Takes & Tickers – June 5, 2025 [Edition 3]: The Impact of Rare Earth Metals on EV Manufacturing and Robots in the Last Mile
Why China’s grip on Dysprosium and Terbium threatens upstream EV production, and how Veho’s pilot with Rivr opens new possibilities in last-mile delivery automation.
Welcome to the third edition of Takes & Tickers — a weekly lens on the most strategic moves shaping the global supply chain.
This Thursday, I break down two key stories from the week—decoded with quotes, sourced links, and strategic takes you won’t get from headlines alone.
Your Voice Matters!
Help me tailor content that truly fits your needs and skills by taking our quick survey. It takes less than 1 minute — and makes a big difference.
Please take the survey now, click here
Here is what I unpack this week:
Rare Earth Metals and China’s Strategic Leverage: China’s dominance in refining heavy rare earths like Dysprosium and Terbium has exposed a critical vulnerability in the EV supply chain. Without vertical integration, the U.S. remains dependent on Chinese processing, making onshoring economically and technically uncompetitive, for now.
Veho and Rivr’s Robot Pilot in the Last Mile: Veho and Rivr’s robotic delivery pilot offers a glimpse into how automation may reshape last-mile logistics through labor augmentation and micro-efficiencies. While early-stage, the move signals intent to build operational fluency as Physical AI matures and scales.
1. Rare Earth Metals and China’s Strategic Leverage: Why Your EV Might Be Caught in the Middle
Four major automakers are racing to find workarounds to China’s stranglehold on rare-earth magnets, which they fear could force them to shut down some car production within weeks.
Electric vehicles are running into a roadblock, and it has nothing to do with batteries or software. The problem is magnetic, specifically, Dysprosium (Dy) and Terbium (Tb), two heavy rare earth metals that help electric motors survive high temperatures without losing power.
China controls over 90 percent of global Dy and Tb supply, and in late 2023, it placed restrictions on exporting magnets containing these metals. These are not fringe materials. They are essential to high-performance NdFeB (Neodymium-Iron-Boron) magnets used in everything from EV traction motors to F-35 jets. The strategic consequences are real.
The Tariff Workaround No One Wants
Some automakers considered creative loopholes. For example, manufacture motors in the U.S., ship them to China to install the magnets, then import the finished units back. Since tariffs target magnets, not assembled components, this sidestep is technically legal.
But economically, it’s a dead end. The time, cost, and complexity of such a workaround make it unsustainable. And as Ford learned in May, when it halted production of the Ford Explorer in Chicago due to a Dy/Tb shortage.
How the Rare Earth Motor Supply Chain Really Works
The supply chain for rare earth-based EV motors spans continents but revolves around a single axis: China.
Mining: Dy, Tb, Nd, and Pr are extracted mostly in China, Myanmar, and Australia.
Refining & Separation: Dy and Tb require multi-stage solvent extraction, a process China has refined over decades.
Magnet Manufacturing: Refined oxides are alloyed into NdFeB magnets, often using Dy or Tb to boost coercivity for high-temperature durability.
Motor Assembly: These magnets are installed into rotor assemblies by global Tier-1 suppliers, including some in the U.S.
EV Integration: Motors are integrated into vehicles at Tesla's facilities in Austin or Fremont.
Without control over steps 2 and 3, no nation can claim true independence in EV motor production.
The Science of Staying Magnetic
NdFeB magnets are incredibly powerful, but they degrade at high temperatures, above 100 to 150°C. In EV traction motors, internal temperatures often exceed this threshold. That’s where Dy and Tb come in. See chart below, where adding a coat of Dy or Tb, improves coercivity of NdFeB.
These heavy rare earths substitute into the magnet crystal structure,
They increase magnetic anisotropy, which helps retain magnetization under heat,
And they protect against thermal demagnetization, maintaining torque and power output.
Without Dy/Tb, magnets lose strength, and motors lose efficiency. High-performance EVs like Teslas rely on this margin.
Can the U.S. Catch Up?
In theory, yes. In practice, not without major investment and patience.
China’s dominance is not based on patents. It’s built on tacit industrial knowledge. Solvent extraction of heavy rare earths involves hundreds of stages, with precise pH, temperature, and flow rates. Minor errors lead to yield loss or cross-contamination.
Reverse-engineering this is hard because:
Every ore body is different,
The refining recipes are not published,
And the labor cost and regulatory burdens in the U.S. make trial-and-error expensive.
Plus, according to S&P Global, it takes 29 years on average to bring a new U.S. rare earth mine to production. And even then, most U.S. mines can’t separate Dy/Tb economically.
The Missing Link: Integration
Even when rare earths are available, they often occur in low concentrations or as by-products. To justify extraction, there must be downstream demand—refining, magnet making, and motor building.
Right now, the U.S. lacks:
Commercial heavy REE separation,
Industrial-scale NdFeB magnet fabrication, and
Dense local demand clusters (like those in Guangdong or Chongqing).
