A new printer, funding round, facility or multi-year partnership is not production capacity by itself. Capacity becomes a useful business commitment only when equipment, material, people, post-processing, inspection and demand can jointly deliver accepted parts. This guide provides an evidence chain for customers, suppliers and investors to distinguish investment intent from operating, qualified output.

1. Define what capacity means

Build volume, laser count or kilograms per hour is not a capacity metric on its own. Define the product family, material, quality requirement, shift pattern, available days and acceptance rule. One machine may make many simple prototypes or a few critical parts with long post-processing; annual unit counts are not directly comparable.

A useful definition is first-pass accepted parts per unit time against a stated requirement. Mass, build volume and machine hours are supporting metrics. When products share a machine, document product mix and nesting assumptions.

2. Build a state model from announcement to release

An investment passes through at least seven states: funded or ordered; delivered; installed; safe to operate; technically accepted; qualified for a process; and repeatedly producing. Each state needs different evidence. A purchase order is not delivery, factory acceptance is not site performance, and one coupon is not serial capability.

Assign an owner, date, input, output and exit criterion to every gate. Press releases often compress several states into one sentence; due diligence must separate them.

3. Three levels of equipment acceptance

Installation qualification establishes that the machine, utilities, extraction, gas, network and safety systems match the approved configuration. Operational qualification checks calibrated ranges, alarms, repeatability and software. Performance qualification uses representative materials and parts to show that the complete cell meets requirements.

Metal powder-bed fusion may require laser power and field alignment, oxygen, recoating and multi-position coupons. Polymer VPP needs optical dose, resin temperature, washing and post-cure; FGF needs drying, mass flow, thermal mapping and shrinkage. The checklist changes by technology, but the logic is the same.

4. The bottleneck is often outside the printer

Measure preparation, cooling, unpacking, depowdering, washing, heat treatment, support removal, machining and inspection alongside print time. If a new printer can create three builds per day while the furnace or CMM closes only one, system output does not triple.

Map waiting time, changeover, batch size and rework. Buffers can conceal the issue: printer utilisation may be high while inventory grows before the next operation.

5. People, recipes and digital provenance

Each shift needs trained operators, maintenance, material control and release authority. Record a competence matrix, supervised build requirement and cover for absence. A process dependent on one expert may have high nominal capacity and low resilience.

Machine file, slicer or build-processor version, material lot, parameters and post-process recipe must remain traceable. Distributed manufacturing must prove the same approved package and deviation control at each site; file portability is not process equivalence.

6. Capacity starts with yield

For a practical estimate, take scheduled available time, subtract maintenance and changeover, and divide by total cycle time. Multiply by first-pass yield and a nesting factor reflecting actual product mix. Repeated build failures, reprints and inspection losses consume capacity directly.

Report nominal and demonstrated capacity separately. Demonstrated capacity should come from a stable period under a defined requirement and measurement system. Use a range and uncertainty for a new installation, not a single annual unit number.

7. Reading financial evidence correctly

Backlog indicates future work, revenue recognised performance, and cash payment. None is the same as qualified physical capacity. A few machine deliveries can create triple-digit growth in a small printer segment; inspect absolute value, base, segment weight and shipment timing.

For a funding round, ask how much goes to equipment, facilities, staff, development and working capital. Without that allocation, the headline amount cannot be converted into productive assets. Treat a company plan as a plan until equipment, headcount and output become verifiable.

8. Measurement and statistical release

Dimensional, density, mechanical and surface data across builds and positions support a capability estimate. This needs capable calibrated measurement, pre-defined sampling and change control. A showcase part or best result is not a distribution.

A minimum evidence package records build ID, machine and software version, material lot, environment, post-process route, nonconformity and release decision. Only then can a capacity claim be audited.

9. Change control during ramp

Supplier, refresh ratio, resin, nozzle, furnace, software or staffing may change during ramp. Each change needs impact analysis, proportionate revalidation and rollback. Fast expansion often moves several variables at once, making failure causes difficult to isolate.

A weekly ramp board should separate availability, performance rate, first-pass yield, total lead time, safety events and customer deviations. OEE is useful only when its definition of a good part matches the application requirement.

10. Procurement and audit checklist

Request evidence of machine state, acceptance records, approved material and recipe lists, post-processing capacity, measurement capability, staffing coverage, first-pass yield, maintenance and capacity assumptions. Delivery commitments should state product mix, priority rules, contingency route and change notification.

Warning signs include machine count, build volume or peak speed without yield; growth percentages without a base; future machines counted as operating; and printing separated from post-processing and inspection.

Editorial conclusion

An AM expansion becomes credible when an auditable chain links financing to repeatedly accepted parts against a requirement. The best comparative metric is not machine count but first-pass accepted product through the full process during a stable period, with uncertainty. This protects customers from inflated promises and manufacturers from committing volume too early.

Sources and scope

This guide draws on ISO/ASTM industrial AM quality-management principles, NIST measurement and digital-thread work, and the risk-based qualification logic of ASTM E2500. Current Roboze, Prodways and NAMIC announcements are case studies, not certifications. Application, contract, material and sector regulation determine the final validation plan.