MP Materials (NYSE: MP): Rare Earth Magnets for Robots and Defense
MP Materials owns Mountain Pass, the only operating rare earth mine in the United States, and it is now doing the harder half of the job: turning that ore into finished neodymium magnets on American soil. It is the rare commodity company that stopped being a commodity company — because the U.S. government guaranteed it a price.
Who they are
Mountain Pass, in California's Mojave Desert, was once the center of the global rare earth industry. Then it wasn't — the mine went bankrupt in 2015, China took the market, and the site sat mostly idle. MP Materials is the company that bought the pieces, restarted the mine, and spent the following decade trying to rebuild everything that used to sit downstream of it.
Rare earths aren't actually rare. What's rare is wanting to do the chemistry. Separating seventeen chemically similar elements is an unglamorous, solvent-heavy, capital-hungry process that China spent thirty years getting very good at and very cheap at. By most counts China controls roughly 70% of global mining and around 90% of processing, which means everyone else's supply chain has a single point of failure sitting in it.
MP is the American attempt to remove that point of failure, and its business now spans two segments: Materials (Mountain Pass — concentrate, separated oxides, metal) and Magnetics (the Independence plant in Fort Worth, Texas, plus the much larger 10X facility under construction in Northlake).
What they do
The operational story of 2025 and 2026 is vertical integration finally arriving. Mountain Pass produced a record 2,599 metric tons of NdPr oxide in 2025, up 101% year over year, alongside 50,692 metric tons of rare earth oxides in concentrate. In December 2025 the company produced its first NdFeB permanent magnets on commercial equipment at Independence — the moment MP stopped being a mining company with a magnet slide deck.
Qualification is the gate now, not capacity. Magnets have been delivered to General Motors for in-car qualification and regulatory testing, with initial commercial deliveries targeted for the fourth quarter of 2026. That sequence — sample, qualify, then ship in volume — is the same gate every component in the robot supply chain has to pass, and it is why announcements and revenue are often years apart.
Meanwhile the capacity plan is deliberately enormous: Independence is expanding from roughly 1,000 to 3,000 tonnes of annual magnet capacity, and the 10X facility is designed to lift the total to about 10,000 metric tons per year — hence the name. Construction broke ground in the first quarter of 2026 and management describes it as accelerating.
How they make money
Second quarter 2026 revenue was $108.5 million, up 89% year over year, with adjusted EBITDA of $28.5 million. Capital expenditure in the same quarter was $230.3 million. Read those two numbers next to each other and you have the whole shape of this business right now: it earns real money, and it is spending multiples of that money building the thing it actually wants to be.
What makes MP structurally unusual is the July 2025 public-private partnership with the Pentagon. The headline terms: a 10-year price protection agreement setting a floor of $110 per kilogram of NdPr — more than double MP's realized price of about $51/kg in 2024 — a commitment that 100% of the magnets produced at 10X will be purchased by defense and commercial customers for a decade, a $400 million equity investment, a $150 million loan for heavy rare earth separation, warrants that could take the government to 15% ownership, and profit sharing above defined EBITDA thresholds. JPMorgan and Goldman Sachs separately committed $1 billion of financing for 10X construction.
Commercial customers followed the government. Apple committed to a long-term agreement reported at up to $500 million for recycled-feedstock magnets, including a $32 million prepayment in February 2026 toward a 3,000 t/yr expansion. GM has a long-term agreement for U.S.-sourced rare earths, alloys and magnets. In 2026 MP also launched Project Swarm, an initiative to aggregate demand and standardize magnet specifications for the drone industry, and signed a long-term gadolinium supply deal with an aerospace and defense customer.
Where it sits in the value chain
MP sits at the very top of the robot supply chain, in Stage 1 — Magnets & critical materials. Nothing downstream works without it: a humanoid's torque comes from permanent magnet motors, and industry estimates put roughly 2 to 3 kg of NdFeB in a single humanoid — several times the magnet content of a typical EV, because the magnet is distributed across 25 to 30 joints instead of one or two traction motors.
Robotics is not yet MP's main demand driver — EVs, wind, defense and data center cooling pumps are all bigger today. But it is the demand curve with the steepest theoretical slope, and unlike most of this chain, magnets are consumed per unit rather than licensed. Every robot built is magnet demand that does not come back.
The bigger trend
Magnets became a policy instrument in 2025, and 2026 is the year the policy started biting on both sides. China tightened export controls on rare earth technology and heavy elements; the U.S. moved in the other direction, with defense sourcing rules extending restrictions on Chinese-origin NdFeB across the supply chain from January 1, 2027. For any Western robot program that wants to sell into defense or government-adjacent markets, magnet provenance is quietly becoming a design requirement rather than a procurement detail.
That is the tailwind. The counterweight is that MP is still a young manufacturer attempting to do, in a few years, what China's magnet industry took three decades to learn. Yield, qualification cycles and cost per kilogram are engineering problems that price floors don't solve. The floor buys time; it doesn't buy competence, and MP will be judged on whether the magnets coming out of Texas actually pass in customers' products.
What to watch
- Whether GM's in-car qualification converts into the initial commercial magnet deliveries guided for Q4 2026 — the first real proof that Independence output meets automotive-grade spec.
- 10X construction milestones in Northlake, and whether the 10,000 t/yr target holds its schedule as capex runs at $500–600 million a year.
- Heavy rare earth separation (dysprosium and terbium) coming online at Mountain Pass — the elements that make magnets survive heat, and the part of the chain China dominates most completely.
- Whether Project Swarm produces named, standardized drone and robotics magnet orders or stays a demand-aggregation concept.
- Political durability of the price protection agreement, including how its annual cost is appropriated.
Related companies
🇯🇵Buys the motors MP's magnets go into — the reducer half of the same robot joint.
🇰🇷Actuator maker one stage downstream, where magnet supply becomes a joint.
🇰🇷End customer class: every cobot and humanoid it builds is magnet demand.
Also in this stage but not yet profiled: Lynas Rare Earths (ASX: LYC), USA Rare Earth (Nasdaq: USAR) and JL MAG Rare-Earth (SZSE: 300748).
FAQ
Does MP Materials actually make magnets, or just mine rare earths?
Both, now. It has mined and separated rare earths at Mountain Pass for years, and in December 2025 it began manufacturing NdFeB permanent magnets on commercial equipment at its Independence plant in Fort Worth, Texas. Magnet volumes are still small relative to the materials business, with the much larger 10X facility under construction.
What is the $110 per kilogram price floor?
It is the core of a 10-year price protection agreement signed with the U.S. Department of Defense in July 2025. The government guarantees an effective floor price of $110/kg for the NdPr content in MP's products, whether sold, stockpiled or used internally for magnets, and shares in the upside if market prices run above that level. For context, MP's realized price in 2024 was roughly $51/kg.
How much rare earth magnet does a humanoid robot need?
Industry estimates commonly cite around 2 to 3 kg of NdFeB per humanoid, spread across 25 to 30 actuated joints. That is several times the magnet content of a typical electric vehicle, which is why robotics is treated as a structurally attractive long-term demand channel for magnet makers even though the current unit volumes are tiny.
Is MP Materials profitable?
It generates positive adjusted EBITDA — $28.5 million in the second quarter of 2026 on revenue of $108.5 million — but it is spending far more than it earns on construction, with capital expenditure guided at $500 to $600 million for full-year 2026. Reported bottom-line results move around with start-up costs at the new magnet facilities.