Knowledge Base

Buy Equipment or Outsource? Run the Utilisation Math, Not Your Gut

2026-08-21

Short answer: if you can keep a machine busy for more than roughly a third of working hours, in-house usually wins on cost per part; below that, outsourcing almost always does. But utilisation is not the only variable — confidentiality, iteration speed, and whether you need a process only once can each decide the answer on their own.

The most common error here is comparing "machine price ÷ expected parts" against an outsourced unit price. That arithmetic omits the largest components of ownership cost, which together usually exceed depreciation itself.

What Ownership Cost Actually Includes

Operator labour — not just pressing start. Nesting, support design, parameter tuning and reprinting failures are skilled hours
Post-processing labour — washing, post-curing, support removal, sanding, coating. Heavy for resin and metal, and routinely omitted entirely
Consumables and spares — beyond resin/powder/filament: LCD panels, nozzles, recoaters, filters, inert gas
Facility and compliance — extraction, dust protection, metal-powder explosion control, noise. Universities and large firms often need a safety review too
Maintenance and downtime — service hours, spare-part lead times, and what happens to the work while a machine is down
Learning curve — the failure rate in the first months is not zero; scrapped material and missed dates in that period are real costs

Consideration	Leans in-house	Leans outsourced
Demand stability	Steady, predictable monthly volume	Lumpy, project-based, on and off
Utilisation	Can fill over a third of working hours	Only a few builds a month
Iteration speed	Three revisions a day; cannot wait on shipping	One revision a week; transit time acceptable
Confidentiality	Cannot leave the building (defence, unreleased products)	An NDA is sufficient
Range of processes	One or two processes used repeatedly	A bit of everything — buying all of it is unrealistic
Capital structure	Capital budget exists; trade capex for low unit cost	Protecting cash flow; per-part payment is more flexible
Staffing	Someone can learn and own the machine long term	No dedicated person — an owned machine will idle
Facility compliance	Can provide extraction, explosion control, pass safety review	Plant conditions do not allow it, especially metal powder

A Practical Middle Path

Many customers end up phasing it: outsource first to establish process parameters and yield and to measure real monthly volume, then decide which machine to buy. Two benefits — the capital decision rests on measured data rather than an estimate, and the parameters and failure lessons accumulated while outsourcing carry straight over, shortening the learning curve once the machine lands.

The reverse combination also works: own one machine for the process you use daily and keep outsourcing the occasional exotic one. That is far more realistic than buying full coverage — no shop needs its own FDM, SLA, SLS and SLM, yet plenty of shops encounter all four kinds of work.

We provide both — equipment sales and a processing service — so we have no stake in pushing you towards a machine. Tell us monthly volume, part types and accuracy requirements, and we will run the numbers both ways for you.

Frequently Asked Questions

Where does the one-third utilisation threshold come from?

It is an experience-based order of magnitude, not a formula: below that level, fixed costs like depreciation and facility spread over each part exceed the margin you would pay a subcontractor. The real threshold shifts with machine price, local labour cost and available quotes, so treat it as a pointer for which direction to model in detail rather than a verdict.

How is design confidentiality handled when outsourcing?

Standard practice is an NDA, an agreement to destroy files after delivery, and where needed sending only the geometry required to machine rather than the full assembly. If your case is one where regulation forbids the data leaving at all — some defence and medical work — that consideration overrides the cost maths outright: buy.

What if we buy the wrong machine?

This is exactly why phasing through outsourcing helps. The two most common mistakes are buying too small a build envelope (parts you regularly make will not fit) and buying the wrong process (SLM bought for work that SLS should do). Both are avoidable with real data from an outsourced period. Sending us the list of parts you actually made in the last six months beats reading spec sheets.

How does this work for a university lab?

Teaching changes the equation: a machine's value is not only the parts it produces but students getting hands-on with the process principle, which outsourcing cannot buy. So teaching needs usually favour ownership — but the selection logic shifts from throughput to how many process principles are covered. Occasional metal parts for research projects, by contrast, are good candidates for outsourcing.