It Started With a $200 Plastic Part
Last year, one of our lead engineers came to me with a request: a custom bracket for a prototype assembly. He wanted it in two days. My first instinct was to send it to our usual CNC shop – they’ve been reliable for years. But the shop quoted $350 per unit with a 10-piece minimum and a two-week lead time. The engineer laughed. “Can’t we just 3D print it?” he asked.
And that’s the moment I fell into the classic procurement trap: thinking the decision is about which process, when it’s really about what you’re optimizing for.
The Surface Problem Everyone Thinks They Understand
Most people assume the trade-off is simple: 3D printing is fast but expensive per part; CNC machining is cheaper at volume but has setup costs. That’s true – as far as it goes. But it misses the deeper layer.
What I hear from engineers all the time: “3D printing is perfect for prototypes, but for production you need CNC.” And from finance: “3D printing is a waste of money unless you’re doing one-offs.” Both statements have a kernel of truth, but they’re also dangerously oversimplified.
In my experience, the real problem is that people (procurement included) default to what they know. I knew CNC, so I almost ordered a $3,500 minimum order for something that only needed one part. That’s not a cost problem – it’s a framing problem.
The Hidden Driver: Total Cost of Ownership (TCO) – And the Causal Reversal
People think 3D printing is expensive because the per-unit cost is higher than CNC at scale. Actually, the causation runs the other way: 3D printing is often cheaper for small batches because it eliminates the fixed costs of CNC (programming, fixturing, minimum order quantities). I’ve seen buyers reject 3D printing because “$50 a part is crazy,” while happily approving a $2,000 setup fee for CNC – then only ordering 20 parts, making each part effectively $150.
Here’s a real numbers example from a project I managed last fall:
- CNC turning & milling: $200 setup + $12 per part, minimum 50 parts = $800 total for 50 parts ($16/part). We only needed 5 parts – total $860 ($172/part).
- 3D printing (Markforged Onyx): $0 setup, $18 per part, order exactly 5 = $90 total. (Way cheaper – seriously, a no-brainer.)
The catch? The 3D-printed parts had lower surface finish and couldn’t hold ±0.001” tolerances. For the prototype, it was fine – but if we had needed tight tolerances, CNC would have been mandatory regardless of cost.
The Cost of Getting It Wrong
Last year, I approved a rush 3D printing order for a part that needed to withstand high torque. The vendor (using a budget printer, not a Markforged) delivered in 48 hours. Part looked great. But during testing, it cracked under load. We lost a week of testing and ended up ordering CNC-machined replacements – at triple the cost because of the expedited machining. That week of delay cost us more in engineering hours than the part itself.
I should have asked: “Is this part going to be load-bearing? Does it need to survive 10,000 cycles?” If yes, then 3D printing (even industrial-grade) might not be the right play – or at least you need the right material and printer. Markforged’s carbon-fiber reinforced Onyx or Metal X could handle it, but a hobby-grade machine? Big red flag.
“I have mixed feelings about rush decisions. On one hand, being fast makes engineers happy. On the other, fast without context creates rework. I’ve learned to ask two extra questions before choosing: ‘What’s the tolerance?’ and ‘What’s the use case?’”
So When Does 3D Printing (Markforged) Actually Shine?
Let’s be honest: I’m not here to sell you on 3D printing for everything. The bottom line is this:
- 3D printing wins when you need complex geometry, low volume (1–100 parts), fast iteration, or consolidation of assemblies into one part. Markforged printers, with their continuous carbon fiber and metal capabilities, push the envelope further into end-use parts.
- CNC turning & milling wins when you need tight tolerances (below ±0.005”), high volume (hundreds or thousands), superior surface finish, or specific materials that aren’t available for 3D printing (e.g., high-temperature alloys).
- 6kW laser cutting is a different beast – it’s ideal for sheet metal, quick flat patterns, and prototypes in metal when the design is 2D. If your part is a bracket that can be cut from a flat sheet, laser is cheaper and faster than 3D printing or CNC.
Here’s where Markforged fits in my workflow today: I use their Onyx/FX series for functional prototypes, jigs, fixtures, and low-volume production parts that don’t need sub-micron tolerances. Their Metal X is my go-to when I need metal parts but can’t wait 2–3 weeks for CNC – and the part geometry is complex enough that machining would require multiple setups. But if I need a simple shaft or threaded part? I still go CNC. (Because seriously, 3D-printed threads are still a pain – not a deal-breaker, but a limitation.)
What I Wish I Knew From Day One
If you’re in procurement or manufacturing engineering and you’re on the fence between 3D printing and traditional machining, here’s my advice (from someone who’s been burned both ways):
- Don’t make the decision based on per-part cost alone. Calculate total cost including setup, minimums, lead time, and scrap risk.
- Ask yourself: is the part tolerance-critical or geometry-complex? Tighter tolerance → CNC. Complex geometry → 3D print.
- Use 3D printing for stages where speed matters most: prototyping, tooling, bridging production while CNC tooling is being made.
- Be honest about when 3D printing doesn’t fit. That’s the real power – knowing when to say “this one should go to CNC” makes your 3D printing recommendations more credible.
I still have mixed feelings about Markforged. I love the capability, but I’ve seen engineers try to use it for everything – and that’s a recipe for disappointment. The best approach is a hybrid model: keep a reliable CNC partner for precision work, and use Markforged for the stuff that makes traditional shops groan (complex undercuts, lightweight lattices, quick turnarounds).
“The assumption is that 3D printing replaces machining. The reality is that both have unique strengths. The smart buyer knows the difference – and sticks to the right tool for the job.”
At the end of the day, my job isn’t to pick a technology. It’s to get the right part to the engineer at the right time and cost. And that sometimes means saying: “This part? Use the 6kW laser cutter. That part? Go CNC. And that one? Markforged it.”
No single process wins every time. And once you accept that, choosing becomes a whole lot easier (and way less stressful).