A quantum program can consume indefinite money if every milestone is framed as “research progress.” Executive governance converts curiosity into staged investment.
The purpose is not to make quantum conservative. The purpose is to fund uncertainty deliberately.
50.1 Investment thesis
View diagram source
flowchart LR
Thesis[Investment thesis] --> Option[Strategic option value]
Thesis --> Capability[Internal capability]
Thesis --> UseCase[Use-case discovery]
Thesis --> Risk[Quantum risk reduction]
Thesis --> Product[Future product platform]Executives should state which thesis they are funding. A program designed for post-quantum cryptography migration is different from one designed for chemistry research, vendor evaluation, cloud platform development, or fault-tolerant service readiness.
50.2 Stage gates
View diagram source
flowchart LR
Learn[Learn] --> Prototype[Prototype]
Prototype --> Validate[Validate]
Validate --> Productize[Productize]
Productize --> Scale[Scale]
Scale --> Operate[Operate]| Stage | Funding question |
|---|---|
| learn | what capability must we build to make better decisions? |
| prototype | can we run meaningful workflows end to end? |
| validate | can evidence beat relevant baselines or reduce risk? |
| productize | can we expose a reliable internal or external service? |
| scale | can demand, cost, and quality scale together? |
| operate | can we govern, monitor, and improve the service? |
Qiskit Serverless, Braket reservations, and Azure Quantum jobs are examples of cloud-era execution and orchestration primitives. Executive roadmaps should evaluate how such primitives reduce internal build burden while preserving evidence, governance, and portability. [R76] [R77] [R78]
50.3 Portfolio model
A mature quantum portfolio has multiple lanes.
View diagram source
flowchart TB
Portfolio[Quantum portfolio] --> Risk[PQC and security risk]
Portfolio --> Research[Scientific research]
Portfolio --> Platform[Platform capability]
Portfolio --> Vendor[Vendor evaluation]
Portfolio --> Product[Product opportunities]
Portfolio --> Talent[Workforce development]Do not mix these lanes into a single ROI claim. Some lanes are defensive, some are exploratory, some are infrastructure, and some may become products.
50.4 Kill criteria
A program without kill criteria will continue because it exists.
investment_gate:
stage: prototype | validate | productize | scale
continuation_criteria:
- evidence threshold
- cost threshold
- capability threshold
kill_criteria:
- baseline not beaten after defined attempts
- cost exceeds option value
- vendor dependency unacceptable
- platform risk unmanaged
pivot_options:
- change use case
- change vendor
- simulate instead of execute
- convert to training programView diagram source
flowchart LR
Gate[Investment gate] --> Continue{Criteria met?}
Continue -- yes --> Fund[Continue funding]
Continue -- partially --> Pivot[Pivot]
Continue -- no --> Stop[Stop or archive]
Pivot --> NewGate[New gate]Stopping a workload is not failure. Continuing a workload with no evidence is failure.
50.5 Executive dashboard
View diagram source
flowchart TB
Dashboard[Executive dashboard] --> Capability[Capability maturity]
Dashboard --> Evidence[Evidence quality]
Dashboard --> Cost[Cost and capacity]
Dashboard --> Risk[Security and compliance risk]
Dashboard --> Vendor[Vendor dependence]
Dashboard --> Talent[Talent readiness]The dashboard should avoid vanity metrics.
| Avoid | Prefer |
|---|---|
| total qubits | validated workload capacity |
| number of experiments | accepted evidence packages |
| vendor demos | contract-tested benchmarks |
| press claims | reviewed external claims |
| raw spend | spend per qualified learning milestone |
| roadmap promises | gate-based roadmap confidence |
50.6 Fault-tolerant readiness
Fault-tolerant readiness is not waiting for a magic machine. It is preparing the organization to use logical services when they become real.
View diagram source
flowchart LR
Today[NISQ and early utility] --> Logical[Logical-qubit experiments]
Logical --> Services[Fault-tolerant service candidates]
Services --> Products[Qualified quantum products]Google’s below-threshold and dynamic-surface-code work shows that the field is making concrete progress on error-correction operations, but a business roadmap still needs gates: logical error rate, decoder latency, logical operation set, resource estimates, service availability, cost per useful result, and application-level advantage. [R90]
50.7 Board-level questions
Executives and boards should ask:
- What decisions has the quantum program improved?
- What risks has it reduced?
- Which workloads have evidence packages?
- Which claims are we willing to make externally?
- Which vendors are strategically important and why?
- What is our PQC migration status?
- What would cause us to reduce, redirect, or increase investment?
- What capability do competitors or partners have that we lack?
View diagram source
flowchart TB
Board[Board review] --> Thesis[Confirm thesis]
Board --> Evidence[Inspect evidence]
Board --> Risk[Inspect risk]
Board --> Spend[Inspect spend]
Board --> Decision[Fund, pivot, or stop]50.8 Closing principle
The quantum program should produce three assets:
- evidence,
- capability,
- optionality.
If it produces only excitement, it is under-governed. If it produces only papers, it is under-integrated. If it produces only infrastructure, it is under-directed. The operating discipline of the book is designed to keep all three assets in view.