Quantum Computing

A practical guide for business leaders: what it is, where it matters, and how to move forward without wasting money or time.

1. What quantum computing actually is

2. How it actually works, without the physics lecture

3. The current state: NISQ, fault tolerance, and the honest timeline

4. The algorithm landscape — what quantum computers can and cannot do

5. Where it could matter for your business, by sector

Classical computers — the laptop you're using, the servers your company runs on, the phone in your pocket — all process information in bits: zeros and ones. Every calculation, no matter how complex, eventually reduces to a sequence of bit operations. This model has worked brilliantly for 70 years, but it has a ceiling. Certain problems grow so fast in complexity that no classical machine, however powerful, can solve them in any reasonable time. Simulating a medium-sized molecule, optimising a global supply chain with millions of variables, or breaking certain cryptographic codes fall into that category.

Quantum computers use a fundamentally different unit of information: the qubit. A qubit can exist not just as a 0 or a 1, but in a combination of both at the same time — a state called superposition. When you connect multiple qubits together through a property called entanglement, their combined state grows exponentially in richness: two qubits can represent four states simultaneously, ten qubits can represent 1,024, and 300 qubits could, in principle, represent more states than there are atoms in the observable universe.

This is not a faster version of a classical computer. It's a different kind of machine, good at a specific class of problems. For most of what your business does — running ERP, storing data, serving web pages, training language models — a quantum computer offers no advantage and never will. But for a small set of high-value problems, quantum computing promises answers that no classical machine could ever produce.

Understanding which of your problems belong in that small set is the entire point of engaging with this technology.

First, quantum computers are probabilistic. Unlike classical computers, which always return the same answer for the same input, quantum computers return a distribution of possible answers. You typically run the same circuit thousands of times and analyse the results statistically. This shapes how problems must be formulated and how results are interpreted.

Second, quantum computers are fragile. Qubits are exquisitely sensitive to heat, vibration, electromagnetic noise — essentially any disturbance. Today's machines need massive refrigeration systems (colder than outer space for superconducting qubits) or ultra-high vacuum chambers (for trapped-ion qubits). This fragility is the single biggest engineering constraint in the field.

Third, errors accumulate fast. Because qubits are fragile, every operation you perform on them introduces a small probability of error. Today's best machines can only run circuits a few hundred operations deep before errors overwhelm the signal. Useful algorithms often need millions of operations. Closing that gap is what the field calls "fault tolerance," and it's the central technical challenge.

Fourth, quantum computers don't replace classical ones — they complement them. Real quantum workflows are hybrid: classical computers handle data preparation, parameter tuning, and post-processing; the quantum computer handles only the narrow sub-problem where it has an advantage. This is how every serious quantum application works today and will continue to work for the foreseeable future.

The field is in what researchers call the NISQ era — Noisy Intermediate-Scale Quantum. Today's machines have between 100 and 1,200 physical qubits, with error rates around 0.1% to 1% per operation. That's enough to run interesting experiments, publish impressive papers, and explore applications — but not enough to solve problems that classical computers can't.

The next milestone is fault tolerance: the ability to combine many imperfect physical qubits into one reliable "logical" qubit through error-correcting codes. This is the regime where genuine commercial advantage is expected. The best current estimates suggest we'll see the first useful fault-tolerant quantum computers (a few hundred logical qubits) somewhere between 2029 and 2035, with industrially relevant scale arriving later.

Different companies are pursuing different hardware approaches, each with distinct trade-offs:

• Superconducting qubits (IBM, Google, Rigetti): the most scaled approach today; fast operations but short coherence times; requires extreme cooling.

• Trapped ions (IonQ, Quantinuum): very long coherence times and high-fidelity operations, but slower and harder to scale.

• Neutral atoms (QuEra, Pasqal, Atom Computing): highly scalable and gaining ground rapidly; flexible connectivity between qubits.

• Photonic (PsiQuantum, Xanadu): room-temperature operation and natural compatibility with fibre networks; still proving it scales.

• Topological (Microsoft): potentially the most stable approach if it works; still at the physics-validation stage.

