Welcome to The Gigacasting Newsletter
New here? Subscribe for weekly updates
The global automotive industry is accelerating the deployment of humanoid robots in factories, shifting from pilot tests to trials in real production lines. Major players like BMW, Toyota, Tesla and Xiaomi have announced integrations for tasks such as precision assembly, high-voltage battery work, and die-casting post processes.
On March 2, 2026, Xiaomi announced via its official Weibo that its humanoid robot completed three consecutive hours of autonomous operation at a self-tapping nut installation station in its EV Gigacasting workshop.
Equipped with a five-fingered dexterous hand, the robot achieved over 90% success while meeting the production line's 76-second cycle time. These are initial tests and I’m sure this is not an acceptable success rate for a relatively simple automated task.
It picks nuts from an automatic feeder, places them in positioning fixtures, and coordinates with conveyors for automated tightening on the rear Gigacasting of the Xiaomi vehicles. Xiaomi notes that precise alignment and reliable engagement remain key technical challenges.
BMW has launched a European pilot at its Leipzig plant, deploying "Physical AI" humanoid robots for high-voltage battery and component assembly. This builds on earlier U.S. trials at Spartanburg and underscores BMW's commitment to exploring humanoid potential in mass production.
Toyota, meanwhile, has moved to commercial deployment of Agility Robotics' Digit at its Woodstock plant in Ontario, Canada. A key site for RAV4 production. Following a successful pilot, Digit now supports employees in logistics and manufacturing under a flexible Robots-as-a-Service (RaaS) model, rather than outright purchase.
Tesla is repurposing existing vehicle lines and building new dedicated capacity to scale its Optimus humanoid robot:
Fremont factory officially started Optimus Gen 3 production in January 2026. The company ended Model S and Model X production to convert that factory space entirely to Optimus. Targeted run-rate: 1 million units per year once fully ramped.
Gigafactory Texas: A massive new dedicated Optimus manufacturing facility is under construction. Long-term goal: 10 million units per year. This site is expected to handle even higher-volume Gen 4 production.
Production of Gen 3 is slated to begin in earnest this summer (2026). High-volume output is targeted for 2027. Tesla’s goal is ~$20k–$30k cost at scale.
Tesla’s CEO, Elon Musk has repeatedly called Optimus potentially Tesla’s biggest product ever, larger than the vehicle business and capable of increasing global GDP by an order of magnitude once at scale. The robots are already being tested for real factory and office tasks, with internal deployment expanding.
Though humanoid robots excel in demos for movement and interaction, commercialization remains challenging, particularly in reliably replacing human labor.
Industrial manufacturing, focused on clear needs like sorting, assembly, and handling, offers the most practical path to scale and stands as the primary breakthrough area for humanoid industrialization noted YZWeekly.com.
Morgan Stanley forecasts the global humanoid robot market reaching $5 trillion by 2050, highlighting its immense long-term potential.
Why is a newsletter focused on Gigacastings and magnesium Thixomolding suddenly talking about humanoid robots?
Simple: I'm no AI expert, I honestly have no idea whether the software will ever get good enough to make truly capable humanoid robots a widespread reality.
But on one point I'm completely confident: if that software breakthrough happens, it means multi-billion $ opportunity for the die-casting industry in North America and Europe.
We're talking lightweight, high-precision structural parts in aluminum and magnesium, perfect for agile, affordable mass manufactured robots.
I want you positioned to capture as much of that upside as possible.
P.S. Decision-makers across the supply chains of humanoid robot startups and companies are already subscribed to The Gigacasting Newsletter. Do the same now
