How to choose a 3D printer in 2026 and when buying one is a mistake

Choose a 3D printer on four things: which materials you want to print (ordinary PLA and PETG are fine on an open-frame machine, ASA and nylon need an enclosure), how large the biggest part is, how much time you are willing to spend on maintenance, and how much you will actually print. That last criterion decides whether buying makes sense at all, below roughly five kilograms of material a year, outsourcing is cheaper even after your own time is counted.

Specific models change faster than an article can keep up with. So what follows are the criteria that hold regardless of the year, plus a calculation you can run for your own case.

First, the question most people skip

Before you start comparing machines, answer this: how many parts will I really print, and how often?

An answer of "I don't know, we'll see" means a few dozen hours of printing a year in practice. At that volume a printer is an expensive toy, which is entirely fine if you admit it upfront and are buying it for the pleasure of the process. The problem starts when you buy it as a saving.

The calculation is in the payback section below.

The four criteria that matter

1. The materials you will print

This is the most important and most frequently underrated decision, because it determines the entire category of machine.

If you only plan to print PLA and PETG, an enclosed machine is money wasted. If you ever want to try ASA, buy enclosed straight away. Retrofitting a chamber to an open-frame printer is not worth it.

2. Build volume

Common machines are around 220 × 220 × 250 mm, which covers most parts. More volume costs disproportionately more and lengthens calibration.

More practical than buying a large machine is learning to split parts and bond them, besides the saving, it usually improves layer orientation relative to the load. If you need a large part once a year, order it.

3. Kinematics and speed

Today's market splits into two designs:

• Bedslinger (the bed moves back and forth), cheaper, wobbles on tall parts.
• CoreXY (the bed only moves down), more expensive, more stable at higher speeds, easier to enclose.

Treat quoted mm/s figures with scepticism. Real print time is set by acceleration, cooling quality and whether the machine handles supports well, not by the catalogue top speed.

4. Maintenance and parts availability

A printer is a machine with a nozzle, belts, bearings and a heated bed. All of it wears out eventually.

Ask: are spare parts available in the EU? Are the part models published so you can print them yourself? How long does a nozzle take to arrive? With open-ecosystem brands the answer is simple; with closed systems it can mean weeks of waiting.

Machine categories and who they suit

Prices are indicative bands as of August 2026. Track specific models and their tests on sites that actually measure them, for example Tom's Hardware or TechRadar. Skip reviews that carry affiliate links and no measured figures.

Payback: when the machine pays for itself

This is the calculation you will not find in printer reviews, because it does not sell anything.

Take an enclosed machine at €800. Good filament costs around €25 / kg. Our price for printing medium-sized parts is €0.12 per gram, that is €120 per kilogram of finished prints (machine time, setup and material included).

Eight and a half kilograms is roughly 60 to 150 pieces for ordinary parts. If you print that much in a year, buying makes sense.

The calculation does, however, leave out three items:

• Your time. Model preparation, slicing, supervision, support removal and cleanup. Realistically 20 to 45 minutes per part. At €20 an hour that is €7 to €15 per part, often more than the material difference.
• Failed prints. In our experience, somewhere between a tenth and a fifth of prints fail in the first months. Material and time both go in the bin.
• Running costs. Nozzles, belts, PEI sheets, a dryer for nylon, electricity.

Once time is counted, the real threshold moves to roughly five to ten kilograms a year. Below it, outsourcing is cheaper. Above it, and especially if you enjoy printing, the machine is worth it.

When not to buy a printer even if you can afford one

• You need a part now and then. An idle machine never pays for itself, and after six months of standing still it needs recalibrating anyway.
• You need one material your machine cannot handle. Buying a second printer for one nylon gear costs more than ordering 50 pieces.
• You need accuracy or a surface FDM cannot deliver. Then you need a different technology, not a different FDM machine. See FDM vs SLA.
• You need parts to a deadline with no backup. When your only printer fails the day before handover, there is no plan B.
• You are going into series production. At tens or hundreds of pieces, small-batch production is in a different league both organisationally and commercially than one machine in a garage.

The sensible middle option

Plenty of companies land on the combination that works best: a cheaper machine for quick prototypes and shape checks, outsourced printing for end-use parts.

You print a PLA prototype overnight and immediately know whether the dimensions fit. Once the shape is settled, you order the end-use part in ASA or nylon, without buying an enclosed machine and a dryer for five pieces a year.

Before you decide

Try a simple test: have the parts you would print in your first month made for you. You will learn your real material consumption, the model preparation time and whether the result is good enough, all for a fraction of the price of a machine.

If it turns out you print a lot, you will buy a printer knowing exactly which one you need. If it turns out you print little, you have saved €800. Send us a model and you will have a quote within 24 hours; indicative prices are published upfront.