Ten industries where useful quantum computing could create — or destroy — enormous value. Each comes with an honest timeline, a realism check, and the players already working on it. Drug discovery and materials lead. Quantum ML is mostly hype today. Cryptography is the urgent risk.
See also: the risks page for the other side of these forecasts.
Drug discovery & medicine
Simulate how molecules bind to proteins at quantum-mechanical accuracy — pre-screen drug candidates in silico before any wet lab.
Why it matters
Pharma companies spend an average of $2.6B to bring one drug to market and 90% of candidates fail. Most of that failure happens because classical computers cannot accurately predict how a drug molecule interacts with the target protein. Quantum computers naturally represent the quantum states underlying chemistry — that's the entire reason Richard Feynman proposed building one in 1982.
What changes
Drug candidates get screened against targets virtually, with accuracy that classical molecular dynamics can't match. Antibody design accelerates. Personalized medicine becomes computationally tractable. Rare-disease research becomes economically viable because cost-per-candidate plummets.
Timeline
Useful quantum simulation of small molecules: 2028–2030. Drug-receptor binding at therapeutic scale: 2030–2033. Wide pharma adoption: post-2033.
Qubits needed
100–1,000 logical qubits with <10⁻⁷ logical error rate.
Early wins to watch
· FeMoco active-site simulation (nitrogenase — also unlocks fertilizer)
· ATP synthase ground-state energy
· Cytochrome P450 drug metabolism predictions
· Antibody-antigen binding affinities
Already working on it
· AstraZeneca + IonQ
· Boehringer-Ingelheim + Google Quantum AI
· Pfizer + AWS Braket
· Cleveland Clinic + IBM Quantum
· Roche + Cambridge Quantum (Quantinuum) on InQuanto
Bull case
McKinsey expects pharmaceuticals to be the single largest commercial quantum revenue category through 2035. The economics work because pharma already spends $200B/yr on R&D — a 5% acceleration is $10B in annual value.
Realism check
AlphaFold solved protein structure prediction with classical AI; quantum needs to leapfrog that bar, not just match it. Most "quantum advantage" claims in drug discovery so far are on tiny molecules with <100 atoms. The leap to therapeutic relevance is large.
Materials science
Design new superconductors, batteries, catalysts, and aerospace alloys by simulating electrons and lattices accurately for the first time.
Why it matters
The Haber-Bosch process for nitrogen fixation (industrial fertilizer) consumes 1–2% of global energy. Nitrogenase, a bacterial enzyme, does the same chemistry at room temperature — we simply do not understand how. The active site (FeMoco) has 100+ correlated electrons, well beyond classical simulation. Quantum computing could solve this and revolutionize agriculture.
What changes
Higher-temperature superconductors (which would transform the power grid). Solid-state batteries with 3-5× current density. Catalysts for cheaper hydrogen production. Carbon-capture materials that work at industrial scale. Lighter, stronger aerospace alloys.
Timeline
Useful materials simulation: 2028–2032. First commercially-designed quantum-discovered material: 2030–2035.
Qubits needed
100–500 logical qubits.
Early wins to watch
· FeMoco / nitrogenase active site (fertilizer revolution)
· Cuprate superconductors
· Battery cathode materials (Li-S, Na-ion)
· Catalysts for green hydrogen
Already working on it
· BASF + Pasqal
· Mitsubishi Chemical + Quantinuum
· BMW + Quantinuum (battery chemistry)
· JSR + IonQ (semiconductor materials)
Bull case
McKinsey estimates materials/chemistry will produce $43-71B of quantum-driven value by 2035. Industrial chemistry alone is a $5 trillion industry where even single-digit efficiency gains compound enormously.
Realism check
Most "first useful quantum chemistry" claims to date have been benchmarkable against classical methods within hours on a laptop. Genuinely quantum-hard chemistry (multi-reference, strongly correlated electrons) is the threshold to clear.
Climate & clean energy
Better catalysts for carbon capture, more efficient solar cells, faster path to fusion materials, optimized grids.
Why it matters
Climate change is fundamentally a chemistry and physics problem at scale: we need better catalysts, better photovoltaic materials, better batteries, better grid optimization. Each of those is computationally bound today.
What changes
CO₂-to-fuel catalysts at industrial cost. Perovskite solar cells at higher efficiency. Fusion-reactor wall materials that survive plasma. Grid optimization that handles intermittent renewables better. Direct-air-capture chemistry orders of magnitude cheaper.
Timeline
Catalyst design wins: 2028–2032. Grid optimization (already partial via D-Wave annealing): incremental from now. Fusion-relevant simulations: 2030+.
Qubits needed
100–500 logical qubits for chemistry; classical-quantum hybrid for optimization (today).
Climate solutions are the highest-stakes near-term application. If quantum accelerates carbon capture by even 10x, the impact dwarfs every other use case combined.
Realism check
Most "quantum + climate" announcements today are pilot programs with little commercial deployment. Real grid optimization works today on annealers, but the big chemistry wins are still 5+ years out.
