Cost Modelling for Additive Manufacturing: A Practical Framework

The most expensive mistake in AM procurement is applying a simple $/kg or $/hour rate to a quote and calling it a cost model. AM cost is non-linear in ways that confuse every first-time buyer and many experienced engineers: a part that fits 50 to a build plate costs 1/50th the machine time of running it alone; a support-heavy orientation can double material consumption; a post-processing step costing $15/kg on a 2 kg part is irrelevant, but on 200 kg of parts it is your second-largest line item. This guide gives you a structured five-driver cost model with real numbers and a worked example you can adapt to your situation.

1. Why Naive Cost Models Fail

The Volume-Fill Interaction

AM cost depends on how much of the build volume is filled. A 200 × 200 × 300 mm build volume filled with one large part and a 200 × 200 × 300 mm build volume filled with 80 small brackets have identical machine time but dramatically different per-part costs. This means that part count per build, not part complexity alone, drives per-part machine cost.

The implication: when quoting or evaluating AM costs, the number that matters is cost per cubic centimetre of part volume — but only after optimising build packing. Quoting based on a half-empty build plate overstates cost by 2–5×.

Non-Linear Support Volume

Support structures in LPBF and DED are printed material that is subsequently machined or manually removed. They consume:

• Machine time (laser scanning the support geometry)
• Material (which is wasted or recycled at degraded quality)
• Labour (removal, often by hand with chisels, pliers, and abrasive tools)
• Post-machining (supported surfaces need machining after removal)

A poorly oriented part with 60% of its volume as support structure costs nearly twice as much as the same part oriented to 10% support. The support volume estimator and orientation advisor on additive.tools directly address this cost driver before any material is spent.

Machine Utilisation

AM machine cost is dominated by CapEx amortisation — typically 50–70% of the hourly machine rate. A machine running at 30% utilisation costs 3× more per part than the same machine at 90% utilisation. Utilisation rates in job shops are typically 40–65%; in high-volume production facilities, 70–85%. This is why in-house AM rarely makes economic sense for fewer than ~300–500 builds per year: at lower volumes, the fixed cost of the machine cannot be amortised.

2. The Five Cost Drivers

(a) Machine Cost per Hour

The hourly machine rate is the single largest cost driver for small, complex parts with moderate material volume.

LPBF machine CapEx:

• Entry-level (single laser, 250×250 mm): $0.3–0.6M
• Mid-range (single/dual laser, 400×400 mm): $0.6–1.0M
• Production (multi-laser, 400×400 mm+): $1.0–2.5M
• EBM (Arcam Q20plus): $0.7–1.5M
• DED (large format): $0.5–1.5M

From CapEx to hourly rate (LPBF mid-range example):

Add operator labour ($30–60/h depending on region and skill level) and overhead (typically 150–200% markup on direct labour in a job shop), and the all-in machine + labour + overhead rate for a mid-range LPBF system is typically $80–150/h. High-end multi-laser systems or low-utilisation single-machine shops can reach $200–300/h.

(b) Material Cost

Metal powder for AM commands a significant premium over bulk material:

Utilisation rate (buy-to-fly ratio): In a typical LPBF build, the part volume is 30–80% of the bounding box volume, and support structures add 10–60% of part volume in additional consumed material. The effective buy-to-fly ratio (purchased powder/final part weight) is typically 1.5:1 to 4:1 for LPBF, versus 5:1 to 30:1 for CNC machining of billet.

Recycling: Unfused LPBF powder is sieved and recycled, but each cycle degrades particle morphology and can increase oxygen content. Most specifications allow 3–5 recycles before powder retirement. Effective utilisation improvement from recycling: 15–30% reduction in net powder cost per part.

(c) Labour and Setup

• Machine setup (powder change, build preparation, file preparation): 1–4 h per build; $50–200
• Build monitoring: typically automated for production runs; 1–2 h operator time for manual monitoring
• Post-build depowdering: 0.5–3 h depending on part complexity; $25–150
• Support removal: highly variable; $50–500+ per part for complex support structures; the most cost-variable step
• Inspection (dimensional, visual): 0.5–2 h per batch; $25–100

Setup cost amortises across all parts in a build. Running 50 parts per build at $200 total setup adds $4/part. Running 1 part adds $200/part. Batch optimisation is the highest-leverage cost reduction lever after support minimisation.

(d) Post-Processing

Post-processing cost is systematically underestimated in early-stage AM cost models. Common post-processing steps and indicative costs:

For a medical Ti-6Al-4V implant, the post-processing cost stack (HIP + machining + electropolish + CT + CMM) often exceeds the printing cost by 2–3×. This is not exceptional — it is normal for regulated industries.

(e) Overhead and Amortisation

Overhead includes quality management (ISO 9001/AS9100 certification maintenance: $30,000–80,000/yr for a small shop), sales and quoting, engineering support, and facility overhead not captured in the machine rate. Standard job shop overhead multiplier: 150–250% on direct labour. For purpose-built AM production facilities with high automation, overhead can drop to 100–150%.

3. Build Time Model

Build time drives machine cost. For LPBF:

Build time = (Layer count × Layer time) + Recoat time + Setup/cooldown

Where:

• Layer count = Part height / Layer thickness (e.g., 50 mm / 0.04 mm = 1,250 layers)
• Layer time per cross-sectional area = area / (scan speed × hatch spacing) — for a 10 cm² cross-section at 800 mm/s and 0.1 mm hatch: 10,000 mm² / (800 × 0.1) = 125 s/layer
• Recoat time = 5–15 seconds/layer (machine-dependent)
• Total for 1,250 layers: (125 + 10) s × 1,250 = 46.9 h

This model shows why tall parts cost proportionally more: each additional 1 mm of height adds ~135 seconds of build time regardless of cross-section area (just the recoat and minimum scan overhead). The build time estimator on additive.tools implements this model with machine-specific parameters.

