Industry Evaluation, Governance and Future Directions

Introduction

Quantum computing is moving from laboratory research toward a period of industrial evaluation. This transition requires more than identifying impressive demonstrations. Organisations must determine whether a quantum method solves a significant problem, whether the required hardware and expertise are available, whether the cost and risk are acceptable, and whether evidence supports the claim.

For computer science professionals, quantum opportunity is therefore both a technical and governance question. A technically valid algorithm may still be commercially unsuitable because the data-loading cost is too high, the device is unreliable, the classical alternative is superior, or the result cannot be audited. Conversely, a modest near-term use of quantum technology—such as workforce development, algorithm prototyping, or participation in a hybrid workflow—may be strategically valuable even before a measurable production advantage exists.

This chapter presents a disciplined framework for evaluating quantum use cases. It also examines responsible claims, security, ethics, sustainability, standards, hybrid quantum–classical computing, quantum artificial intelligence, networking, and fault-tolerant roadmaps.

Learning Objectives

By the end of this chapter, students should be able to:

1. Evaluate a proposed quantum use case using feasibility, resources, cost, risk, and benchmark evidence.
2. Distinguish a proof of concept, a technical advantage, and a business advantage.
3. Explain governance requirements for quantum software, data, claims, and procurement.
4. Analyse security implications, especially cryptographic migration and quantum-resistant cryptography.
5. Describe the roles and limitations of hybrid computing, quantum AI, quantum networking, and fault-tolerant architectures.
6. Construct a responsible roadmap for adopting quantum technologies.

1. From Quantum Possibility to Industrial Value

A use case is not automatically valuable because a quantum algorithm exists for it. Industrial evaluation begins with the complete problem definition:

• What decision or computation is required?
• What inputs are available, in what format and at what quality?
• What output is useful to a person or operational system?
• How quickly must the result be produced?
• What classical method is currently used?
• What evidence would justify replacing or augmenting that method?

A quantum proposal should be described as a workflow rather than as a circuit alone. The workflow commonly includes data preparation, encoding, quantum execution, error mitigation or correction, classical post-processing, validation, and deployment. If any one of these stages dominates cost or error, the complete workflow may fail to provide value.

1.1 Three different meanings of advantage

The word “advantage” must be used precisely.

Computational advantage means that a quantum procedure performs a defined computational task better than a specified classical procedure under stated conditions. “Better” may refer to runtime, sample complexity, memory, energy, solution quality, or another metric.

Practical advantage means that the complete implemented workflow is superior in an operational setting, including data movement, queueing, compilation, error handling, and integration.

Business advantage means that the organisation obtains economic, strategic, regulatory, scientific, or social value. A practical advantage may not be a business advantage if the value is too small relative to deployment cost.

A credible claim identifies the problem size, hardware or simulator, baseline, metric, uncertainty, and evaluation date. Claims such as “quantum will transform optimisation” are strategic hypotheses, not evidence.

2. A Feasibility and Evidence Framework

A useful evaluation separates five dimensions: problem fit, technical feasibility, resource feasibility, economic feasibility, and organisational risk.

2.1 Problem fit

Quantum methods are most plausible when the problem has a structure that a known quantum algorithm can exploit. Examples include certain simulation tasks, sampling problems, amplitude-estimation settings, and combinatorial optimisation formulations. However, a mapping to a quantum formalism is not sufficient. The mapping must preserve the objective, constraints, precision requirements, and relevant data relationships.

A problem-fit review should ask:

• Is the problem naturally represented by a Hamiltonian, unitary process, oracle, or amplitude-estimation model?
• Is the input size large enough for a quantum method to matter?
• Is the required output a sample, an expectation value, a decision, or a full classical description?
• Does the problem have a known lower bound or classical algorithm that limits the possible benefit?
• Are there patents, safety requirements, or domain constraints affecting deployment?

2.2 Technical feasibility

Technical feasibility includes qubit count, connectivity, gate fidelity, coherence, measurement quality, compilation overhead, circuit depth, parallelism, and access model. A circuit with 100 logical qubits cannot be treated as equivalent to a circuit with 100 noisy physical qubits. Logical qubits are error-protected abstractions; physical qubits are hardware elements subject to noise.

Important questions include:

• Can the circuit be executed within the coherence and error budget?
• How many repetitions, or shots, are needed to estimate the output?
• Does error mitigation increase sampling cost substantially?
• Is the proposed device available at the required time and service level?
• Can results be reproduced across hardware generations and vendors?

2.3 Resource feasibility

Resources include specialist staff, cloud access, software tooling, classical high-performance computing, data engineering, cryogenic or laboratory infrastructure, and time for verification. Quantum programmes often fail because the organisation budgets for hardware access but not for classical preprocessing, integration, benchmarking, and training.

A simple resource inventory can classify items as existing, purchasable, partner-provided, or unavailable. The “unavailable” category is important: it identifies assumptions that must be tested before a pilot begins.

2.4 Economic feasibility

Total cost of ownership includes direct and indirect costs. A useful model is:

Here, C_access is quantum hardware or cloud-access cost; C_classical is the cost of classical computation and storage; C_development is algorithm and software development; C_integration is connection to operational systems; C_validation is testing and benchmarking; C_governance covers security, compliance, and audit; and C_operation is continuing maintenance and support. All terms should be measured over the same period and in the same currency.

A basic net-benefit estimate is:

where B is the monetised benefit over the evaluation period. This expression is a decision aid, not a complete investment model. Benefits may be uncertain, delayed, or non-financial. A discounted cash-flow analysis can be used when timing and risk are material.

For uncertain outcomes, expected net value can be represented as:

where p_i is the probability of scenario i, B_i is its benefit, and C_i is its cost. The assumptions are that the scenarios are sufficiently defined, probabilities sum to 1, and benefits and costs are expressed on a comparable basis. This model should be accompanied by sensitivity analysis because probability estimates may be subjective.

2.5 Risk feasibility

Risks include technical failure, vendor dependency, information-security exposure, regulatory non-compliance, misleading claims, project delay, skills shortages, and reputational damage. A risk register should identify the event, cause, consequence, likelihood, impact, owner, mitigation, and trigger for escalation.

A common qualitative risk score is:

where L is a chosen likelihood score and I is an impact score. This product is useful for ranking risks only when the scales are defined consistently; it is not a probability of loss unless the scores have been calibrated accordingly.

3. Benchmarking Quantum Proposals

Benchmarking must compare complete, relevant systems. A quantum circuit’s gate count or qubit count alone is rarely meaningful. A benchmark protocol should specify:

1. The problem distribution and instance sizes.
2. The classical baseline, including hardware, software, and tuning effort.
3. The quantum hardware, compiler, error-control method, and number of shots.
4. The metric, such as runtime, energy, approximation error, success probability, or cost.
5. Statistical uncertainty and repeatability procedures.
6. The point at which data preparation and post-processing are included.

For an estimator based on repeated independent measurements, the standard error of a sample mean is commonly written as:

where s is the sample standard deviation and m is the number of shots. This relationship assumes observations are sufficiently independent and that the sample standard deviation is an appropriate estimate of variability. Increasing shots reduces statistical error, but it does not remove systematic hardware bias.

A benchmark should also report confidence intervals or uncertainty bounds where appropriate. A result that is 2% faster but varies by 5% between runs does not establish a reliable advantage.

Benchmark traps

Several practices produce misleading conclusions:

• Comparing a tuned quantum prototype with an untuned classical baseline.
• Excluding data encoding, queue time, or classical post-processing.
• Selecting only favourable problem instances.
• Treating simulator results as hardware results.
• Reporting a speedup without specifying whether it is circuit time or wall-clock time.
• Measuring a noisy output against an ideal answer without accounting for application tolerance.
• Using a vendor-specific metric as if it predicted business performance.

Metrics such as quantum volume, layer-based fidelity measures, and application-oriented benchmark suites can provide information, but none is a universal predictor of value. Benchmark selection must follow the application.

4. Governance for Quantum Computing

Quantum governance is the system of responsibilities, controls, evidence, and decisions that keeps quantum activity lawful, secure, reproducible, and aligned with organisational objectives.

4.1 Governance objects

Governance should cover:

• Data: classification, consent, provenance, retention, and access.
• Algorithms: source control, versioning, assumptions, and licensing.
• Circuits and configurations: compiler versions, device settings, random seeds, and calibration dates.
• Results: raw measurements, metadata, uncertainty, and validation records.
• Claims: who approved public statements and what evidence supports them.
• Suppliers: service availability, data location, audit rights, portability, and incident obligations.
• People: role separation, training, and accountability.

Reproducibility requires more than saving a circuit. A suitable record may include the problem instance, encoding method, circuit, transpilation output, backend identifier, calibration information, shot count, mitigation method, software versions, and classical baseline.