Cryptography & security
Shor's algorithm breaks RSA/ECC. NIST has already standardized post-quantum replacements. The race is who migrates first.
Why it matters
Essentially every secure connection on the internet — HTTPS, SSH, bank wires, government communications — relies on factoring or elliptic-curve discrete-log problems that quantum computers can solve in polynomial time. The transition to post-quantum cryptography (PQC) is the largest cryptographic migration in history.
What changes
RSA-2048 and ECC-256 become obsolete. NIST-standardized ML-KEM, ML-DSA, SLH-DSA become the new defaults. Hardware (HSMs, TPMs) must be replaced. Symmetric crypto (AES) survives but with longer keys. New "quantum-safe" products emerge across networking and identity.
Timeline
NIST PQC standards finalized August 2024. Federal deadline: 2030–2035 (varies). Practical RSA-breaking quantum computer: 2030+ (uncertain).
Qubits needed
Breaking RSA-2048 requires ~20M noisy qubits or ~thousands of logical qubits (Gidney/Ekerå 2021).
Early wins to watch
· NIST PQC migration in federal systems (NSM-10, 2022 mandate)
· Quantum Key Distribution for sovereign comms (limited scope)
· Apple iMessage PQ3, Cloudflare PQC TLS — already live
Already working on it
· NIST (algorithm standards)
· Cloudflare, Apple, Google, AWS, Microsoft (deployment)
· SandboxAQ, PQShield, Arqit, ID Quantique (vendors)
· NSA CNSA 2.0 mandate
Bull case
PQC is the rare cybersecurity category with a hard federal deadline and a non-discretionary buyer. Total migration spend likely exceeds $40B globally over a decade.
Realism check
Most spend goes to incumbents (Entrust, Thales, IBM, hyperscalers) — pure-play startups capture a thin slice. QKD remains niche; NSA explicitly does not recommend it for national security systems.
Logistics & optimization
Quadratic speedups on hard combinatorial problems — airline routing, supply chain, portfolio construction.
Why it matters
Many of the most economically important problems in business — routing, scheduling, resource allocation, portfolio optimization — are NP-hard combinatorial problems. Even modest speedups translate to billions in saved operations.
What changes
Airlines route fleets more efficiently. Supply chains rebalance in real time. Financial portfolios get rebalanced with constraints classical heuristics miss. Traffic systems optimize at city scale.
Optimization is the only quantum domain with real, deployed commercial value today. As annealer scale grows, more use cases tip from "classical wins" to "quantum wins."
Realism check
The "quantum advantage in optimization" story has cried wolf for a decade. Classical solvers (Gurobi, CPLEX) keep getting better. Most "quantum-only-solvable" problems turn out to be solvable classically with enough engineering.
If quantum machine learning works, it could accelerate AI training, enable new model architectures, and handle data that's naturally quantum (chemistry, materials). If it doesn't — which is the consensus view today — quantum + AI will mean classical AI orchestrating quantum subroutines.
What changes
Quantum kernels for SVM-like classifiers. Quantum-enhanced sampling for generative models. Quantum reservoir computing. Most realistically: classical AI doing the heavy lifting with quantum coprocessor calls for specific kernels (chemistry, optimization sub-problems).
Timeline
Limited commercial QML: 2027–2030. Genuine quantum advantage on real ML benchmarks: highly uncertain.
· NVIDIA Ising — open quantum AI models for QPU calibration (Apr 2026)
· Hybrid quantum-classical training pipelines
Already working on it
· NVIDIA (CUDA-Q, NVQLink, Ising)
· Google, IBM, Quantinuum research arms
· Xanadu PennyLane (open-source library)
· SandboxAQ (quantum-inspired ML)
Bull case
If QML works even narrowly on chemistry or financial data, the combined quantum+AI market is the largest in tech. Even a small QML coprocessor add-on to NVIDIA stacks could be massive.
Realism check
No demonstrated quantum advantage on a real-world ML benchmark exists today. The theoretical case is contested. Be skeptical of "quantum AI" marketing — it's almost always hype layered on hype.
Finance & risk
Monte Carlo speedups on derivative pricing, faster risk calculations, portfolio optimization.
Why it matters
Financial models — option pricing, value-at-risk, credit risk — are dominated by Monte Carlo simulations. Quantum amplitude estimation provides a quadratic speedup on Monte Carlo, which translates to dramatically faster overnight risk calculations or finer-grained intraday models.
What changes
Banks run risk calculations in seconds rather than overnight. Derivative pricing becomes more accurate. Fraud detection patterns surface that classical algorithms miss. Real-time portfolio rebalancing under constraints.
Timeline
Production financial quantum use: 2028–2032.
Qubits needed
100–500 logical qubits for meaningful Monte Carlo speedups.
Early wins to watch
· JPMorgan + IBM credit risk modeling
· Goldman Sachs derivative pricing pilots
· HSBC commercial advantage paper (34% accuracy lift on optimization, Nov 2025)
Wall Street is the most willing-to-pay customer base in technology. Even modest speedups on Monte Carlo justify large procurement contracts.
