MP Materials Corp. (MP)

The only fully integrated US rare-earth-to-magnet producer — the NdFeB motor-magnet bottleneck behind every humanoid robot — backstopped by the Pentagon, Apple and GM, and the one name in this batch trading on a calm tape.

Investment Research · Photoncap-style deep dive · v1 of "MP Materials" · May 14, 2026

What MP Materials physically does

MP Materials sits at the chokepoint of the robotics supply chain that no compute or software company can engineer around: the motor magnet. Every humanoid robot, every robot arm, every actuator is built around electric motors, and the highest-performance electric motors use neodymium-iron-boron (NdFeB) permanent magnets — the strongest commercially available permanent magnets, which let a motor deliver more torque in a smaller, lighter package. A humanoid robot can contain dozens of motors; the magnet content is real, recurring, and physically irreplaceable with current motor architectures. The rare-earth elements at the heart of those magnets — neodymium and praseodymium (NdPr), plus heavy rare earths like dysprosium and terbium for high-temperature performance — are the binding material constraint, and China has dominated their mining, separation and magnet manufacture for two decades.

MP Materials is the Western answer to that dependency, and uniquely it spans the entire chain. It owns and operates Mountain Pass in California — the only operating rare-earth mine in the United States — where it mines and now also separates rare-earth ore into purified NdPr oxide and metal. It then runs the Independence facility in Fort Worth, Texas, where it produces rare-earth metal, alloy and — as of December 2025 — finished NdFeB permanent magnets. No other US company is fully integrated mine-to-magnet. MP describes itself, accurately, as the only fully integrated rare-earth producer in the United States covering the whole supply chain from mining and processing through metallization and magnet manufacturing.

Why is this the binding constraint of the robotics cycle in a way the chip names are not? Because compute has multiple credible Western suppliers (NVIDIA, Qualcomm, Ambarella, Lattice) and the bottleneck there is competitive, not physical. Magnets are different: the material is geologically concentrated, the separation and magnet-making expertise is concentrated in China, and China has demonstrated — through its 2025 export controls — both the willingness and the ability to weaponize that concentration. If humanoid robots scale, the magnet supply chain is a hard physical constraint, and MP is the only listed pure-play that owns the entire Western alternative. The honest framing, developed below: MP's robotics revenue today is essentially zero — magnets only started in December 2025 and the customer relationships disclosed so far are Apple and GM, not robot OEMs — but the structural position is the most genuinely "bottleneck" of any name in this compute-and-materials batch.

Product roadmap

MP's roadmap is a vertical-integration build-out across two segments. The Materials segment is the established business: mining at Mountain Pass and the ramp of on-site separation, producing NdPr oxide and metal. Q1 2026 marked record NdPr output — the separation ramp is the near-term execution story. The Magnetics segment is the new leg: the Independence facility in Fort Worth commenced NdFeB permanent-magnet manufacturing in December 2025, and MP has said it expects initial magnet revenue in the second half of 2026. Beyond Independence, MP is building the "10X Facility" — a much larger magnet plant — and is advancing a magnet-recycling line at Mountain Pass tied to its Apple agreement, which would create an incremental domestic feedstock source containing both light and heavy rare earths. MP is also pursuing heavy-rare-earth expansion (dysprosium, terbium), the elements needed for high-temperature magnet performance and historically the hardest part of the chain to source outside China.

The dated milestones that matter: magnet manufacturing started December 2025; initial magnet revenue is guided for H2 2026; the recycling line has completed conceptual design and is moving through engineering and procurement; the 10X Facility is in build-out. What MP does not do is make motors, actuators or robots — it makes the magnet, the alloy and the separated oxide. It also does not yet have a disclosed robot-OEM customer; its anchor offtake relationships are in consumer electronics (Apple) and autos (GM), with the Department of Defense as a funding-and-offtake backstop. The robotics demand is a thesis about where magnet demand goes, not a current contracted book.

The financial print

MP Materials reported Q1 2026 on May 7, 2026: total revenue of $90.6 million, up 49% year-on-year, driven by higher NdPr oxide and metal sales as separation ramped and on stronger pricing. Including price-protection-agreement income, consolidated revenue plus PPA income was $132.9 million ($90.6 million revenue plus $42.3 million PPA income). Adjusted EBITDA swung to $36.6 million, a $39.3 million year-on-year improvement, helped by the revenue growth and the PPA income. The newly reporting Magnetics segment produced $21.1 million of revenue and $9.6 million of segment adjusted EBITDA — a real first contribution from the magnet business, though MP still posts a consolidated GAAP net loss as the build-out absorbs capital.

