Unique Research · Insight · October 30, 2024 · Source: Sam Korus / ARK Invest (CN editorial; “APK” in CN = ARK)
Elon Musk (Tesla Q1 call): a humanoid that perceives reality and executes on demand removes soft caps on economic scale. ARK: even before superhuman robots, scaled humanoids could unlock ~$24T revenue — household and manufacturing markets roughly split (see CN charts).
U.S. manufacturing labor math
~12M manufacturing workers · ~230B hours/year · ~$785B total pay · ~$2.4T output. Extreme case: robots working 16h/day → ~5.9M robots (~half the headcount) could hold current output.
NPV breakeven
Early humanoids will be expensive and far from human capability. ARK NPV threshold: at $16,000 unit cost, need >5% productivity lift vs. a human worker to pencil. Musk: despite higher complexity per mass, eventual cost “shouldn’t exceed half a car.”
Two firm variables
Company size: Large manufacturers organize around specialized, already-automated tasks — productivity historically >2× smallest firms; lower labor share of revenue; higher wages on automated tasks. Small firms juggle multi-task roles; general-purpose automation (humanoids) helps them more because task-specific automation has lagged.
Labor share of revenue: 2021 range from tobacco ~3% and grain/oilseed milling ~6% to machine shops ~37% and apparel knitting ~40% (BLS via ARK). ~40% of U.S. manufacturing employees work at firms <500 people — higher labor cost burden → larger relative gain from general robots.
Tailwinds
Post-COVID supply shocks, onshoring, and labor shortages may lower price sensitivity and speed adoption. ARK: general robotics could be a multi-tens-of-trillions market. Open question: how fast AI software delivers human-like performance across manufacturing subsectors.
Not investment advice. Charts in CN package.
— END —