Tesla Optimus Economics: Replacing 100 Blue Collar Workers
Tesla Optimus keeps turning up in labour-cost arguments, and the argument usually runs on a single number: the robot's price. The price is the easy part. What it costs to own a fleet, power it and run it for a decade is where the interesting decisions sit.
I did the arithmetic for the most-cited version of the claim: 100 blue collar workers replaced by 100 robots. All figures are in Australian dollars.
The headline numbers
| Amount | |
|---|---|
| Annual salary replaced | $6,000,000 |
| Total capital expenditure | $4,727,664 |
| Annualised capital cost (10 years) | $472,766 |
| Annual net savings | $5,527,234 |
$6M of payroll replaced by $472,766 a year of capital. The savings come to 92% of the labour bill they replace.
Where the $4.7 million goes
| Component | Cost | Share |
|---|---|---|
| 100 Optimus bots | $3,825,000 | 81% |
| 52 Powerwall 3 units | $780,000 | 16.4% |
| Solar array | $122,664 | 2.6% |
| Total | $4,727,664 |
The robots are 81% of the capital cost. The whole energy system is 19%, and that is with the power sized at three times what the bots actually consume.
The assumptions
| Assumption | Value | Rationale |
|---|---|---|
| Annual salary per blue collar worker | $60,000 | Average from salary surveys for roles such as warehouse or plant operators, excluding additional costs like superannuation |
| Purchase price per Tesla Optimus bot | $38,250 | Projected $25,000 USD converted at 1.53 AUD/USD |
| Number of bots | 100 | Direct replacement for 100 personnel |
| Energy consumption per bot | 2.3 kWh per day | Based on Tesla specifications for a full day's operation |
| Annual operation days | 365 | Year-round operation to align with annual salary equivalence |
| Total annual energy for 100 bots | 83,950 kWh | 2.3 kWh/day/bot × 100 × 365 |
| Solar and battery oversizing factor | 3x | For redundancy |
| Annual energy generation needed from solar | 251,850 kWh | 3x the required annual energy |
| Average solar insolation | 4.5 kWh per kW per day | Typical average across Australia |
| Solar panel installed cost | $800 per kW | 2025 pricing for commercial-scale systems |
| Tesla Powerwall capacity | 13.5 kWh per unit | Usable energy per Powerwall 3 |
| Number of Powerwalls required | 52 | ceil(690 kWh daily storage ÷ 13.5 kWh) |
| Cost per Powerwall (installed) | $15,000 | Approximate 2025 installed cost in Australia |
| Total battery cost | $780,000 | 52 × $15,000 |
| System lifespan for amortisation | 10 years | Estimate for bots, batteries and solar panels |
| Exchange rate (USD to AUD) | 1.53 | Average rate in July 2025 |
Where the power math comes from
A bot uses 2.3 kWh a day, so 100 of them need 230 kWh a day and 83,950 kWh a year. Size the power system at three times that for redundancy and the array has to cover 690 kWh a day, or 251,850 kWh a year.
At Australian average insolation of 4.5 kWh per kW per day, 690 kWh a day needs about 153 kW of panels. At $800 per kW installed, that comes to $122,664.
Storage has to match the same three-times figure. 690 kWh of daily storage at 13.5 kWh usable per Powerwall 3 is 52 units, and 52 × $15,000 is $780,000.
Megapacks would cover the job and then some. One Megapack holds 3.9 MWh for about $1.59M AUD, more than five times the storage this fleet needs and roughly a third of the entire capital budget in a single unit. Powerwalls are the right tool at this scale.
What the model leaves out
None of the numbers above are wrong. The gaps are in what was left out.
Maintenance. Nothing in the ten-year amortisation covers keeping a robot running. Actuators, joints, grippers and battery packs all wear out. Whatever that costs lands on the annual side, and the annual side is what decides the case.
The humans still needed. 100 robots on a floor still need supervising, and someone has to service them. The comparison counts the saving on the replaced workforce and nothing on the other side of the ledger.
Discounting. The annualised cost is $4,727,664 divided by ten, with no discount rate. That makes the capital look more expensive than it really is, so the saving is understated, as long as the bots survive ten years of duty. Which is the maintenance question again.
Uptime. The bots run 365 days a year and are charged against a salary that also covers 365 days. The assumption gives the robots no shift changes, no breaks and no spare unit standing by while another is serviced. A human roster does not run 365 days either.
Superannuation. $60,000 excludes super, so the labour cost being replaced is understated too.
Currency. The bot price is $25,000 USD converted at 1.53. A stronger Australian dollar raises the capital cost and weakens the case.
So what is the case
Taken at face value, $472,766 a year of capital buys the labour equivalent of about eight people. That is the version of the claim that gets quoted, and it is hard to argue with on a spreadsheet.
What I would not do is write $5,527,234 into a business case as savings. It holds the robots to a 365-day duty cycle with no maintenance, no operators, no downtime and a ten-year service life, and none of those have been demonstrated at this scale. The capital cost is knowable today. The operating cost is the open question, and it is the larger half of the answer.