Report: Apple started a robotics lab in 2012 to reduce human labor needed to assemble products, but abandoned it in 2018 when the robots couldn't match humans
William Gallagher / AppleInsider : Source: The Information .
Context & Ripple Effects
The 2012–2018 robotics lab is the earliest documented attempt in Apple's decade-long effort to cut humans out of product assembly — an effort driven partly by labor costs and partly by the concentration of that labor in China. As far back as 2015, some operations executives wanted to move part of manufacturing to Vietnam, and senior management rebuffed the relocation, keeping production tied to a workforce the lab was meant to make dispensable.
The lab's quiet shutdown matters because it explains the gap before the current push: reporting on Apple's renewed drive to fully automate iPhone assembly lines frames automation as a way to reduce reliance on China, acquiring startups like DarwinAI along the way. Read together, the arc runs from failed labor substitution, to rejected diversification, to automation recast as a geopolitical hedge.
First-order effects
- Through 2018 the lab's failure left Apple's assembly volumes dependent on human line workers at existing Chinese contract manufacturers, reinforcing exactly the concentration the Vietnam proposal had tried to dilute.
- The abandonment handed Apple's operations organization proof of technical limits: dexterity-sensitive assembly tasks stayed human, so any near-term cost or supply-chain flexibility gains had to come from elsewhere.
Second-order effects
- When Apple returned to assembly automation around 2024 via startup acquisitions like DarwinAI, the objective had changed from labor savings to China de-risking — meaning success is now measured against supply-chain exposure, a bar the 2012 program never had to clear.
- Contract manufacturers who absorbed the failed-automation risk in 2018 gained six more years of human-labor-based volume, while Apple's internal AI dysfunction reported elsewhere raises the odds that the second attempt inherits the same execution problems.
Third-order effects
- The lab joins a pattern — the eight-year car program undone by shifting goals and a revolving door of leaders, and AI/ML work bogged down by organizational dysfunction — suggesting Apple's structural weakness is sustaining long-horizon hardware moonshots that don't ship through a near-term product team.
- If the pattern holds, Apple's manufacturing footprint stays anchored where skilled human assembly already exists until robot dexterity improves independently of Apple's internal efforts, keeping supplier geography — not Apple's labs — as the binding constraint.
The trend: Apple's assembly-automation ambition is cyclical, reviving whenever China exposure becomes acute even though robot dexterity has twice failed to match its line workers.