Multi-laser scaling: A quad-laser machine running four lasers simultaneously on the same cross-section does not simply divide build time by four. Lasers must avoid one another's melt pools, and edge regions are typically handled by a single laser. Effective speedup is 2.5–3.5× for typical cross-sections. For very small cross-sections (< 200 mm²), laser interference limits reduce the speedup to < 2×.

4. Break-Even Analysis: AM vs CNC vs Casting

The AM break-even against conventional manufacturing depends on three variables: part complexity (geometric freedom benefit), volume (amortisation of tooling/setup), and lead time tolerance.

General break-even rules of thumb:

Use the make vs buy tool on additive.tools to compute break-even for specific part geometries with your cost inputs.

5. The Support Cost Trap

Support structures are the hidden cost multiplier that transforms an apparently reasonable per-part estimate into an unprofitable job. Consider a simple housing with internal channels printed upright: support volume might be 15% of part volume. Rotate it 90°: support volume rises to 55%. The material cost doubles. The removal labour doubles. The machining of supported surfaces doubles.

Quantified support cost impact for a representative LPBF 316L part (~200 cm³ bounding box, 80 cm³ solid volume):

At $80/h labour + machining, poor orientation costs $295 more per part than optimised orientation. At 100 parts, this is $29,500 — enough to justify a dedicated design-for-AM review. The orientation advisor and support volume estimator on additive.tools quantify this before any material is spent.

6. Post-Processing Cost Model

For a complete cost model, post-processing must be itemised, not estimated as a percentage of print cost. The percentage heuristic (e.g., "add 30% for post-processing") fails when post-processing costs are dominated by minimum charges:

• An HIP run with a minimum charge of $1,000 applied to a single $200 bracket makes the part $1,200 — 6× more than print cost. Applied to 100 identical brackets in the same HIP run, it adds only $10/part.
• A CMM first-article inspection at $300/part changes economics entirely for low-volume aerospace parts.

Build a line-item post-processing cost table for every job:

The post-processing time estimator on additive.tools helps model the schedule and cost of multi-step post-processing workflows.

7. Worked Example: 100 × Ti-6Al-4V LPBF Brackets

Part: Aerospace bracket, 80 × 60 × 40 mm bounding box, 45 cm³ solid volume, 15 cm³ support volume (well-optimised orientation), 60 g solid mass (Ti-6Al-4V density 4.43 g/cm³ → 45 cm³ × 4.43 = 199 g ≈ 200 g/part). 24 parts per build (4 × 6 grid), so 5 builds for 100 parts (with some buffer builds assumed → 5 builds).

Build Parameters

• Machine: mid-range LPBF, $100/h all-in rate
• Layer thickness: 60 µm
• Part height: 40 mm → 667 layers
• Average cross-section per build: 24 × 45 cm² / 40 mm depth × density factor ≈ build time ~18 h/build
• Total machine time: 5 builds × 18 h = 90 h → $9,000 machine cost

Material

• Part volume: 100 × 45 cm³ = 4,500 cm³ = ~20 kg solid Ti-6Al-4V
• Support volume: 100 × 15 cm³ = 1,500 cm³ = ~6.6 kg support (recycled but degraded)
• Net powder consumption (accounting for 80% recycle): ~24 kg equivalent purchase
• Powder at $200/kg: $4,800 material cost

Setup and Labour (print side)

• 5 builds × $150 setup/build: $750
• Depowdering, 5 builds × 2h × $60/h: $600
• Print-side labour: $1,350

Post-Processing (per 100 parts)

• Stress relief (batch, 20 kg parts @ $3/kg): $60
• HIP (100 parts, $1,000 minimum run + volume): $1,200
• CNC machining (3 surfaces, 0.5 h/part × 100 × $120/h): $6,000
• Inspection CMM (10% sample, 10 parts × $250): $2,500
• Post-processing: $9,760

Overhead and Margin

• Engineering/quoting overhead (20% of direct): ($9,000 + $4,800 + $1,350) × 0.20 = $3,030
• Profit margin (25% on total): markup applied below

Summary

At higher volumes (500+ parts), the 24 parts/build density improves, HIP minimum charge amortises further, and machine utilisation increases; the landed price typically falls to $220–280/part for this geometry. At 10 parts (1 partial build), it rises to $800–1,200/part due to setup and HIP minimum charge dominance.

This worked example matches real-world quotes from LPBF service bureaux for comparable Ti-6Al-4V aerospace structural parts in 2024–2025.

8. Reference: additive.tools Calculators

The worked example above can be reproduced and adapted using the following tools:

• Cost per part — basic per-part cost from machine rate, material, and build time
• Total cost of ownership (TCO) — full CapEx + OpEx model for in-house vs outsourced AM
• Make vs Buy — break-even analysis against CNC, casting, and AM alternatives
• Build time estimator — layer-by-layer build time from part geometry and machine parameters
• Support volume estimator — estimate support mass and cost penalty by orientation
• Post-processing time estimator — model multi-step post-processing schedule and cost
• Multi-machine cost — compare per-part cost across different machine options for a given volume

Further Reading

• AM for automotive: scale economics — high-volume AM cost structures in automotive
• Conformal cooling design guide — cost-benefit analysis of conformal cooling vs conventional
• Topology optimisation for AM — weight reduction that also reduces material cost and build time
• Orientation advisor — orient parts to minimise support volume and surface roughness penalty