During a recent town hall, Tesla management informed workers at the Austin Gigafactory that the company will commence data collection to train its Optimus humanoid robots on the Texas manufacturing lines as early as February. This represents the second major robotics training site for Tesla, complementing the more than year-long training program already underway in Fremont, California. The move leverages the scale of Tesla’s Austin workforce—numbering in the thousands—to accelerate data gathering in a facility that produces both electric vehicles and the upcoming Cybercab robotaxi.

In Fremont, Tesla employs several dozen dedicated Optimus trainers who record themselves performing routine tasks—such as organizing vehicle parts and operating conveyor belts—while wearing large camera-equipped helmets and heavy backpacks. Video footage captured from multiple angles feeds the company’s machine-learning models, teaching robots how to mimic human motions. Tesla has experimented with lighter data-capture rigs, including fanny-pack cameras, but insiders indicate the helmet-and-backpack system remains the primary tool for consistent, high-quality recordings.

At the World Economic Forum in Davos, CEO Elon Musk confirmed that Optimus prototypes are now executing “simple tasks” on the Fremont factory floor, with plans to tackle more complex operations by year’s end. He projects public sales of the humanoid robot by late 2027. However, Musk cautioned that production of both Optimus and the Cybercab robotaxi in Austin will be “agonizingly slow,” underscoring the technical and logistical challenges of scaling novel robotics hardware alongside mass-market electric vehicles.