• Silicon spin (Intel, Diraq): leverages existing semiconductor manufacturing; early but promising.

No single approach has clearly won. A reasonable five-year outlook is that two or three of these modalities reach commercial fault tolerance, and the others fade. Betting on one specific hardware vendor today would be premature.

The press often frames quantum as "exponentially faster than classical computing." This is misleading. Quantum computers are exponentially faster for specific problem structures — and for everything else, they are slower, more expensive, and more error-prone than a classical laptop. The business question is whether your problems fit one of those structures.

Shor's algorithm factors large numbers exponentially faster than classical methods. This is what threatens RSA and ECC encryption. It needs a large, fault-tolerant machine that doesn't yet exist but likely will within a decade. Relevant to security strategy (see our Quantum-Safe Security pillar), not to business operations.

Grover's algorithm provides a quadratic speedup for searching unstructured data. Impressive in theory, rarely decisive in practice — a quadratic improvement often doesn't justify the cost of quantum hardware.

Variational Quantum Eigensolver (VQE) and related methods are the most promising near-term tools for simulating molecules and materials. They run on today's noisy hardware and are already producing publishable results in chemistry and materials science. Commercial advantage over classical methods is not yet clearly demonstrated but is the most credible near-term finish line.

Quantum Approximate Optimization Algorithm (QAOA) targets combinatorial optimisation — scheduling, routing, portfolio construction. Results so far are mixed: classical algorithms keep catching up. A useful area to monitor, but no clear advantage yet.

Quantum Machine Learning (QML) — addressed in detail in section 8 below.

The honest summary: no quantum algorithm has yet demonstrated clear commercial advantage over the best classical methods on a problem anyone cares about. That moment — often called "quantum utility" — is expected this decade, most likely in chemistry or materials simulation first.

The value of quantum computing is concentrated in industries where a small number of very hard computational problems sit at the core of the business.

Pharmaceuticals and life sciences are the leading candidate. Drug discovery depends on simulating how molecules interact with proteins — a quantum-mechanical problem by nature. Classical approximations are good but imperfect, and the gap costs the industry years and billions per drug. Every major pharma has a quantum team exploring this. Realistic commercial impact: 5–10 years, compounding.

Chemicals and materials face the same class of problem: designing new catalysts, batteries, fertilisers, or polymers requires simulating molecular behaviour. Fertiliser production (the Haber-Bosch process) alone uses 1–2% of world energy; a better catalyst discovered via quantum simulation would be transformative. Major players include BASF, Dow, Covestro, Mitsubishi Chemical, and their quantum partners.

Finance has explored quantum for portfolio optimisation, derivative pricing (Monte Carlo acceleration), risk analysis, and fraud detection. Results so far are more exploratory than transformative, and classical methods remain competitive. Goldman Sachs, JP Morgan, HSBC, BBVA, and Santander have active quantum teams, mostly positioned as "watch and wait" rather than "deploy now."

Logistics and operations — routing fleets, scheduling factories, optimising supply chains — are natural fits in principle, but in practice classical heuristics are extraordinarily good. Quantum advantage here is the least certain of any major sector. Worth monitoring, not worth betting on for the next five years.

Energy and utilities can use quantum simulation for better battery chemistries, more efficient solar cells, and grid optimisation under high-renewable penetration. Timeline aligns with the materials-simulation curve.

Automotive and aerospace benefit from both materials simulation (lighter alloys, better batteries) and optimisation (manufacturing, routing, CFD). Airbus, Boeing, BMW, Volkswagen, and Daimler all have active programmes.

Defence and intelligence care deeply about quantum for both offensive and defensive reasons: cryptography, sensing, secure communications, and signal processing. A separate conversation with its own rules.

Insurance, retail, telecom, media, professional services — for most businesses outside the sectors above, quantum computing will be irrelevant for the next decade or longer, except through the security implications covered in the Quantum-Safe Security pillar.

If your business is not in one of the high-fit sectors listed here, the right posture is literacy and vigilance — not investment.