Tesla Optimus price target: $20,000. Is that real?

Elon Musk said Tesla Optimus will cost $20,000.

I’ve heard enough Musk predictions to apply the standard correction factor (multiply the timeline by 2-3, take the specifics with salt). But the number itself is interesting enough that I wanted to do my own estimate.

The bill of materials estimate

I went through publicly available information about humanoid robot components and estimated costs at volume production (10,000+ units per year).

Actuators and motors: 28-40 actuators (depends on hand dexterity). At scale, high-quality brushless DC motors with harmonic drives run about $50-150 each. Call it $100 average across 35 actuators. $3,500.

Sensors: Cameras (8-12 units), IMU, force sensors in hands and feet, LIDAR or depth sensors. Camera modules at volume are cheap ($5-15 each). The overall sensor package, including processing boards. $400-600.

Compute: Tesla uses their own FSD chip. At volume, including memory and power management. $300-500.

Battery: Estimated 2-3 kWh for 8+ hours of operation. Using Tesla’s own cells at their internal cost. $400-600.

Frame and shell: Aluminum and plastic structural components. Machined at scale. $800-1,200.

Wiring, connectors, PCBs: The nervous system of the robot. $300-500.

Total hardware BOM at scale: roughly $5,700-6,900.

The gap to $20,000

A BOM of $6-7K suggests a retail price of $20,000 is in the range of possibility with Tesla’s manufacturing scale. The typical hardware markup to cover assembly, software development, warranty, logistics, and margin is 2-3x BOM.

At 2x: $12,000-14,000. Aggressive but Tesla-like. At 3x: $18,000-21,000. More typical for complex hardware.

So $20,000 isn’t fantasy. But it requires things that don’t exist yet: a mature supply chain for humanoid robot components, Tesla-scale manufacturing (hundreds of thousands of units), and software capable enough that the robot is worth $20,000 to a consumer.

That last part is the real constraint. A humanoid robot at $20,000 that can only fold laundry slowly isn’t worth $20,000. The hardware could hit the price point. The software determines the value.

My actual prediction

First consumer version: $40,000-50,000. Limited capability. Basic household tasks. 2027-2028 if Tesla is aggressive, 2029 if realistic.

Second generation: $25,000-30,000. Better software, more capable hands, learned a wider range of tasks.

Third generation: maybe $20,000. If the manufacturing learning curve follows Tesla’s automotive curve.

The $20,000 robot isn’t a first-generation product. It’s a third-generation product sold at first-generation volume. That’s how Tesla priced the Model 3 after the Roadster and Model S.

Is it real? Not yet. But the physics of the BOM and the pattern of Tesla’s pricing strategy make it more plausible than I expected. And plausible is farther than I thought I’d get when I started the spreadsheet.

Related thinking: