Highlights

• Hyundai's Atlas targets factory deployment by 2028-2030 while Tesla's Optimus pursues vertical integration, but Chinese firms already shipped 90% of 13,000 humanoid robots globally in 2025 at aggressive sub-$15,000 price points.
• Morgan Stanley projects a multi-trillion-dollar humanoid market by 2050, yet every robot depends on NdFeB permanent magnets where China controls 85-95% of refining and manufacturing—a supply chokepoint Musk admitted affected Optimus production.
• The humanoid boom is fundamentally a rare-earth magnet demand story: long-term winners will be those securing NdPr/Dy/Tb supply outside China, not those with the flashiest demo reels or stock rallies driven by AI speculation rather than robot ROI.

---

A Bloomberg feature (opens in a new tab) casts this as Hyundai’s Atlas (opens in a new tab) versus Elon Musk’s Optimus (opens in a new tab)—a showdown in “physical AI.” Hyundai (via Boston Dynamics) is moving Atlas from demos toward factory deployment, while Tesla pushes vertical integration with Optimus. Investors see a massive opportunity, as Rare Earth Exchanges has reported Morgan Stanley projects a multi-trillion-dollar humanoid market by 2050. But the subplot matters more: humanoids run on high-performance motors, and the magnet supply chain that feeds them remains heavily concentrated in China.

Hyundai acquired (opens in a new tab) Boston Dynamics from SoftBank in 2021.

What we can verify about Atlas

Hyundai/Boston Dynamics has published specs and timelines. Atlas is slated for parts sequencing in 2028 and broader assembly work by 2030, with Hyundai targeting substantial annual manufacturing capacity by 2028. Boston Dynamics lists 56 degrees of freedom, ~50 kg peak lift (30 kg sustained), ~4 hours battery life, swappable batteries, and an industrial operating range from -20°C to 40°C. These are credible, enterprise-grade specifications.

Where the narrative gets wobbly

Bloomberg links Hyundai’s sharp share rally to Atlas momentum. Reuters adds nuance: analysts cited broader AI-related speculation—including potential Nvidia collaboration—rather than robots alone. That distinction matters. “AI halo” enthusiasm and bankable robot ROI are not the same.

China isn’t the future rival—it’s the present market

On shipments, Chinese firms lead. Bloomberg reported (opens in a new tab) in January that China accounted for roughly 90% of ~13,000 humanoid robots shipped globally in 2025. Unitree and AgiBot price aggressively, with some entry-level humanoids listed under $15,000. Volume doesn’t equal productivity—many units are pilots—but it builds supplier scale, data, and cost compression.

The rare-earth choke point inside every robot joint

Humanoids rely on dense electric motors, and NdFeB permanent magnets remain the performance standard. Multiple industry sources (IEA, USGS, DoE, and market research) consistently show that China dominates rare-earth separation and magnet manufacturing—often cited at 85–95% for refining and magnet output. That concentration is already a geopolitical lever. Reuters reported Musk acknowledged Optimus production was affected by Chinese export licensing for rare-earth magnets.

Where Western bets actually rank

Hyundai looks strongest on hardware maturity plus a factory roadmap. Tesla owns the vertical-integration narrative but has encountered friction with magnet suppliers. China leads on scale, pricing, and component ecosystems.

The uncomfortable investor takeaway: the humanoid boom is quietly an NdPr/Dy/Tb demand story. The long-term winners may be those who secure magnet supply outside China—not those with the flashiest demo reel.