The funding structure is a core part of the financial story. MP received $32 million in Apple prepayments in February 2026, bringing total Apple-related prepayments on the balance sheet to $72 million. It has a Department of Defense partnership providing project funding and offtake support, GM prepayments, an $11.0 million Section 48C tax credit and 45X production credits — a stack of government and customer capital underwriting the Independence and 10X facilities and the heavy-rare-earth expansion. Sell-side analysts forecast roughly 67% full-year 2026 revenue growth with full-year EPS around $0.33; forward P/E is approximately 48.6x against a roughly $10.6 billion market cap.

Crucially, the technicals: MP prints an RSI of 44.3 and sits just +0.9% above its 50-day moving average. That is the single most important fact for the trade — MP is the only name in this entire compute-and-materials batch that is not extended. It has consolidated, not melted up, and that gives an entry an actual margin of safety the chip names and the memory names do not offer. The binary event is Q2 2026 earnings, expected early August 2026 (MP reports on a roughly quarterly cadence; August 6, 2026 is the working estimate, to be confirmed), with the H2 2026 initial-magnet-revenue guidance the milestone to track.

Customer mix today

MP discloses by segment and by anchor relationship rather than by named-customer percentage. The Materials segment's revenue is NdPr oxide and metal sales — historically a meaningful share of MP's rare-earth concentrate went to Chinese processors, but MP halted all rare-earth exports to China as of April 2025, redirecting volume to its own domestic separation and to Western customers. The Magnetics segment, brand-new, has its offtake anchored by Apple (the $72 million of prepayments is tied to a magnet-supply agreement) and GM (which also has prepayments in place), with the Department of Defense as the strategic funding-and-offtake backstop.

The 2024-to-2026 change is the entire story here. In 2024, MP was effectively a rare-earth concentrate miner whose output flowed substantially into the Chinese processing chain — a price-taking, single-product, geopolitically exposed position. By Q1 2026, MP has stopped exporting to China, is separating its own material domestically (record NdPr output), has started making finished magnets (December 2025), and has converted Apple, GM and the Pentagon into prepaying, capital-providing anchor partners. That is a transformation from "commodity miner exposed to China" to "integrated, government-backed, Western-supply-chain magnet producer." On robotics specifically: there is no disclosed robot-OEM customer. The robotics exposure is a demand thesis — robotics is forecast to become the single largest driver of NdFeB magnet consumption by 2040 per industry research — not a contracted relationship today. The honest framing: the customer mix is rapidly de-risking geopolitically and moving up the value chain, but the named customers are a phone maker and an automaker, with robotics as the structural tailwind behind the whole magnet build-out rather than a current line item.

What's actually happening at the Independence facility and the demand side

Two mechanisms to watch. On the supply side, the execution story is the Independence facility ramp: magnet manufacturing started December 2025, and the question is whether MP can scale finished-magnet output and convert the H2 2026 initial-magnet-revenue guidance into a real, growing line. The Q1 2026 Magnetics segment print — $21.1 million revenue, $9.6 million segment EBITDA — is the first proof point that the magnet business can contribute, supported by magnetic-precursor production while finished-magnet volume builds. The recycling line at Mountain Pass, tied to the Apple agreement, is the feedstock-security mechanism — completing engineering and procurement now, it would give MP a domestic source of both light and heavy rare earths independent of mined supply.

On the demand side, the mechanism is the structural shift in where NdFeB magnets get consumed. The industry research is consistent that robotics — and the humanoid sector specifically — becomes the dominant driver of magnet demand as production scales toward potentially billions of units over the long horizon, and global rare-earth magnet demand is already around 385,000 tonnes a year worth roughly $19 billion, growing in the high-single-digits. Be specific and skeptical, though: that demand is a long-horizon forecast (the "by 2040" framing matters), humanoid volumes today are tiny, and MP has not disclosed a single robot-OEM offtake agreement. There is also a near-term demand nuance worth flagging — MP management itself, per Bloomberg's May 7, 2026 reporting, has signaled it sees heavy-rare-earth demand potentially falling as motor technology advances and magnet designs use less heavy rare earth. That is an honest complication: the company that should be talking its book on heavy-rare-earth scarcity is instead flagging that technology could reduce it. The robotics demand thesis for MP is real and large in the long run, but it is a structural tailwind, not a 2026 revenue catalyst, and the demand mix within rare earths is itself evolving.

The competitive threat / Lynas, USA Rare Earth and China

MP's competitive picture has a Western front and a Chinese front. On the Western front, the named competitors are Lynas Rare Earths — the largest rare-earth producer outside China, which in March 2026 began producing samarium oxide and is a commercial supplier of both light and heavy rare earths — and USA Rare Earth, which owns the Round Top deposit in Texas, is building a sintered-neo-magnet plant targeting first commercial production around Q1 2026, but likely will not open its mine until 2028. Against these, MP's edge is that it is the only fully integrated, currently operating, mine-to-magnet US producer — Lynas is stronger on heavy rare earths and is Australian; USA Rare Earth is earlier-stage and not yet integrated. There is no IP litigation defining the competitive timeline; this is a capacity-and-integration race.