Realism check
Most financial Monte Carlo workloads parallelize trivially on GPUs. The quadratic speedup from quantum needs to outpace the constant-factor advantage of massive classical compute farms. The math is closer than vendors imply.
Quantum sensors exploit superposition and entanglement to measure magnetic fields, gravity, rotation, and time at sensitivities 10-1000× better than classical instruments. Unlike quantum computing, much of this is already deployed in defense pilots today.
What changes
Submarines and aircraft navigate without GPS. Brain MEG scanners that fit on a helmet replace $3M cryogenic suites. Mineral and oil exploration via sub-surface gravity sensing. Cardiac MCG diagnostics in ambulances.
Timeline
Defense PNT (positioning, navigation, timing): deploying now. Medical OPM-MEG: commercial by 2027. Civilian PNT backup: 2028+.
Qubits needed
Different physics — sensing uses single-atom or single-spin systems, not qubit arrays.
Early wins to watch
· Q-CTRL Ironstone Opal — 111× better positioning than INS in GPS-denied trials
· Infleqtion + Royal Navy underwater optical clock
· SBQuantum NV-diamond magnetometers in defense
· Cerca Magnetics OPM-MEG nearing regulatory filing
Quantum sensing has real product-market fit *today* in defense PNT. GPS denial in Ukraine, Red Sea, and Indo-Pacific is a forcing function. Medical imaging upside is 10× displacement story.
Realism check
TAMs remain small ($0.5–2B near-term) and dominated by lumpy government contracts. Many "quantum" sensors compete with mature classical alternatives (MEMS gyros, GPS+INS fusion, cesium clocks) that keep getting better and cheaper.
Fundamental science
Simulate quantum chromodynamics, condensed-matter exotica, possibly quantum gravity. The Feynman use case.
Why it matters
Feynman's 1982 motivation: simulating quantum systems on classical computers takes exponential resources, but quantum computers naturally represent quantum states. High-energy physics, condensed matter, cosmology, and quantum gravity all benefit.
What changes
Lattice QCD calculations done with quantum computers. High-temperature superconductor mechanisms understood. Topological phases of matter mapped. Black hole information paradox studied via quantum simulation. Cosmology and inflation models tested.
· Various university physics departments via IBM/AWS/Azure
Bull case
Fundamental science discoveries enabled by quantum computers will reshape physics curricula. Nobel-grade results in 2030s. The original Feynman use case is also the most physically defensible.
Realism check
Doesn't generate revenue. But generates the validation results that legitimize the whole field. Watch the peer-reviewed physics papers, not the press releases.
National security & defense
Codebreaking. Stealth detection via quantum radar. Secure communications via QKD. Quantum-AI hybrid intelligence systems.
Why it matters
Whoever achieves cryptographically relevant quantum computing first gains an enormous intelligence advantage — they can read previously-secured adversary communications while their own are protected. This drives the largest single source of quantum R&D funding globally.
What changes
Adversaries' encrypted intelligence backlog becomes readable. Quantum sensing reveals submarines and stealth aircraft. Quantum-secured government links become the standard. Cyber capabilities asymmetric.
Timeline
Most classified. Public-track: PQC migration mandates by 2030. Quantum-enabled SIGINT: timeline opaque.
Qubits needed
Same as cryptography — thousands of logical qubits for Shor at RSA scale.
· DOE National QIS Centers ($625M renewal Nov 2025)
· AFRL contracts to IonQ, Q-CTRL, Atom Computing
· AUKUS Pillar 2 quantum technologies
Already working on it
· DARPA, DOE, NSA, AFRL, NRO (US)
· GCHQ + UK NQCC
· China national quantum program (15th Five-Year Plan)
· NATO quantum initiatives
Bull case
National security is the largest single source of quantum revenue today (~40-50% of sector). Government funding will continue through 2035 regardless of commercial outcomes.
Realism check
The geopolitical implications are sobering. First-mover advantage in cryptographically-relevant quantum computing creates real, hard-to-counter intelligence asymmetry — which is part of why the US, China, and EU treat this as a strategic priority.
The pattern
Look across all ten domains and a pattern emerges: chemistry-flavored problems (drug discovery, materials, climate, fundamental science) cluster around 2028–2032 as the credible first-wave use cases. Cryptography is the urgent counter-current — you don't need useful quantum computing to break RSA, you just need some quantum computing eventually, and adversaries are collecting today.
Optimization wins today on annealers but lacks a proven gate-based path to asymptotic advantage. Quantum ML is mostly hype on real benchmarks. Sensing is the dark horse — already deployed, real revenue, lower TAM than computing but much closer to market.
For investors, the highest-conviction near-term thesis is that 2–3 vendors capture 70%+ of chemistry-driven quantum revenue by 2035. For policymakers, the urgent thesis is post-quantum cryptography migration before the cryptographically-relevant quantum computer arrives. For everyone, the realistic thesis is that quantum complements classical computing — same way GPUs do today — for a narrow set of structured problems where it actually matters.