As a result, raw materials often boomerang back to China for processing. Until the U.S. builds a vertically integrated value chain, it won’t capture the economic upside of its own resources.
The Cost Gap Is Huge
Let’s talk price. According to industry data:
In China, Dy/Tb oxides cost $11–15 per kilogram.
In Brazil, the cost rises to $35–40.
In the U.S., estimates exceed that, due to lack of scale, higher wages, and strict environmental regulations.
This is why most non-Chinese players either license tech from China or accept long lead times.
So What Now?
There are only a few viable paths:
Motor redesign: Tesla hinted in 2023 Investor Day that it may phase out rare earth magnets entirely. That would require moving to ferrite-based or induction motor architectures, which are heavier and less efficient.
Material innovation: Companies are experimenting with Grain Boundary Diffusion (GBD) to reduce Dy/Tb usage by up to 90 percent, without performance loss.
Allied sourcing: Australia, Canada, and Vietnam have reserves, but lack processing scale. They will need investment and political coordination to compete.
The Strategic Takeaway
This is a materials war, but also an economic one. China wins not just because it mines rare earths, but because it processes them, integrates them, and uses them across industries. The U.S. holds some of the raw material, but without the midstream, it remains dependent.
EV makers must now decide: Rebuild the supply chain, redesign the motor, or remain exposed.
2. Rivr and Veho Partner to Deploy Robots for the Last Mile
From Veho blog:
“We’re partnering with RIVR, a physical AI and robotics company, to test AI-powered delivery robots that will support driver-partners on their delivery routes—starting this month in Austin, Texas.
… This isn’t automation for automation’s sake. It’s a thoughtful experiment with real-world impact. If successful, the pilot has the potential to reduce physical strain on driver-partners by allowing robots to handle more of the repetitive or physically demanding work. It could also significantly improve route efficiency, improving on time delivery for customers. With improved efficiency comes the opportunity to lower delivery costs for the shippers, all while maintaining speed and consistency in getting parcels to customers’ doors.”
Can robots create the Aha! moment the last mile has been waiting for?
While the announcement from Veho and Rivr is over a week old, the implications are still worth exploring. It’s easy to dismiss such a move as PR or hype. But that would be an oversimplification. At its core, this pilot addresses a real operational need. It aims to reduce physical strain on delivery partners by outsourcing repetitive or physically demanding tasks to autonomous ground robots. There’s also potential to improve route efficiency and delivery consistency, especially in residential zones.
From that perspective, it’s a logical experiment. And while it may still be early, it is not without strategic merit.
The Value Proposition Today
The promise lies in how robots can augment human labor, not replace it. The current pilot likely operates under semi-structured conditions. Delivery partners stay nearby and collaborate with the robot in real time. In these use cases, robots can reduce walking time for drivers, handle short-range doorstep drop-offs, and help drivers cover more ground without fatigue.
Even if the robot’s performance is limited by speed or environmental constraints, it opens up an additional delivery thread. This thread runs in parallel with the main route and can ease some operational pressure.
The Technical Horizon: When Physical AI Matures
Where this becomes transformational is when Physical AI advances. At that point, robots could manage autonomous routing and real-time decision-making. That would unlock new capabilities.
Robots could recalculate paths at the level of sidewalks, gates, and driveways. They could integrate live terrain, pedestrian, and obstacle data. They could even synchronize precisely with delivery vans, allowing for dynamic hand-offs mid-route.
This is not science fiction. We already see early examples in warehouses and closed campuses. But scaling to residential streets demands thorough training, robust safety systems, and adaptable intelligence.
Where the Challenge Lies
The real challenge is not just technical. It is systemic.
Training robots to navigate curbs, cluttered sidewalks, or unmarked entrances takes time. It requires a real-world learning loop fed by data from thousands of neighborhoods.
Liability and safety remain complex. Parked cars, children playing, and unpredictable pedestrian behavior introduce risk. Companies have not fully solved for these scenarios yet.
High-rise apartments and gated communities add another layer of difficulty. These environments are less suited for early-stage robotic delivery.
So while the use case may not scale universally today, that does not make it fruitless. It should be viewed as a testbed. This is a space for learning, refining, and understanding what robotic deployment really demands.
Looking Forward, Not Hyping Forward
Is this pilot revolutionary? Not yet.
Is it premature? Possibly, in certain conditions.
Is it strategically valuable? Yes. It contributes to a broader investment in the design of the last mile.
We should see the Veho and Rivr collaboration not as a finished product, but as a signal of intent. It marks a step toward rethinking how last-mile delivery works. Much like early autonomous vehicle pilots, the value here lies in data, workflow experimentation, and learning from real-world use.
As Physical AI improves, these early moves give companies a foundation. They can build operational knowledge, test risk models, and prepare for wider deployment. We are not there yet. But we are closer than ever.
