?
Yes, simplified models such as spin chains and the Hubbard model, on devices with tens to over a hundred qubits. Useful, industrial-grade simulations are still out of reach.
How many qubits are needed for industrial catalysts?
Estimates range from thousands of logical qubits to millions of physical qubits with days of runtime, depending on the system and assumptions.
What is quantum-centric supercomputing?
A workflow, promoted by IBM and others, in which quantum processors handle the hardest part of a calculation and classical supercomputers handle the rest.
When will materials benefit?
Scientific insights are emerging now. Industrial design impact is likely in the 2030s at the earliest.
Keep exploring
- Algorithm: Quantum Phase Estimation (QPE)
- Algorithm: Trotter-Suzuki Decomposition
- Algorithm: Variational Quantum Eigensolver (VQE)
- Algorithm: Quantum Imaginary Time Evolution (QITE)
- Hardware: Google Willow
- Hardware: Quantinuum H2
- Hardware: IBM Heron (r1, r2, and r3)
- Company: Phasecraft
- Company: Quantinuum
- Company: Google (Google Quantum AI)
- Research: Higher-Accuracy Molecular Energy Estimation Using Adaptive VQE Ansätze
- Research: Automated Active Space Selection Improves Practical VQE Chemistry Simulations
- Research: Independent Verification Attempt of a Recent Quantum Advantage Claim
- Industry: Quantum Computing for the Chemical Industry
- Industry: Quantum Computing and the Semiconductor Industry
- Industry: Quantum Computing for Pharmaceuticals and Drug Discovery
- Industry: Quantum Computing for Energy and Utilities
- Industry: Quantum Computing for Agriculture and Food
This page is written for educational purposes and summarizes public announcements, published research, and published resource estimates (last reviewed September 30, 2026). Vendor and customer claims have not all been independently verified, timelines are estimates, and nothing here is investment or business advice.