4.2 Stage-gate decision process

A practical programme can use five gates:

A stage gate prevents a compelling demonstration from being treated as a production commitment.

4.3 Standards and conformity

Organisations should monitor relevant national and international standards bodies, cybersecurity guidance, software quality standards, risk-management frameworks, and sector-specific regulation. The precise applicable requirements depend on jurisdiction and industry. Standards should be used to define controls and evidence, not as a substitute for technical understanding.

In procurement, contracts should address service-level objectives, data handling, confidentiality, auditability, incident reporting, portability of circuits and results, and discontinuation plans. Vendor claims should be independently tested where the claim affects a major decision.

5. Security, Ethics, and Sustainability

5.1 Cryptographic transition

A sufficiently capable fault-tolerant quantum computer could threaten widely used public-key systems based on integer factorisation or discrete logarithms through Shor’s algorithm. Symmetric cryptography is affected differently: Grover’s algorithm gives a quadratic search speedup in an idealised setting, so increasing symmetric key sizes is a common mitigation principle. The exact security assessment depends on the algorithm, implementation, attack model, and security level.

The immediate security issue is not that such a machine is already available. It is the “harvest now, decrypt later” risk: adversaries may store encrypted traffic today and attempt decryption when capabilities improve. Organisations should inventory cryptographic use, prioritise long-lived sensitive data, establish crypto-agility, and migrate toward standardised post-quantum cryptographic mechanisms according to authoritative security guidance.

Quantum key distribution is not a universal replacement for post-quantum cryptography. It requires specialised infrastructure, has operational constraints, and does not by itself solve endpoint compromise, authentication, denial of service, or poor key management.

5.2 Ethics and responsible claims

Ethical evaluation includes fairness, privacy, explainability, accountability, accessibility, and social impact. Quantum optimisation applied to public services, lending, employment, or policing can reproduce or amplify unfair objectives if the underlying data and constraints are biased. A quantum method does not make an objective ethically neutral.

Responsible communication should distinguish:

• a theoretical speedup from an observed speedup;
• a laboratory demonstration from production readiness;
• a quantum subroutine from an end-to-end quantum solution;
• a vendor forecast from independently verified evidence; and
• a future possibility from a present capability.

Claims should state scope, assumptions, comparator, uncertainty, and limitations. Marketing language should never imply guaranteed advantage without evidence.

5.3 Sustainability

Quantum systems may require substantial cooling, control electronics, laboratory infrastructure, and classical computation. Sustainability assessment should include embodied equipment, facility energy, cooling, network use, repeated shots, failed experiments, and the carbon intensity of electricity. A quantum method should be compared with the full resource use of the best relevant classical method, not merely with the energy used by a quantum processor during a circuit.

Useful practices include workload scheduling, simulator filtering before hardware execution, reuse of calibration data where valid, transparent energy accounting, and selecting the least resource-intensive method that meets the requirement.

6. Hybrid Quantum–Classical Computing

Near- and medium-term quantum systems are generally hybrid. Classical computers prepare data, optimise parameters, schedule jobs, compile circuits, perform error mitigation, and interpret results. The quantum processor performs a selected subroutine.

A variational workflow illustrates this arrangement. A parameterised circuit prepares a state, measurements estimate an objective, and a classical optimiser updates parameters. In abstract form:

Here, θ_k is the parameter vector at iteration k, and f(θ_k) is an estimated objective obtained from quantum measurements. The equation describes a control loop, not a guarantee of convergence. Noise, barren plateaus, optimiser choice, parameter concentration, and measurement cost can all limit performance.

Hybrid design principles include:

• keep data movement between classical and quantum systems small;
• identify which subroutine truly requires quantum processing;
• measure total latency rather than processor time alone;
• design for backend substitution where possible;
• retain a classical fallback;
• validate outputs against exact or high-quality approximate methods on smaller instances.

7. Quantum Artificial Intelligence

Quantum AI includes several distinct areas: quantum machine learning, quantum-enhanced optimisation for machine learning, quantum simulation of learning models, and the use of AI to control or calibrate quantum systems. These should not be conflated.

Potential approaches include quantum feature maps, variational quantum classifiers, quantum kernels, and quantum generative models. Their practical value depends on data loading, trainability, sample complexity, noise, and the quality of the classical baseline. Classical machine learning is highly optimised, so a small circuit demonstration does not establish an advantage on realistic data.

