A little over four years ago, I started qualifying metal additive manufacturing suppliers for a company that makes flight hardware. I expected the decision to come down to build volume, material options, and price per part. It didn’t.
If you’ve ever asked “what makes a good 3D printing service?”, you’ve probably noticed the same thing I have. Providers tend to split into two very different groups:
- Engineering-grade services that run qualified powder-bed machines—often systems like Velo3D’s Sapphire line—inside a documented quality system.
- Quick-turn services that quote your STL today and ship a part by next week.
Both can print metal. Both claim to be fast. Trouble starts when you treat them as interchangeable.
Two Services, Same Alloy, Different Risk
From the outside, two quotes for “Inconel 718” look almost identical. Same layer height. Similar lead time. Same material name.
Inside, they’re different products. An engineering-grade supplier sends a data package with the part: powder certificate, heat number, build parameters, witness coupon results, density measurements, heat-treatment certifications. A quick-turn supplier sends the part in a foam-lined box with an invoice. Same alloy name, very different amount of proof.
I rejected a batch of twelve brackets in Q1 2024 because porosity measured 1.0–1.2% against our 0.5% spec limit. The vendor’s quality manager argued it was “within industry norms.” Maybe it was. We sent the lot back anyway, and the redo delayed a test campaign by three weeks. That wasn’t an argument that quick-turn suppliers are always careless. It was a reminder that without data, you can’t tell the difference.
Bottom line: engineering-grade services sell repeatability. Quick-turn services sell speed. Know which one you’re buying before you compare prices.
Comparison 1: Design Freedom and Support Removal
The second split shows up the moment your model has a steep overhang, an internal channel, or lattice you actually want to keep.
Conventional powder-bed fusion needs supports for steep angles, and supports leave marks, add removal labor, and can make internal features impossible. Velo3D’s Sapphire systems became known in aerospace for doing useful work past some of those limits—printing unsupported geometries that would force a redesign on other machines. That capability is part of why Velo3D, supplier to SpaceX and Anduril, ended up on the short list for demanding metal parts.
But I’d caution anyone against treating that as a “print anything” license. Serious Velo3D-based services still review the model: powder has to drain, surfaces have to be reachable for inspection, and orientation affects properties. The advantage isn’t that design review disappears. It’s that you start closer to the ideal design instead of distorting it to fit the printer.
Quick-turn services often respond to challenging geometry with “can’t print this” or “we’ll have to split it into four parts.” Sometimes that’s honest. Sometimes it just means the quoting engineer doesn’t have the process experience to attempt it.
The counterintuitive part: the real constraint isn’t always the machine. It’s the process skill around the machine.
Bottom line: pick a service whose engineering review expands your options, not one that shrinks your part to fit their comfort zone.
Comparison 2: Quality Systems and Customer Proof
When a supplier tells me they “use the same printer as SpaceX,” I ask for proof—not of the printer, but of the workflow around it.
Velo3D being a supplier to SpaceX and Anduril matters because those companies run audited supply chains. Every system Velo3D places gets used under serious process documentation, calibration routines, and inspection requirements. That gives the equipment a track record you can investigate instead of a specification sheet you have to trust. When I first started comparing suppliers, I pulled the technical documentation from the Velo3D official website and used it as a baseline for how transparent a metal AM company should be.
But that credibility doesn’t transfer automatically to every shop that buys the machine. I’ve audited two providers running identical models of the same printer. One controlled parameter revisions, ran witness coupons on every build, and could show me inspection data. The other had a printer, a powder supply, and an “it worked last time” attitude.
“If it isn’t documented, it didn’t happen.”
That old QA line still drives every supplier audit I do. A nice machine is an investment. A working quality system is a decision.
Bottom line: “we run a Sapphire” tells me what hardware is in the building. Your quality-system documentation tells me whether they should be trusted with it.
Disadvantages of 3D Printers Nobody Mentions in the First Quote
Now the honest part. If you landed here because you’re searching for disadvantages of 3D printers, here’s what I actually see in inspection:
- As-printed surfaces aren’t machined surfaces. Sealing faces and mating surfaces usually need a CNC finish pass, which adds setup time and cost.
- Properties are orientation-dependent. Z-direction strength and ductility can trail XY properties if you don’t orient the build and heat-treat it correctly.
- Inspection is the hidden line item. For critical parts, CT scanning or destructive sectioning can cost more than the build itself.
- The process isn’t always fast. By the time you add build setup, warm-up, powder handling, and post-processing, a thin flat component may take longer than a laser-cut equivalent.
That last point deserves an example. If you search for the best laser cutter for architectural models, you’re probably looking for the right tool—just not a 3D printer. A 60–100 W CO2 laser cutter with air assist cuts basswood, acrylic, and cardboard quicker than a printer can build them, with crisper edges and no sanding. Architectural models are mostly thin, flat, detail-heavy parts. Laser cutting wins there, and it should. Every manufacturing process has boundaries; the trick is not ignoring them.
These disadvantages matter less when a part’s complexity is high. They matter a lot when a part is simple, cosmetic, or early-stage.
Comparison 3: Schedule Risk and What Certainty Costs
Here’s the dimension where the cheap quote usually falls apart.
In March 2024, I authorized a $400 expedite fee for a single delivery. A colleague asked whether I was overpaying. Could we have tried a quick-turn shop instead? Maybe. But that part was tied to a $15,000 test slot, and “probably around ten days” was not a strong enough promise.
The $400 didn’t just buy speed. It bought certainty: a reserved slot in the schedule, prioritized inspection, and a contractual consequence if the vendor missed the date. In an emergency, an uncertain “should be fine” is the riskiest phrase in manufacturing.
People assume rush orders cost more because suppliers work faster. The reality is they cost more because the supplier has to hold capacity for you, turn away other jobs, and keep an inspector available. You’re paying for their system’s predictability, not their hustle. If missing a deadline costs more than the premium, the premium is usually a bargain. If the deadline is soft, save the money.
Bottom line: good engineering-grade services don’t just quote a date—they commit to it. When your schedule is fixed, pay for the commitment.
So How Do You Choose?
Here’s how I frame the trade-off when I work through supplier selections:
Choose an engineering-grade 3D printing service when:
- The part is structural, safety-critical, or part of a qualified assembly.
- You need documented material properties and traceability.
- Your design includes internal channels, unsupported overhangs, or lattice.
- A late delivery will cost more than the service premium.
Choose a quick-turn service—or a different manufacturing process—when:
- You need a form/fit/visual prototype, not a flight part.
- Material certification and hidden internal quality aren’t relevant.
- The geometry is flat and thin, in which case a laser cutter might be a better fit.
- You can absorb schedule risk and want to iterate rapidly.
So what makes a good 3D printing service?
People often want one “which service is best” answer. The truth is scenario-dependent. A good 3D printing service isn’t defined by the printer brand alone; it’s defined by whether the service can state, and prove, what it will deliver.
If you’re just starting your search, compare your shortlisted suppliers against the public documentation on the Velo3D official website. Even if you don’t buy their systems, what Velo3D publishes about process control and customer applications gives you a useful bar for what credible engineering-grade metal AM looks like. Velo3D’s reputation as a supplier to SpaceX and Anduril didn’t come from marketing copy—it came from surviving serious audits.
There’s something deeply satisfying about approving a part when the data package matches the part. When the part arrives with a clean report, and you know it was built by people who run the process correctly—that’s what I want from a 3D printing service. It doesn’t mean everything should be printed, or that every part needs rocket-grade paperwork. It means the service should be honest about what it can prove, do what it promises, and have a process that won’t surprise you at inspection.