Highlights

• Tesla, Chery, Seres, XPeng, and Hyundai are racing into humanoid robotics in 2026, repurposing EV factories and supply chains as China’s auto price war pushes carmakers toward higher-value businesses.
• Tesla targets mass production of Optimus Gen 3 by end-2026 with 1 million robots/year capacity, while Chery deploys traffic-police robots and Seres partners with ByteDance on AI-driven robotics.
• Humanoid robots depend critically on rare earth magnets (neodymium, terbium, dysprosium) for actuators and motors, creating new supply competition for heavy rare earths mostly processed in China.

---

As we enter Chinese New Year the nation’s carmakers are sprinting out of the EV price war—and into robots.

A state-owned Chinese financial outlet (Shanghai Securities News / cnstock.com, under Xinhua) reports that, in early 2026, Tesla, Chery, Seres, XPeng, Hyundai, and major auto-supply-chain firms are racing into “embodied intelligence”—humanoid and task robots meant for factories, showrooms, and public-facing services. The piece frames the unfolding via a Chinese idiom meaning clearing out old capacity to make room for a higher-value new business (think: repurposing auto factories and supply chains for robotics).

What makes this a real business news item?

1) Tesla is (again) signaling a hard pivot—not just a demo.

The Chinese report says Optimus Gen 3 (opens in a new tab) will appear in 2026, with a target of mass production by end-2026 and an eventual ~1 million robots/year capacity, with Fremont’s Model S/X line repurposed for robots. This is not just Chinese media talk: recent Western reporting echoes the same timeline, capacity ambition, and the revised hand/dexterity emphasis as a gating hardware constraint. Translation: Tesla is presenting Optimus as a production program, not a science-fair prototype.

2) Chery is showcasing “robots in the street,” not just on stages.

Chery’s Mo Jia robot “Wuyou” traffic-police R001 reportedly began duty at a Wuhu intersection on January 10, 2026, integrating with traffic signals and assisting police. Separate reporting corroborates the deployment date and location. The Spring Festival TV cameo is marketing, but the operational claim—real-world fielding—makes it notable. See more in EEWorld (opens in a new tab).

3) Seres is building a second engine beyond Huawei-linked autos.

Seres (opens in a new tab) formed a new Shanghai entity (RMB 50M) focused on robotics/AI software and partnered with ByteDance’s Volcano Engine on robot decision/control tech. The strategic subtext is clear: keep EV cash flow while buying an option on robotics margins. (Still: corporate registrations and MOUs are not revenue.)

4) Hyundai’s U.S. factory robot push is framed as cost-defense.

Hyundai is described as deploying humanoids in U.S. plants from 2026 and scaling Atlas (Boston Dynamics) usage—positioned as a response to margin pressure and competitiveness. That’s a very “boardroom” motive: labor substitution + throughput resilience.

Investor-grade skepticism: where the story may overreach

• “Stop multiple best-selling models” is dramatic language. Tesla’s S/X are important flagships, but not its volume core; treat the “sacrifice” framing as narrative.
• Million-unit robot capacity is an ambition, not a delivered run-rate. Musk himself has cautioned early production will be slow.
• China’s model is the point: many entrants rush in, margins collapse, then a few scalemonsters emerge. The West should read this as a supply-chainmobilization story, not a guaranteed humanoid profit story—yet.

Rare Earth Elements/Magnets Not Optional

Rare earth elements—especially those used in high-performance NdFeB (neodymium-iron-boron) permanent magnets—are foundational to modern robotics and humanoid systems. Every precision actuator, joint motor, servo drive, harmonic reducer, and compact electric drivetrain inside a robot depends on dense, lightweight magnets to deliver high torque in small form factors with low heat loss. Neodymium and praseodymium (NdPr) provide the magnetic strength, but the real strategic choke point lies in the heavy rare earths—notably terbium (Tb) and dysprosium (Dy)—which are added in small percentages to maintain magnet performance at high temperatures and under repeated mechanical stress.

In humanoids, where dozens of tightly packed motors operate continuously in limbs, hands, and mobility systems, thermal stability is not optional; without Tb and Dy, magnets demagnetize under load. That means heavies are not bulk materials, but they are performance-critical multipliers. As robotics scales from prototypes to mass production, demand will concentrate on high-coercivity magnet grades—the same materials essential to EV drivetrains, drones, and defense systems—intensifying competition for limited heavy rare earth supply, most of which is currently processed in China. In short, the humanoid revolution is also a magnet story—and therefore a rare earth story.

Disclaimer: This item originates from multiple Chinese news sources, including Shanghai Securities News / cnstock.com, a state-owned outlet under Xinhua; all claims—especially production targets and factory conversions—should be verified with independent sources.