Tesla had spent years promising an electric motor that could dispense with rare-earth metals. On September 3, Elon Musk attached that promise to the Cybercab.
“The Cybercab motor uses no rare earth metals, but maintains the same range!” he wrote on X. Tesla had announced its rare-earth-free permanent-magnet ambition in 2023. The new claim offered a glimpse of a response to China’s mineral power that receives much less attention than new mines: change the machine so it stops needing the material.
Washington can negotiate an export license, and a manufacturer can stockpile magnets while the talks drag on. Engineers have another option. They can change what goes into the next motor, and every vehicle built around that design carries the decision forward.
That gives the rare-earth investment story an unusual split. Some companies stand to earn more by supplying scarce material outside China. Others could earn more by helping customers use less of it. Both are attracting money. Their long-term economics will look very different.
The factory between the mine and the motor
Rare earths are a family of 17 elements found in deposits around the world. Their availability depends on much more than their abundance in the ground. The hard part is turning the right elements into products a manufacturer can use, at a price and purity it can accept.
For electric motors, the central ingredients are neodymium and praseodymium, usually shortened to NdPr. Combined with iron and boron, they make powerful permanent magnets that fit into small spaces. Dysprosium and terbium can help those magnets retain their performance at high temperatures. A shortage of these less abundant heavy rare earths can therefore hold up a much larger volume of motors.
China’s advantage grows as the material moves toward the finished product. The International Energy Agency’s 2026 rare-earth report puts China’s 2024 share of magnet rare-earth mining at 60%, its share of refining at 91%, and its share of sintered permanent-magnet production at 94%. Those figures describe different steps. Finding another deposit addresses only the first.
Rare earths tend to occur together and behave similarly in chemical reactions. Separating them requires carefully controlled processing, often through many stages of solvent extraction. The resulting oxides still have to become metals, alloys and magnets. Then a customer has to establish that those magnets will perform reliably inside its equipment.
A defense contractor cannot replace a qualified magnet by ordering a vaguely similar one from a new supplier. An automaker has to test heat tolerance and consistency across large production runs. China has accumulated the plants, specialist equipment and operating experience behind those processes. Replicating them takes more than a mining permit.
The export controls introduced in April 2025 exposed that dependence. They required licenses for specified rare-earth products, including materials used in high-performance magnets. The later pause on China’s expanded October 2025 measures, running through November 10, 2026, eased the immediate pressure but left the earlier licensing system in place. A license can allow a shipment through while leaving the buyer dependent on the next approval.
America’s response has consequently spread well beyond prospecting. It includes financing the missing factories, buying from allies, recovering material already above ground, and paying engineers to develop alternatives.
Tesla changes the bill of materials
A permanent magnet supplies a magnetic field without continuously spending electricity to create that field. Rare-earth magnets do this exceptionally well for their size. But motors can also generate magnetic fields electrically. Tesla used induction motors before it introduced permanent-magnet motors with the Model 3. Going without rare earths is possible; doing so while preserving the size and efficiency advantages of a permanent-magnet design is the harder assignment.
Ferrite magnets offer one route: retain permanent magnets, using a cheaper and more abundant material that generally provides less magnetic strength for a given volume. Induction motors and electrically excited designs take another route by changing the motor architecture. These are examples of the available engineering choices, not an identification of Tesla’s undisclosed design.
Every route changes the rest of the design. A weaker magnet may require a larger motor. Electrical excitation can introduce additional losses or cooling requirements. Engineers can compensate through rotor geometry, power electronics, gearing and better thermal management. The practical question is whether the complete drive unit meets the vehicle’s requirements at an acceptable cost.
Cybercab is a useful test because a purpose-built robotaxi gives Tesla control over those requirements. A manufacturer designing around a defined use case has room to choose a different balance of acceleration, efficiency and packaging.
A robot wrist has very little room to hide a larger motor. Extra mass in a hand increases the load on the elbow and shoulder, which can require larger actuators farther up the arm. Saving money on the magnet can mean paying for it elsewhere in the robot. A joint that must hold position or work repeatedly also faces a different thermal problem from a traction motor driving through a changing speed range. A substitute that works in a car does not automatically work in a dexterous hand.
The likely result is selective adoption. Larger joints and stationary industrial machines may accept alternatives that tightly packed hands cannot. Better cooling can reduce the need for dysprosium and terbium even when a motor keeps its neodymium magnets. Some systems may move motors away from the joint and transmit force mechanically. Each successful redesign reduces the amount of controlled material needed for a useful hour of work.
