Cost-aware BOM engineering

Robotics is hardware. Hardware has a Bill of Materials. The BOM determines whether your robot can ship at $200 or $20000 — and whether your business model works. Cost-aware BOM engineering is unglamorous but it's the difference between an expensive demo and a real product. Here's the working knowledge that separates engineers who build cheap robots from those who quote 4× more than they need to.

The 80/20 rule

For most robots, 5–10 components account for 80% of the BOM cost. Spending engineering effort on cheaper alternatives for those few items lifts the whole project's affordability. The rest of the BOM (resistors, screws, brackets) doesn't matter — it's a few percent total.

The big-ticket items in 2026 hobby robots:

• Compute: Jetson Orin Nano $499; Pi 5 $80. ~10× delta if you can use Pi.
• Cameras: RealSense $300; phone-based ToF $5 in volume. 60× delta.
• BLDC motors + drivers: ODrive $200/axis; SimpleFOC custom $30/axis.
• Battery: $30–$200 depending on capacity and chemistry.
• Lidar: $200 (single-line) to $3000 (multi-line); 15× delta.
• Mechanical structure: $50 in 3D-printed PLA vs $500 in machined aluminum.

Decide what budget tier you're targeting; pick from the right column.

The sourcing landscape

Where you buy from matters as much as what you buy:

Source	Price	Lead time
Amazon	~3× cost	2 days
DigiKey / Mouser	1.2× cost	2–3 days
AliExpress	1× cost	2–4 weeks
Alibaba (volume)	~0.7× cost	4–8 weeks + customs
LCSC	1.1× cost	5–10 days

For prototypes: DigiKey + LCSC. For production: Alibaba + AliExpress. For "I need this on Tuesday": Amazon, but expect to pay.

The pricing tricks

Wholesalers hide pricing tiers. Always:

• Click "see volume pricing" — many components drop 30%+ at 100 units.
• Use parametric search to find equivalents — a "STM32F407VGT6" might cost twice as much as a "STM32F407VET6" with one fewer pin.
• Check distributor stock before committing — if 1 distributor has 10000 in stock and others zero, that's pricing leverage.
• Ask for samples for prototype builds; manufacturers will often send free.
• Look at industrial-grade vs commercial-grade — sometimes industrial is cheaper if it's the production-volume part.

The cost-engineering process

For a new robot project:

1. Write the BOM with placeholder line items.
2. For each item, estimate cost from previous projects or quick web searches.
3. Sort by cost; the top 5 are your optimization targets.
4. Spend a half-day on each top item: alternatives, suppliers, volume pricing.
5. Update the BOM; total it.
6. If still too expensive: scope reduction (which features can go?) before more sourcing.

Discipline: never skip the top-5 review. It's where 90% of cost reduction lives.

The volume curve

Component cost vs quantity is non-linear. A typical chip:

• 1 unit: $10.
• 100 units: $4.
• 1000 units: $1.50.
• 10000 units: $0.60.

If your business plan is 1000 robots/year, your unit cost won't be the prototype cost. Estimate the production-volume cost when planning.

Custom vs off-the-shelf

Off-the-shelf at low volume; custom at high volume.

The crossover usually sits at 100–500 units, depending on the component. Custom PCBs at 10 units cost $50 each (PCBA included); at 1000 units, they're $5. Custom mechanical brackets are similar.

For prototype + small production: ~3D printed structure + off-the-shelf electronics. For volume: PCBA + machined / injection-molded structure.

The depreciated cost trap

"This robot costs $5000 to build, but with $200 in materials" — said by every consumer-robot startup. The $5000 includes labor, overhead, design amortization, support, returns, etc.

For pricing strategy:

• BOM cost: just parts.
• Cost of goods sold (COGS): BOM + manufacturing labor + assembly + QC + shipping.
• Fully-loaded cost: COGS + design amortization + warranty + support.
• List price: typically 2.5–4× fully-loaded cost.

A $200 BOM robot has a $400–$700 COGS and a $1500–$3000 retail price. The "BOM cost" headline is misleading.

Make-vs-buy decisions

For each component, ask: should I design this myself or buy it?

• Buy: solved problems with mature suppliers (sensors, motors, batteries, connectors).
• Make: differentiated parts (custom geometry, novel actuators, specialized PCBs).
• Modify: 3D-printed enclosures, custom firmware, branded plastics.

The make-vs-buy decision drives engineering time. Custom mechanical part = 2–4 weeks; off-the-shelf = an hour.

The IP angle

If your robot has unique IP, protect it:

• Custom PCBs in your name.
• Trademark / brand on the housing.
• Patent the differentiating mechanism (if it's actually novel).

For most hobbyist projects, IP doesn't matter. For a startup, it's part of the BOM exercise.

A worked example: $200 quadruped

For a tabletop walking robot:

• 12× MG996R servos: $40 (volume from AliExpress).
• Servo controller (PCA9685): $5.
• ESP32-S3: $7.
• 3000 mAh 2S LiPo: $20.
• BMS + battery management: $10.
• 3D-printed structure: $5 (PLA).
• Power regulator: $5.
• USB cable + connectors: $8.
• IMU (MPU6050): $4.
• Misc (screws, jumpers): $10.
• Total BOM: ~$115.

At $115 BOM, ship for ~$300 retail. For a desk-sized walking robot, this is achievable in 2026.

Common cost mistakes

• Over-spec'd components: a $500 industrial servo where a $40 hobby servo would do.
• Branded over generic: Adafruit modules cost 3× the equivalent AliExpress part.
• Single-source on critical parts: supplier discontinues; you're stuck.
• Ignoring shipping and import duty: a $20 part can land at $35 after international shipping.
• Forgetting the assembly time: at $30/hr labor, 2 hours of assembly = $60 added to COGS.

Exercise

Pick a robot project. Write the BOM in a spreadsheet. Sort by cost. Spend an hour on the top 3 items: find at least one alternative for each. Update the spreadsheet; total. The reduction will surprise you. This same exercise scales from $100 hobby projects to $50000 industrial robots.

Next

Multi-robot coordination — when one robot isn't enough.