The Deep Tech Scale Readiness Model

Four stages and a changing standard of proof

Technology readiness levels provide a vocabulary for the maturity of a technology in its intended environment. They do not describe every condition needed for a viable business. NASA's framework concentrates on technical maturity, while the US Department of Energy's adoption-readiness approach also examines demand, resources and wider barriers to deployment. Manufacturing-readiness practice adds another essential test: whether production capability is supported by evidence.

At the prototype stage, the company is establishing that the critical functions work and that they solve a worthwhile customer problem. Engineering judgement and manual intervention are expected. The important management task is to make those dependencies visible. A prototype build should reveal which materials, tolerances, skills and operating conditions are likely to become difficult as volume increases.

Industrialisation translates the design into a controlled way of making and delivering the product. Architecture and interfaces are baselined, suppliers are qualified, test methods are established and work instructions are tried by the people who will use them. A baseline permits disciplined change. It does not prevent learning or force a premature final design. The exit question is whether the company can undertake a defined first production run with appropriate controls and funding.

Low-rate initial production, or LRIP, tests whether that operating system repeats. The company needs evidence from representative builds, operators, material lots and customer deployments. It must understand the difference between units started, units passing first time and units eventually accepted after rework. The quantity required to establish confidence will depend on the process, failure consequences and available data. There is no universal batch size that makes a deep-tech business ready to scale.

Scale then becomes an ongoing management responsibility. The business must sustain performance across the approved rate and product mix while controlling changes, supplier risk, service obligations and liquidity. A new factory, supplier or use environment can invalidate part of the earlier evidence. Readiness should therefore be assessed for a defined product family, production route and market, and revisited when those conditions change.

The model tracks the progression across the original nine operating dimensions, with commercial demand and deployment/service added explicitly. Those two additions matter because a company can manufacture successfully and still fail to convert deliveries into repeat business or cash. Read the columns as cumulative obligations, not as departments completing independent checklists.

The eleven dimensions, stage by stage

Read left to right as the evidence strengthens. Read down to expose the constraint on the next commitment.

What must mature — Prototype: critical functions work. Industrialisation: the production route works. LRIP: delivery becomes repeatable. Scale: performance stays controlled.

Technology and delivery — Product: critical functions demonstrated; architecture baselined and verified; released configuration proven; changes and reliability controlled. Manufacturing: build route and risks understood; DFM and pilot process proven; repeatable accepted output; rate and bottlenecks managed. Supply chain: critical inputs and sources mapped; BOM and suppliers qualified; production supply demonstrated; resilience and capacity secured. Quality: acceptance and failure capture; QMS and control plans operating; traceable release and containment; prevention and improvement. Regulatory: requirements and route mapped; evidence and permissions planned; release permitted for scope; compliance lifecycle controlled.

The operating business — Organisation: accountability beyond founders; functional ownership established; operations team can deliver; management capacity scales. Systems: records and versions controlled; core workflows linked; transactions and traceability reliable; information integrated and trusted. Economics: cost and price hypotheses; BOM and conversion costed; actual accepted-unit cost known; margin and capital productivity. Cash: funding to next evidence milestone; cash needs and commitments modelled; ramp and downside funded; cash cycle and liquidity controlled.

Commercial delivery — Commercial demand: customer problem and buyer tested; paid use case and terms validated; repeatable demand supports ramp; profitable demand and mix managed. Deployment and service: use environment understood; installation and support designed; customer acceptance repeatable; fleet and service costs controlled.

Gate decisions — G1: fund industrialisation. G2: authorise bounded LRIP. G3: authorise a defined ramp. G4: renew or revise scope.

Readiness is limited by the evidence required for the next step. Critical gaps cannot be offset by a strong average.

Readiness lives in the handovers

A design change creates variation. A process change creates rework and retest. A delivery delay means later acceptance. Later acceptance means more cash exposure and more funding required.

The most revealing gaps often sit between functions. Engineering approves a component change, but purchasing has already committed to the previous revision. A supplier reports acceptable quality, but its variation makes final calibration slower. Sales promises an installation date without checking the availability of the only engineer qualified to commission the system. Each team can look competent while the customer experiences an unreliable business.

That is why the model should be reviewed across functions. A product change needs a manufacturing, quality, supply and cash assessment. A production ramp needs a credible plan for test capacity and field support. A lower unit price needs to be reconciled with actual yield and payment terms. The most valuable review follows one customer order through design, procurement, build, release, installation and collection, asking where the evidence or ownership breaks down.

Test and calibration deserve particular attention in precision hardware. Adding assembly capacity creates little benefit if accepted output is limited by a specialist test station or repeated retesting. A useful capacity study follows good units through the whole route, includes downtime and rework, and uses the product mix the company actually intends to sell. The investment decision may then favour a better fixture, a more stable process or improved diagnostics over a larger machine.

Supply resilience requires the same specificity. A second supplier only reduces risk when its output is qualified, its capacity is available and switching is permissible. Where a viable alternative does not exist, a deliberate single-source strategy may be stronger than an unqualified backup. Reserved capacity, ownership of tooling, access to process knowledge, agreed recovery arrangements or a selective inventory buffer can form part of that strategy. Their cost and residual risk should be explicit.