That is the implication for the robot business: companies will compete on how much motion they can get from a constrained supply of magnets. Tesla’s Cybercab disclosure is no proof that Optimus is rare-earth-free, but it shows why magnet chemistry and actuator design belong in any serious discussion of robotics costs.
A Minnesota startup is taking the materials route directly. Niron Magnetics makes permanent magnets from iron and nitrogen. In August it announced a conditional federal loan commitment of up to $150 million, followed by a separate $150 million loan from the Shakopee Mdewakanton Sioux Community. Its first full-scale plant in Sartell is expected to begin production in 2027.
Niron is worth watching for the same reason as the Cybercab motor, although there is no verified supplier connection between them. Its disclosed robotics connection is with Bimotal, a motor startup. Their January 2026 collaboration proposed compact motors for mobility and robotics using Niron magnets. That puts an actual development project behind the robot-joint idea, while leaving customer qualification and volume production ahead. Niron’s next test is producing consistent magnets for demanding customers. Repeat orders would show the product has earned its place.
The mine already above ground
While new magnets are being developed, American manufacturers still need the existing kind. Some of the fastest routes to supply begin with material people have already dug up.
Manufacturing scrap is an attractive starting point. Its composition is relatively well known, it can be collected in concentrated streams, and it has already passed through much of the expensive processing that turns rock into a useful material. Discarded motors and hard drives contain valuable magnets too, although collecting and dismantling them adds work.
Apple has put money behind this approach. Its July 2025 agreement with MP Materials committed $500 million to American-made rare-earth magnets and supported a dedicated recycling facility at Mountain Pass. The companies expect the arrangement to use recycled feedstock, including material from old electronics and industrial scrap. The commercial pull is as important as the processing technology: a buyer is committing to the resulting product.
Industrial waste offers a larger, messier hunting ground. Bauxite residue, the red mud left after extracting alumina, can contain rare earths. So can mine tailings and other processing residues. Recovering them turns a disposal problem into a possible source of revenue, provided the separation process can handle the specific waste stream without consuming more money than the recovered products are worth.
The Energy Department’s June 2026 selections for $134 million in proposed support included a red-mud project at Gramercy, Louisiana, and a Phoenix Tailings demonstration.
Phoenix Tailings addresses the conversion of rare-earth feedstocks into usable metals. It opened a metallization facility in Exeter, New Hampshire, in October 2025, initially focused on NdPr and dysprosium-iron alloy. That sits directly in the gap between ore availability and magnet production. Its proposed Freedom Facility has a conditional $500 million federal debt commitment, with initial operations targeted for 2028. The existing metal-making operation and the planned integrated plant are different stages of the business.
Cyclic Materials is working farther upstream in the recycling process, extracting magnets from discarded equipment and recovering their rare earths. Its August 27 announcement of $75 million in new financing described commissioning work at an Arizona magnet-recovery plant and plans for a larger South Carolina campus. Its appeal is access to material that otherwise disappears into conventional scrap processing. Recovering that material is the first job; turning it into separated elements and finished magnets still requires downstream partners and plants.
Recycling will grow alongside new supply. It cannot immediately recover magnets still working inside vehicles and wind turbines, and a rapidly expanding robot industry needs material before its first generation of machines reaches the scrapyard. But every recovered batch can be processed without opening another mine, and every factory scrap stream gives a recycler feedstock it can use now.
Why the Moon keeps entering the conversation
The Moon does contain rare earths. Lunar scientists even have a name for a geochemical component enriched in them: KREEP, short for potassium, rare-earth elements and phosphorus. A USGS assessment suggests that some local deposits could have concentrations comparable to ores mined on Earth. That is an inference from geological evidence, rather than a proven reserve, but it gives prospectors a serious question to investigate.
The jump from an interesting rock to an ore body is enormous. An ore body needs a useful concentration of the desired elements, a recoverable mineral form and enough accessible material to support the machinery built around it. A lunar operation would also have to supply power, move abrasive dust, process material in vacuum, manage heat and operate with very limited human intervention. Shipping a product back to Earth adds another engineering business on top of the mine.
Putting a refinery on the Moon also recreates the most difficult part of today’s rare-earth problem, in a place without an existing chemical industry. Lower launch costs help transport equipment. They do not establish the grade of the deposit or make neodymium separate itself from its neighbors.
The more plausible first market for lunar mining is local consumption. Water can support life and, after processing, propellant production. Oxygen extracted from lunar material can support spacecraft and habitats. Metals could eventually feed construction and manufacturing in space. A customer on the Moon avoids the cost of lifting an equivalent product from Earth, which changes the economics substantially.
NASA’s resource-utilization work follows that progression: begin with accessible resources and immediate local uses, then develop more elaborate products and markets. Rare-earth shipments to an American magnet factory belong much farther along that path. No operating lunar rare-earth mine is available to relieve today’s shortage.
