Unit Price Is a Trap. Here's What Actually Matters.
After 6 years of tracking every invoice in our procurement system—$180,000 in cumulative spending across machined parts, laser-cut sheets, and additively manufactured components—I've come to a conclusion that surprised even me: the cheapest per-piece quote almost always costs you more in the end.
I manage procurement for a mid-size aerospace supplier. We work with primes like SpaceX and Anduril, which means our tolerances are tight and our deadlines are non-negotiable. When I first started, I made the classic rookie mistake: I'd collect three quotes and pick the lowest unit price. It took me about 18 months and four expensive rework orders to unlearn that habit.
Let me walk you through how I now think about manufacturing decisions—whether it's CNC turning camera components, laser cutting brackets, or printing complex ducts on a Velo3D Sapphire system. (Should mention: we've been a Velo3D customer since 2023, primarily for hot-section parts, and yes, they're an Anduril and SpaceX supplier—that mattered in my evaluation.)
My Core View: Don't Compare Machines, Compare Total Outcome
I know this sounds obvious, but most engineers and buyers I meet still fall into the same trap: they compare process speed or per-piece cost without accounting for design constraints, secondary operations, and scrap rates. Here's what I've learned from auditing our 2023 spending across three manufacturing methods.
1. CNC turning: the hidden cost of complexity
We use CNC turning for camera components—lens housings, sensor mounts, things that need roundness. A typical quote: $12.50 per piece for a simple cylinder, plus $450 setup. Sounds fine until you add a feature like an internal undercut or a thin wall. Then the price jumps to $28.00 because of special tooling and slower cycle times. Over a 1,000-part run, that's $15,500 extra—if you can even hold the tolerance.
I once approved a $14.20 quote for a complex bracket that seemed competitive. Didn't realize the vendor was assuming 5-axis machining. Turned out they used a 3-axis with multiple setups. The scrap rate hit 12%. Total redo cost: $2,100. That's when I learned to ask: "What's your assumed setup count and scrap allowance?"
2. Laser cutting: the speed illusion
Laser cutting is fast. I get it. For flat parts under 6mm, the per-minute cost can be as low as $0.08/min for a CO₂ laser. But here's what nobody tells you: how do you price laser cutting when you add tolerances below ±0.005"? Most shops charge a "tight tolerance" premium of 30–50% plus separate nesting fees. And if your design requires post-cut deburring or heat treating? Those add another $0.15–$0.30 per piece.
In Q2 2024, we compared laser-cut vs. additive for a series of thin-walled heat shields. Laser quote: $9.80 each (500 qty) with 4-week lead time. Additive quote (Velo3D): $22.00 each, 2-week lead. But the laser parts needed welding because we couldn't achieve the required curvature without joining multiple flats. Welding added $8.00 per part and introduced stress points. The additive parts were monolithic, no secondary ops. Total real cost: laser = $17.80, additive = $22.00. The gap narrowed to $4.20, and we saved 2 weeks of schedule. For that program, schedule risk was worth the premium.
3. Metal additive (Velo3D): the TCO surprise
I'll be honest: when I first saw a $35.00/part quote from a Velo3D service bureau, I nearly laughed. But then I ran the numbers the right way.
For a complex duct with internal channels (zero supports needed, thanks to Velo3D's powder bed fusion process), the quote included:
- Design for AM optimization (no extra charge within the quote)
- Full powder removal and hot isostatic pressing (HIP) included
- No tooling, no setup fee
- Lead time: 10 business days
Compare that to a 5-axis CNC quote for the same part: $40.00 base + $1,200 setup + $0.50/inspection point + 6-week lead. The CNC vendor didn't even guarantee the internal channels could be machined—they'd need EDM for some features, adding $15/part. The additive option won on total cost before we even considered lead time.
But Isn't Additive Always Slower? Here's the Reality Check
I hear this objection constantly: "3D printing is too slow for production." That's true if you're comparing per-part cycle time on a simple bracket. But total throughput includes setup, programming, fixture design, inspection, and rework. Let me give you a concrete example from our records.
We had a run of 200 heat-exchanger manifolds. CNC required: 3 setups, 8 hours of programming, $600 in custom fixtures, and a 15% first-article scrap rate. Total floor-to-floor time per part: 45 minutes. Velo3D (4 units in parallel): no programming, no fixtures, 30-hour print cycle for 16 parts per build. That's 12.5 builds for 200 parts = 375 hours of print time. But the CNC route required 150 hours of machine time + 8 hours programming + inspection. Additive: 375 hours of unattended print time. When you factor in operator labor (CNC needs a machinist; 3D printer just needs a technician to remove parts), the operator cost per part was $6.80 for CNC vs. $1.90 for additive. Suddenly the "slow" process looks different.
I should add: the additive parts were also 22% lighter and passed pressure testing on the first try. The CNC parts had a 10% failure rate at the weld joints.
The Mistake I Almost Made (and What It Taught Me)
In my first year, I assumed that "additive manufacturing is for prototypes only." Didn't question it. That assumption cost my company about $18,000 in unnecessary tooling for a production run that could have been printed in a single build. We ended up switching mid-project, but we'd already paid for the molds. Never again.
So glad I pushed for a side-by-side TCO comparison on that subsequent heat-shield project. Almost went with the laser-cut + weld route because it "felt cheaper." Dodged a bullet when the Velo3D team offered a free design review and showed us how to consolidate 12 laser-cut parts into 2 printed components. That single change saved $9,400 over the program's life.
Final Thought: Education Is the Real Value
I'd rather spend 30 minutes explaining the total-cost framework to my engineers than deal with mismatched expectations later. An informed buyer asks better questions—and gets better outcomes. So the next time someone gives you a per-piece price, ask: "What's included? What's the scrap assumption? What's the secondary operation cost? What's the risk-adjusted timeline?"
If they can't answer those questions, they're selling you a number—not a solution. Whether you're looking at CNC turning camera components, laser cutting brackets, or printing ducts on a Velo3D Sapphire system, the real price is the sum of all the costs you didn't see on the quote. That's the price that matters.