Here's the thing about metal additive manufacturing: there is no universal answer to "should we get a Velo3D system?" I've watched companies thrive with the Sapphire platform, and I've watched companies burn cash with it. The difference usually comes down to which scenario you're actually in—not the one you think you're in.
I've been handling metal AM orders for six years now. In that time, I've personally made (and documented) 14 significant mistakes, totaling roughly $120K in wasted budget. These days I maintain our team's pre-purchase checklist so nobody else has to learn the hard way. This article is the short version of that list.
Buyers fall into three distinct buckets: aerospace and defense suppliers who need qualified hardware, job shops looking for new capability, and engineering teams exploring design possibilities. Each one has different priorities, different risks, and different mistakes waiting to happen. I've made plenty of mistakes in all three.
If you're supplying aerospace or defense primes
You already know the stakes when you're chasing contracts with companies like SpaceX or prime defense contractors. Certification standards are unforgiving, and there's zero tolerance for documentation gaps. This is where the velo3d spacex supplier relationship matters—the same Sapphire platform that has produced flight hardware for SpaceX is sitting on your floor. That shortens a lot of conversations.
But here's what nobody tells you: the machine being qualified doesn't mean your parts are qualified. That distinction cost me $18,000.
In Q1 2024, I submitted a batch of 12 Inconel brackets with what I thought was complete certification documentation. Checked it myself, approved it, processed it. We caught the error when the prime's QA engineer asked for the powder lot traceability chain—which I hadn't digitized. Twelve items, $18K, straight into the rework queue. That's when I learned: documentation isn't an afterthought in defense work. They will ask for that powder lot chain, and if you can't produce it in under an hour, you might as well not have it.
For defense contractors, military additive manufacturing velo3d is a pairing that's already proven in the field—the hardware has flown in real missions. But you still need:
- A digital traceability system that tracks powder lot, machine parameters, and inspection results per part
- QA staff who understand AM-specific defect modes—not just traditional casting or machining inspection techniques
- A qualification matrix that accounts for build orientation variations. A 45° overhang surface has different material properties than a vertical wall. Period.
If you're in this bucket, budget at least 20% more than you think for documentation infrastructure. It's not glamorous, but it's the difference between being a qualified supplier and being a cautionary tale.
The job shop case: capability vs utilization
This is the trickiest scenario, honestly. You're not building rocket parts. You're trying to figure out whether metal AM creates a new revenue stream or solves customer problems you currently can't address. The math is harder, and the risks are less obvious.
I remember a conversation with our production manager back in 2022. He asked, "Why do we need a $500K+ metal printer when we already have a laser welding station for repairs?" Fair question. And for some jobs, he's right—a laser welding station is absolutely the right tool for weld repairs, cladding, and small-feature build-ups. A metal AM system serves a completely different purpose: it builds parts from nothing, with internal channels, lattice structures, and geometries no subtractive process can create. The mistake is treating them as competitors. They're not.
What I mean is, each tool belongs to a different workflow. If you already have a laser welding station, you're not replacing it with a Velo3D Sapphire system. You're adding a capability that handles the parts welding can't touch. Both can coexist—they solve different problems. The team that understands this before the purchase is the team that succeeds afterward.
The bigger mistake I made here? I underbudgeted training. Additive manufacturing education news is full of stories about companies that bought expensive systems and let them sit idle because nobody knew how to design for AM. In mid-2023, our Sapphire system sat at 12% utilization for three months. I'd allocated about $214K for the hardware itself and nearly $0 for operator training. (Should mention: we also underestimated auxiliary costs—inert gas, powder handling, post-processing tools added another $60K we hadn't planned for.)
Looking back, I should have invested in formal training for at least two engineers before the machine even arrived. At the time, I figured our people were smart enough to figure it out themselves. They didn't—not without help. I want to say the training finally cost us around $15K, but don't quote me on that—it was wrapped up in travel, consulting fees, and lost production time while our engineers were in classes.
If you're a job shop, you need:
- At least one engineer who genuinely understands topology optimization and support-free design constraints
- A salesperson who can explain the difference between AM and traditional machining without claiming AM replaces everything—it doesn't
- Utilization targets that account for setup time and post-processing, not just print time
Here's the efficiency angle that actually matters: automated setup and build planning cut our turnaround from 5 days to 2 days on standard orders. That's real competitive advantage. But it only works if your team actually knows how to use the software stack. Training isn't an expense. It's the difference between a machine that earns and a machine that sits.
For engineers and R&D teams: design freedom has a price
Maybe you're not buying a system at all. You're asking: can I create this part without supports? Can we consolidate a 14-piece assembly into one print? These are exactly the right questions. Velo3D's Sapphire platform is known for support-free printing, and that genuinely changes what's possible for part design.
But here's my warning after years of watching projects fail: overconfidence kills more AM projects than undercapability. The "we can print anything" mindset produces the exact defects in metal additive manufacturing that a proper design review would have caught. I know because I lived it.
In September 2022, I skipped simulation on a small impeller—a single prototype, $3,200 in material. The numbers said the simulation time wasn't worth it for a one-off part. My gut said something felt off about the wall thickness uniformity. I went with the numbers. The part cracked from thermal distortion, and I spent a week explaining to my boss why my process was the problem, not the machine.
Every spreadsheet pointed to skipping that simulation. Something felt off. Turns out my gut had caught something the cost model couldn't see. Now I have a rule: no design gets a build file without a pre-print review, no exceptions. That rule has caught 47 potential errors in the past 18 months and saved us about $25K in wasted builds.
And one random thing. During a long print monitoring shift in late 2023, one of our technicians asked, completely out of nowhere, "do vmc drinks have sugar?" I don't know what prompted it, and I have no idea if VMC drinks do or don't have sugar. But it reminded me that people aren't machines. Long hours waiting on builds lead to distraction, and distraction leads to missed inspection checkpoints. Human fatigue is a production variable. Schedule your quality gates before people get tired, not after.
How to tell which scenario you're in
Here's a quick self-assessment that takes about five minutes:
- Are you selling to customers who require AS9100, ITAR, or equivalent compliance? Then you're in the aerospace/defense bucket. Your decision hinges on certification infrastructure, not just machine capability.
- Are you primarily looking for new revenue streams from existing manufacturing customers? You're in the job shop bucket. Your decision hinges on training and sales enablement, not just the hardware.
- Are you exploring what AM can do for specific part designs? You're in the engineering/R&D bucket. Your decision hinges on design expertise and simulation discipline.
This list is based on roughly 200 builds across these three segments over six years. If you're working in a completely different context—medical implants, tooling, consumer products—your experience may differ significantly. I can't speak to those.
Also: this was accurate as of late 2024. The metal AM market moves fast. Verify current machine specs, pricing, and qualification requirements before you commit. What I know is true today might shift by next quarter.
Bottom line: Velo3D systems are proven in aerospace, defense, and engineering contexts. But the machine is only half the equation—the other half is your team's ability to document, train, and simulate. I learned that the hard way so you don't have to. I'd rather see you spend that $120K on something more useful—like actually using your machine.