Asteroids create a similar confusion. The high-value metals commonly discussed by asteroid-mining companies are platinum-group metals. Platinum and palladium are not rare earths and cannot replace neodymium in a motor magnet. An asteroid project can be a serious venture while having little direct bearing on China’s control of magnet supplies.
NASA’s OSIRIS-REx mission illustrates both the achievement and the remaining distance. It returned 121.6 grams of material from Bennu in 2023, an extraordinary scientific sample. The mission was designed for science, so its economics should not be used as a commercial mining cost estimate. It still shows the operational gap between collecting extraterrestrial material and supplying factories by the tonne.
AstroForge is worth following as a test of that distance. It is pursuing asteroid resources, particularly platinum-group metals. Its 2025 Odin mission lost contact before achieving its asteroid reconnaissance goal. In June 2026, the company reported assembling the major flight hardware for DeepSpace-2, intended to attempt an asteroid rendezvous and imaging mission. Successful reconnaissance would give the company something much more useful than a theoretical asteroid valuation: information about a target it might eventually mine. It would still be several steps short of returning saleable material.
Interlune has chosen a different commodity: helium-3, an isotope used in applications including the extreme cooling required by some quantum computers. Helium-3 is a gas, not a rare-earth element. In May 2026, Interlune announced a $6.9 million NASA contract to develop a lunar resource-extraction payload, expected to be ready for launch in 2028. The company lists rare earths among its longer-term ambitions, but its funded experiment concerns gases in lunar soil.
Its chief scientist, Elizabeth Frank, put the next step plainly: “The data we collect will also tell us how much power is needed to extract resources like helium-3.” That is a useful way to judge a space-mining startup. Before anyone can build a credible profit forecast, someone has to measure the energy required to recover the product.
TransAstra offers a way to build useful businesses while working toward asteroid mining. It develops telescope systems for detecting objects and inflatable systems for capturing them. The company reports that its capture-bag technology was deployed on the International Space Station in October 2025. Tracking satellites and developing debris-handling technology can attract customers well before asteroid resources reach a refinery. An ISS deployment establishes progress on a subsystem, rather than proof of a complete mining operation.
These three private companies deserve attention on their own terms. None currently supplies an American factory with rare earths from space.
Follow the purchase order
For public-market investors, the strongest near-term candidates are the companies filling an identifiable gap in the supply chain. A shortage can raise selling prices, but it can also starve a downstream factory of inputs. The business model determines which effect reaches the shareholder.
MP Materials (NYSE: MP) offers the most direct American mine-to-magnet exposure in this group. Mountain Pass supplies the ore and separated rare-earth products; Independence in Texas is building the downstream magnet business. The company’s August 2026 results reported 840 metric tons of NdPr production in the second quarter and further magnet deliveries for customer qualification and regulatory testing.
MP is producing separated material now, while its magnet business is still passing through qualification and scale-up. Its reported Magnetics revenue for the quarter came from precursor products. Investors valuing a mature finished-magnet business would be paying in advance for a stage the company has yet to establish.
Washington has also changed MP’s economics. The 2025 government agreement included a ten-year $110-per-kilogram price floor for covered NdPr products and long-term support for purchases from its planned 10X magnet facility. In the second quarter of 2026, MP recorded $17.6 million in price-protection income, separately from its $108.5 million of revenue. Policy support is already visible in the accounts.
That makes MP more than a bet on spot prices rising. Investors are buying exposure to a domestically financed industrial buildout with contracted support. The next increase in business value depends on turning qualification work and new capacity into sustained customer shipments, with the capital spending and share dilution counted along the way.
America also has a practical option in an allied supplier that is already producing. Lynas Rare Earths (ASX: LYC) mines at Mount Weld in Australia and processes rare earths through its Australian and Malaysian operations. Its fiscal 2026 annual report records 7,260 tonnes of ready-for-sale NdPr production, and shipments of separated dysprosium and terbium oxides during the year.
Lynas described its heavy-rare-earth pricing as reflecting demand outside China, rather than simply following a domestic Chinese index. That gives the investment case a concrete mechanism: customers seeking a separate supply chain may pay a different price for material they can obtain reliably. Production experience and existing customer relationships give Lynas a head start that a new deposit cannot buy overnight. Its report also describes a four-year supply agreement with the U.S. government, evidence that an American solution can include a factory in Malaysia.