Systems should make these handovers dependable. Accurate part numbers, revision rules, inventory records and release authority matter before the choice of software. A small company may operate effectively with simple tools if its controls are reliable; a larger business may need integrated planning, quality and finance systems. In either case, quality management should operate through the business processes themselves.

The resulting evidence is more useful than a list of systems installed or roles hired. Can the team identify every unit affected by a suspect component? Can production continue without routine founder intervention? Can it show what an accepted unit cost? Can it commission the units it plans to ship? The answers reveal which capability must improve before the next commitment.

Measure the cost of an accepted unit

The economics of a prototype are easy to misread when the bill of materials becomes a substitute for production cost. Purchased components are only one part of the cost of making an accepted unit. Labour, test, scrap, rework and production overhead can materially change the result. Installation, warranty and service create further obligations after the product leaves the factory.

At a selling price of £1,500, comparing price with the £600 material bill suggests a 60% margin. The production gross margin is actually 33.3%. Add £90 per sold unit for incremental warranty, service and logistics, and the defined margin after fulfilment is 27.3%, before research, selling and corporate costs. The illustration assumes all accepted units are sold and no work in progress or saleable stock remains.

This changes the investment question. Management needs to establish which improvement reduces the observed cost and whether it will survive at the next production rate. Better first-pass yield may free test capacity as well as reduce labour. A cheaper component may increase calibration effort. Automation may remove a labour step while introducing maintenance and integration costs. The business case should connect the proposed investment to measured operating losses, with explicit assumptions for yield, utilisation and product mix.

The cash required to grow

Even a viable unit margin does not establish that a ramp is affordable. Suppliers may require deposits, inventory may sit through a long build cycle, and customers may pay only after installation and acceptance. Growth increases the amount of cash committed to that sequence. The relevant funding question is how much cash the business will need at its lowest point, and whether that funding will be available then.

In the illustration, operating working capital rises from £440,000 at 100 units per month to £1.32 million at 300 units per month — an additional £880,000, excluding capex, tax, deposits, operating losses and cash reserves.

The cash conversion cycle adds inventory days and receivable days, then subtracts payable days. It is a useful timing measure, but the cash amounts must be modelled on their appropriate bases: receivables reflect selling prices, while inventory and payables reflect different cost or purchase values. Multiplying the cycle by one daily cost figure can misstate the requirement.

For a real gate decision, the company should maintain a 13-week cash forecast and a forecast through the next evidence milestone. Lower yield, delayed acceptance, slower collections and capex timing should be tested together where they could occur together. Deposits, milestone billing and staged capacity can reduce exposure, but their feasibility must be established with customers and suppliers. A fundraising intention does not close a funding gap.

Give each gate a capital decision

Each dimension is scored on a simple scale: 0 — absent or failed; 1 — planned; 2 — demonstrated with gap; 3 — demonstrated and accepted.

The model becomes useful when each review authorises something specific. The prototype gate can release an industrialisation budget. The industrialisation gate can permit a capped LRIP quantity. The LRIP gate can approve a defined production rate, product mix and investment tranche. At scale, the review renews that authority or limits expansion when performance deteriorates. The decision should name the scope, spending limit, accountable approver and date for review.

Evidence quality matters as much as the score. A pilot result from one operator or one material lot may provide limited confidence about the next rate. NIST's process-capability guidance ties capability assessment to a stable process and appropriate statistical assumptions and data. Small LRIP populations require transparent counts, test conditions and uncertainty; an attractive capability index should not substitute for that explanation.

A conditional decision should consequently fund bounded learning whose remaining risk is understood. It should not become a standing exception that expires only when a customer complains. If a test fails, a due date slips or the scope changes, the review needs to reopen the authority to proceed.

Build the evidence before expanding the commitment

A leadership team can begin by choosing one forthcoming decision: a larger batch, a new supplier, a production line or a customer commitment. It should define the intended rate and configuration, collect the available build and commercial records, and ask each accountable owner to score the evidence before the group meets. Disagreement is useful when it exposes a different assumption about what the business is being asked to do.

The first review should produce a short sequence of evidence-building work. That might be a representative production run, a supplier capacity trial, a cost reconciliation or a test of customer acceptance. Priority should follow the critical path, the consequence of failure and the cash at risk. Improving an already strong dimension has little value if a different constraint still prevents the decision.

Boards and investors can use the same logic to structure funding. Each tranche should buy a defined reduction in uncertainty or the capacity justified by that reduction. This makes the capital request more intelligible: what is being committed, which assumption the money will test, and what the company will be able to do once the evidence is accepted.

The practical test of scale readiness is whether the company can repeat its promise under the conditions it plans to operate. A working prototype begins that proof. A scalable business sustains it through production, customer use and the collection of cash. The next investment should strengthen the part of that chain that currently limits the business.

Sources and further reading

NASA, Technology Readiness Levels.

US Department of Energy, Adoption Readiness Levels.

US Government Accountability Office, Manufacturing risk management, GAO-10-439, 2010.

NIST, Engineering Statistics Handbook, process capability.

ISO, Quality management.

BDC, Cash conversion cycle.

The stages and scoring rules are an original management framework. All numerical examples are illustrative. Sources reviewed 9 September 2026.