Humanoid Robot Factories Now Build One Per Hour — Here Are The Production Numbers
Rapid Advancements in Humanoid Robot Manufacturing
Increased Production Capacity
- A factory in California has increased its humanoid robot production from one per day to one per hour, achieving a 24-fold increase in throughput within 120 days.
- Figure AI's plant has delivered over 350 third-generation humanoid robots using more than 150 connected workstations and custom manufacturing software.
- The annual capacity of this facility is now at 12,000 units.
Competitive Landscape
- Engine AI, a Chinese startup, opened a factory capable of producing a complete humanoid robot every 15 minutes, targeting an annual output of 10,000 units.
- Other companies like Agility Robotics and Unitree are also ramping up production with similar targets for their respective models.
- Tesla is set to enhance its Optimus Gen 3 production this summer while Boston Dynamics is shipping its Atlas model to Hyundai.
Cost Reductions and Economic Implications
Decreasing Manufacturing Costs
- Goldman Sachs reported that the costs associated with humanoid manufacturing have dropped by 40% within a year, ranging from $50,000-$250,000 down to $30,000-$150,000.
- The cheapest models are already available for under $14,000.
Labor Cost Comparison
- The average cost of employing a US manufacturing worker (including wages and benefits) is approximately $156,000 annually.
- Companies utilizing humanoid robots report payback periods of less than two months due to lower operational costs compared to human labor.
Understanding Takt Time in Production
Definition and Importance
- Takt time refers to the interval between completed units on a production line; it’s crucial for measuring efficiency.
- Figure AI's takt time improved from 24 hours to just 60 minutes; Engine AI's takt time stands at an impressive 15 minutes.
Industry Comparisons
- For context, Tesla produces a Model Y approximately every 40 seconds. This highlights the rapid advancements being made in the humanoid robotics sector.
Quality Control Challenges
First Pass Yield Metrics
- Figure AI reports an initial first pass yield of only 80%, meaning one out of five robots requires rework before shipment—this is below automotive industry standards but acceptable for early-stage production.
Verification Concerns
- The sustained output claimed by Engine AI at their new factory has yet to be independently verified.
Integration vs. Production Volume
Distinction Between Building and Deploying Robots
- While factories can produce robots efficiently, integrating them into existing workflows involves additional challenges such as safety validation and reliability across multiple shifts.
Shift in Focus
- The industry has transitioned from merely building functional robots to optimizing production speed so that hiring human workers becomes optional due to economic advantages.