A responsible evaluation asks whether the quantum representation improves validation performance, sample efficiency, energy, latency, or another operational metric. It should use held-out data and prevent leakage between training and testing. The baseline should include strong classical models rather than a weak comparator selected to favour the quantum method.

AI can also support quantum engineering by detecting calibration drift, designing experiments, optimising compilation, and estimating error. These applications may deliver value earlier than fully quantum AI models because they use quantum technology as the object of optimisation rather than as an unexplained source of predictive power.

8. Quantum Networking

Quantum networking connects quantum processors, sensors, or memories using quantum states and, in some architectures, entanglement. It is conceptually different from sending classical bits over a faster network. Unknown quantum states cannot generally be copied perfectly, and entanglement distribution requires loss management, synchronisation, memory, and suitable interfaces.

Possible applications include distributed quantum computing, secure communication protocols, clock synchronisation, and networked sensing. A network evaluation must consider fidelity, latency, loss, memory lifetime, link rate, repeaters, error correction, and classical control traffic.

The long-term vision of a quantum internet is uncertain and technically demanding. Near-term networking may be more accurately described as specialised point-to-point experiments or interconnect research. Organisations should avoid treating a demonstration of entanglement over a link as evidence of a deployable general-purpose quantum network.

9. Fault Tolerance and Roadmaps

Physical qubits are noisy. Fault-tolerant quantum computing seeks to encode logical qubits into multiple physical qubits so that errors can be detected and corrected without directly measuring the encoded quantum information. A logical operation must be reliable enough that correction does not create more errors than it removes.

A useful conceptual distinction is between:

• physical error rate: the error probability associated with a physical operation or measurement;
• logical error rate: the residual error probability after encoding and correction;
• threshold: a regime in which increasing code size can reduce logical error, subject to assumptions of the code and noise model;
• resource overhead: additional physical qubits, operations, time, measurement, and control required for protection.

A roadmap should therefore report logical performance and resource estimates, not only physical-qubit counts. It should specify target algorithms, logical-qubit requirements, tolerated logical error rates, error-correction code, decoder assumptions, interconnect needs, and expected classical-control burden.

Roadmaps are conditional forecasts, not delivery guarantees. Their uncertainty arises from fabrication yield, correlated errors, leakage, control complexity, software maturity, manufacturing scale, and economic constraints. A professional roadmap includes milestones that can be independently measured and fallback decisions if milestones are missed.

10. Illustrative Case: Quantum-Assisted Portfolio Rebalancing

This is a fictional illustrative case. A financial institution wishes to rebalance a portfolio while satisfying limits on risk, transaction count, sector exposure, and liquidity. A team proposes a quantum approximate optimisation algorithm (QAOA) formulation.

Background and problem

The existing classical workflow uses mixed-integer optimisation and produces a feasible portfolio overnight. The proposed quantum workflow maps binary asset-selection decisions into a cost function, executes a parameterised circuit, and sends measured candidate solutions to a classical feasibility and risk-checking system.

Analysis

The business objective is not “run QAOA”; it is to obtain a portfolio of acceptable quality within the overnight window. The evaluation therefore defines:

• solution objective value;
• feasibility rate after all constraints;
• wall-clock time, including queue and post-processing;
• cost per completed optimisation;
• stability across market scenarios; and
• comparison with the tuned incumbent solver.

The team identifies several risks. Constraint penalties may distort the objective. Hardware noise may reduce feasible samples. Data preparation and classical repair may dominate runtime. Financial decisions require audit trails and model-risk controls.

Approach

The team first generates small instances for which exact classical solutions are available. It then compares QAOA with a strong classical heuristic using identical instances and a pre-registered evaluation protocol. Quantum execution is limited to a pilot budget. Every result records circuit version, backend, calibration date, shots, parameter optimiser, and post-processing.

A deployment gate requires either a statistically credible improvement in the defined operational metric or a separate strategic justification, such as capability building. If the quantum method does not meet the gate, the organisation retains the classical workflow and documents the negative result.

Outcome

In this illustrative case, the pilot does not demonstrate production advantage. The hybrid workflow produces feasible candidates, but classical repair and repeated sampling eliminate the apparent processor-time benefit. Nevertheless, the pilot yields reusable tooling, a cryptographic inventory, trained staff, and a validated benchmark suite. The organisation decides to continue limited research while avoiding a production claim.

Learning points

1. The complete workflow, not the circuit, determines value.
2. A negative benchmark is useful evidence when the protocol is rigorous.
3. Business constraints must be represented and tested explicitly.
4. Governance and documentation are part of technical readiness.
5. Capability development can be a legitimate strategic outcome, but it should not be mislabelled as computational advantage.

11. A Practical Adoption Roadmap

A diploma-level practitioner can structure a roadmap into four horizons.

Horizon 1: Prepare. Establish governance, inventory cryptography, identify high-value problems, train staff, and create classical baselines.

Horizon 2: Explore. Implement small prototypes, use simulators and accessible hardware, test formulations, and record reproducible evidence.

Horizon 3: Validate. Run pre-registered benchmarks, include end-to-end costs, conduct security and ethical reviews, and apply stage gates.

Horizon 4: Deploy selectively. Operate only workflows that meet technical, economic, compliance, and reliability requirements. Maintain a classical fallback and monitor performance after deployment.

Best practice is to treat quantum adoption as a portfolio of options rather than a single irreversible bet. Small, evidence-producing experiments can preserve strategic flexibility while limiting financial and reputational exposure.

Challenges and Limitations

Quantum evaluation remains difficult because hardware changes rapidly, terminology is inconsistent, and many proposed algorithms have uncertain scaling behaviour. Benchmark results may not transfer between devices. Access queues and cloud pricing can change. Some benefits are strategic or scientific and resist simple monetisation. Security guidance must evolve as cryptographic standards and threat assessments develop.

The largest limitation is often not the absence of a quantum processor but the absence of a well-defined problem, strong baseline, reliable data, or operational integration. Sound governance cannot remove physical noise, and excellent hardware cannot compensate for an invalid business case.

Chapter Summary

Industry evaluation of quantum computing requires end-to-end reasoning. A credible use case begins with an important problem and a precise success metric, then evaluates formulation, hardware feasibility, resources, cost, risk, security, ethics, sustainability, and evidence. Computational, practical, and business advantage are different claims and must not be conflated.

Benchmarking should use relevant problem instances, strong classical baselines, complete workflow costs, uncertainty reporting, and reproducible records. Governance covers data, circuits, software, suppliers, results, claims, and accountability. Security planning should include cryptographic inventory, crypto-agility, and migration to appropriate post-quantum mechanisms. Ethical practice requires fair objectives, privacy protection, and responsible communication. Sustainability requires accounting for the whole system.

Hybrid quantum–classical computing is the dominant near-term model. Quantum AI and quantum networking offer important research directions but require disciplined evaluation. Fault-tolerant roadmaps must address logical performance and resource overhead rather than physical-qubit counts alone. A staged adoption roadmap allows organisations to learn, benchmark, and prepare without overstating present capability.

Key Terms

Quantum advantage: Demonstrated superiority on a defined computational task under stated conditions.

Practical advantage: Improvement by the complete implemented workflow, including integration and overhead.

Business advantage: Organisational value produced by a technology or workflow.

Baseline: The reference method against which a proposed method is compared.

Benchmark: A defined, repeatable evaluation using specified instances, metrics, and procedures.

Total cost of ownership: All direct and indirect costs incurred over the relevant period.

Expected net value: Probability-weighted benefit minus cost across defined scenarios.

Risk register: A structured record of risks, impacts, owners, and mitigations.

Governance: Controls and responsibilities that guide lawful, accountable, and reliable use.

Crypto-agility: The ability to change cryptographic mechanisms without redesigning an entire system.

Post-quantum cryptography: Cryptographic methods designed to resist attacks from quantum and classical computers.

Harvest now, decrypt later: Collecting encrypted information today for possible future decryption.

Hybrid quantum–classical computing: A workflow in which classical and quantum processors perform complementary tasks.

Variational circuit: A parameterised quantum circuit whose parameters are adjusted by a classical procedure.

Quantum machine learning: Machine-learning methods involving quantum states, circuits, or quantum subroutines.

Quantum networking: Connecting quantum devices or nodes using quantum communication resources.

Physical qubit: A hardware-level quantum information carrier subject to noise.

Logical qubit: An error-protected qubit encoded using multiple physical qubits.

Fault tolerance: Reliable computation achieved despite physical errors through encoding and correction.

Responsible claim: A statement whose scope, evidence, assumptions, and uncertainty are made clear.

Review Questions

Short Answer Questions

1. Distinguish computational advantage, practical advantage, and business advantage.
2. Why must a quantum benchmark include a strong classical baseline?
3. Name four objects that should be covered by quantum governance.
4. What is meant by “harvest now, decrypt later”?
5. Why are logical-qubit metrics more informative than physical-qubit counts for fault-tolerant roadmaps?