The heavy-rare-earth problem makes Energy Fuels (NYSE American: UUUU) worth a closer look. At White Mesa in Utah, the company has produced rare-earth oxides alongside its broader uranium business. In August 2026, it reported that a Japanese magnet manufacturer had qualified its terbium oxide, following earlier qualifications for NdPr and dysprosium.
That customer approval addresses product quality. Volume is the next step. Construction of White Mesa’s larger heavy-rare-earth operation began in July, with the dysprosium and terbium circuits targeted for completion by the end of 2027. Energy Fuels also completed its acquisition of Australian Strategic Materials on August 28, adding an operating metal-and-alloy plant in South Korea. Investors get exposure to constrained ingredients and another step toward finished magnets, alongside uranium operations and a substantial expansion bill.
USA Rare Earth (NASDAQ: USAR) changed its position in the supply chain just this week. On September 3 it completed its combination with Serra Verde, bringing in an operating mine and processing business in Brazil. That joins its British metal-making business, Less Common Metals, and the magnet-manufacturing buildout at Stillwater, Oklahoma. The company can now combine existing upstream supply with its manufacturing plans, although mixed rare-earth output from Brazil still needs further processing before it becomes a magnet.
Its August results targeted 600 tonnes per year of run-rate magnet capacity at Stillwater in the fourth quarter. Steady output and customer shipments will establish the value of that line. Round Top, its Texas deposit, remains a separate development project. USAR’s opportunity is to connect those businesses into a dependable source of materials and magnets; buying the businesses does not finish the integration work.
Farther downstream, Neo Performance Materials (TSX: NEO) gives investors a different angle. Its European operations include rare-earth processing and a new permanent-magnet facility in Narva, Estonia. In its August 2026 update, Neo said the facility was shipping qualification samples and expected two to three customer programs to enter commercial production during 2026.
Here the potential payoff comes from becoming an approved supplier to customers diversifying away from China. It requires dependable feedstock and a margin on converting that feedstock into a component. Higher raw-material prices alone do not guarantee higher profit. Neo also has a substantial rare-metals business, including gallium, hafnium and tantalum, so its earnings should not be treated as a pure measure of rare-earth magnet demand.
Tesla belongs on the other side of the trade. Tesla (NASDAQ: TSLA) is a user trying to reduce its exposure to scarcity. A successful motor redesign could improve cost control and production reliability, rather than produce a windfall from higher rare-earth prices. The benefit would sit inside a much larger vehicle, energy and autonomy business. Buying Tesla is a very indirect way to express a view about magnets.
A shortage pays for its own competition
I would start with the companies whose customers are already paying to escape the bottleneck. Lynas is shipping material, while MP’s price protection and Apple’s purchase commitment help finance a larger American supply chain. Niron’s factory financing backs a more ambitious route, one that changes the material itself. Those are different reasons to own a business, and none excuses paying any price for its shares.
Space mining belongs on a different investment clock. Successful prospecting, processing demonstrations and returned material would make those ventures progressively more valuable. They would still need to show that the specific commodity and its delivery cost fit a paying market. Lunar oxygen sold to a lander and neodymium sold to a Texas magnet factory are separate businesses.
For the next several years, the United States and its allies are likely to gain far more from better terrestrial processing, recycling and motor design than from extraterrestrial rare earths. Over a longer period, the engineering response could change demand enough that today’s supply shares become a poor guide to tomorrow’s pricing power.
China has spent decades becoming the default source for high-performance magnets. Its customers are now financing other places to buy them. Some are financing ways to stop buying them at all.
Disclaimer: This article is for information and discussion, not personalized investment advice.
Sources
International Energy Agency: Rare Earth Elements, 2026 report
China Ministry of Commerce: suspension of expanded October 2025 export controls
Niron Magnetics: conditional $150 million federal loan commitment
Niron Magnetics: $150 million Shakopee Mdewakanton Sioux Community loan
Niron Magnetics and Bimotal: motor-development collaboration, January 2026
Apple: $500 million U.S. supply-chain commitment and MP Materials agreement, July 2025
U.S. Department of Energy: $134 million in proposed critical-mineral project support, June 2026
Phoenix Tailings: Exeter metallization facility opening, October 2025
Phoenix Tailings: Freedom Facility initiative and conditional federal financing
Cyclic Materials: $75 million growth financing, August 27, 2026
AstroForge: DeepSpace-2 spacecraft assembly update, June 2026
Interlune: $6.9 million NASA lunar resource-development contract
Energy Fuels: Japanese customer qualification of heavy rare earths, August 19, 2026
Energy Fuels: heavy-rare-earth plant construction, July 29, 2026
Energy Fuels: completed Australian Strategic Materials acquisition, August 28, 2026
USA Rare Earth: completed Serra Verde combination, September 3, 2026