Long Answer Questions

1. Explain how feasibility, resources, cost, and risk should be combined when evaluating a proposed quantum use case.
2. Discuss the main sources of misleading quantum performance claims and propose controls against them.
3. Explain the role of hybrid quantum–classical computing and analyse its principal overheads.
4. Compare the security roles of post-quantum cryptography and quantum key distribution.
5. Describe a staged governance and adoption roadmap for an organisation beginning quantum experimentation.

Analytical / Application Questions

1. An organisation estimates a benefit of $120,000 and total costs of $150,000 for a quantum pilot. Analyse why the pilot might still be strategically justified, and identify evidence required before continuation.
2. Design a benchmark protocol for a quantum optimisation proposal. Specify the baseline, instances, metrics, uncertainty reporting, and deployment gate.
3. A vendor reports a large speedup based only on quantum processor execution time. Identify the missing workflow components and explain how you would test whether the claim represents practical advantage.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes business advantage from computational advantage?

A. Business advantage concerns organisational value, while computational advantage concerns performance on a defined task.

B. Business advantage is always measured in qubits, while computational advantage is measured in revenue.

C. Business advantage applies only to hardware manufacturers.

D. There is no meaningful distinction between the two terms.

Correct Answer: A. Business advantage concerns organisational value, while computational advantage concerns performance on a defined task.

Question 2

Which item should be included in an end-to-end quantum benchmark?

A. Only the number of physical qubits

B. Only the quantum processor execution time

C. Data preparation, quantum execution, post-processing, and a strong classical baseline

D. The vendor’s advertised roadmap date

Correct Answer: C. Data preparation, quantum execution, post-processing, and a strong classical baseline

Question 3

In the expression TCO = C_access + C_classical + C_development + C_integration, what does C_integration represent?

A. The cost of connecting the workflow to operational systems

B. The cost of measuring qubit coherence only

C. The cost of symmetric encryption keys

D. The probability of hardware failure

Correct Answer: A. The cost of connecting the workflow to operational systems

Question 4

What is the main concern represented by “harvest now, decrypt later”?

A. Quantum computers may permanently delete stored data.

B. Attackers may store encrypted data now and decrypt it when capabilities improve.

C. Classical computers may stop supporting encryption immediately.

D. Quantum networks cannot transmit classical messages.

Correct Answer: B. Attackers may store encrypted data now and decrypt it when capabilities improve.

Question 5

Which statement about quantum key distribution is correct?

A. It universally replaces post-quantum cryptography.

B. It removes the need for endpoint security.

C. It requires specialised infrastructure and does not solve every security problem.

D. It guarantees protection against denial-of-service attacks.

Correct Answer: C. It requires specialised infrastructure and does not solve every security problem.

Question 6

In a hybrid variational workflow, what usually updates circuit parameters?

A. A classical optimisation procedure

B. A random hardware reset with no measurement

C. A network router alone

D. The cryogenic system without classical control

Correct Answer: A. A classical optimisation procedure

Question 7

Why can a quantum machine-learning demonstration fail to establish practical value?

A. Quantum states cannot contain information.

B. Data loading, noise, training cost, and strong classical baselines may eliminate the apparent benefit.

C. All machine-learning tasks are solved exactly by quantum circuits.

D. Validation data are never needed for quantum models.

Correct Answer: B. Data loading, noise, training cost, and strong classical baselines may eliminate the apparent benefit.

Question 8

What is the purpose of a logical qubit?

A. To represent an error-protected qubit using multiple physical resources

B. To reduce every circuit to one physical gate

C. To eliminate the need for measurements

D. To describe only classical memory

Correct Answer: A. To represent an error-protected qubit using multiple physical resources

Question 9

Which is a responsible technology claim?

A. The technology guarantees transformation of every optimisation problem.

B. A vendor forecast is equivalent to independently verified production evidence.

C. The reported result specifies its baseline, scope, assumptions, and uncertainty.

D. A simulator result automatically proves hardware advantage.

Correct Answer: C. The reported result specifies its baseline, scope, assumptions, and uncertainty.

Question 10

If a risk-ranking score is defined as R = L × I, what do L and I represent?

A. Latency and iteration

B. Likelihood and impact

C. Logic and integration

D. Learning and inference

Correct Answer: B. Likelihood and impact