The far more important competitive front is China. China dominates global rare-earth separation and magnet manufacture, has decades of cost advantage and scale, and demonstrated in 2025 that it will use export controls as policy leverage — which is precisely what created the Western-supply-chain investment thesis and the government backing for MP. But that same Chinese dominance is the structural threat: if Beijing chose to flood the market and crater rare-earth and magnet prices, MP — a higher-cost Western producer still ramping — would be squeezed, and that is exactly why the price-protection agreements, the government funding and the customer prepayments matter so much. They are the contractual and political backstop against a Chinese price war. The competitive reality is that MP wins on Western-supply-chain security and integration; it does not win on cost against China, and its economics depend on a structural willingness — by governments and by customers like Apple and GM — to pay a premium for non-Chinese supply.

The terminal risk

MP's terminal risk has two strands. The first is motor-architecture change: the entire magnet-demand thesis assumes high-performance electric motors keep using NdFeB permanent magnets. There are credible alternative motor designs — induction motors, switched-reluctance motors, externally-excited synchronous motors, and emerging rare-earth-reduced or rare-earth-free architectures — and if robot and EV motor designers move materially toward those to escape rare-earth dependency, NdFeB demand growth slows. The honest, uncomfortable data point here is that MP's own management has flagged that heavy-rare-earth demand could fall as technology advances — an acknowledgment from the most informed party that the demand mix is not fixed. The transition window is long (motor architectures change slowly) and NdFeB is unlikely to be displaced entirely because the performance-density advantage is real, but the magnitude of magnet demand per robot is not a constant.

The second strand is the China price war already described — Beijing using its cost-and-scale dominance to make Western production uneconomic. MP's defense against both is the same backstop stack: government funding and offtake, customer prepayments, and price-protection agreements that insulate the economics from spot-price collapse. The named beneficiaries if the terminal risks materialize are, on the motor-architecture side, the rare-earth-free motor technology developers, and on the price-war side, the Chinese incumbents. The terminal risk constrains the multiple — you cannot pay an unlimited price for a higher-cost producer whose demand thesis depends on both motor architecture and geopolitics holding a particular shape — but it is a slower-burning, more-hedged risk than the socket-absorption or category-absorption risks facing the chip and memory names, because MP's backstop stack is unusually deep and the strategic rationale for Western governments to keep MP viable is unusually strong.

Bull / Gap / Optionality (Photoncap framing)

1. MP is the only fully integrated, operating, mine-to-magnet US producer — a genuine physical bottleneck. Unlike the chip names, where the bottleneck is competitive, the magnet bottleneck is physical and geographic, and MP is the sole listed pure-play that owns the entire Western alternative chain — Mountain Pass mining and separation plus Independence magnet manufacturing (started December 2025).

2. The funding and offtake backstop is exceptionally deep. $72 million of Apple prepayments, GM prepayments, a DoD funding-and-offtake partnership, an $11.0 million Section 48C credit plus 45X production credits — a stack of government and blue-chip-customer capital underwriting the build-out and insulating the economics from a Chinese price war. Few small/mid-caps have a backstop this strong.

3. The financial inflection is real and visible. Q1 2026 revenue up 49% to $90.6 million, adjusted EBITDA swinging $39.3 million year-on-year to $36.6 million, record NdPr output, and a first $21.1 million Magnetics-segment contribution — the integration build-out is converting to numbers, with ~67% full-year 2026 revenue growth expected by the Street.

4. Uniquely in this batch, the tape is calm. RSI 44.3 and just +0.9% above the 50-day moving average — MP has consolidated rather than melted up, which means an entry here carries an actual margin of safety that the extended chip and memory names cannot offer. This is the cleanest entry setup in the compute-and-materials group.

5. The robotics demand thesis, while long-horizon, is structurally enormous. Industry research points to robotics — humanoids specifically — becoming the single largest driver of NdFeB magnet consumption over the long horizon, against an already-$19-billion, high-single-digit-growth magnet market. MP is the Western pure-play positioned for that demand.

Gap

1. Robotics revenue today is zero — it is a demand thesis, not a contracted book. Magnets only started in December 2025, initial magnet revenue is guided for H2 2026, and the disclosed customers are Apple and GM, not robot OEMs. The robotics tailwind is structural and long-dated; it is not a 2026 catalyst.

2. MP is a higher-cost producer that does not win on cost against China. Its economics depend on governments and customers continuing to pay a premium for non-Chinese supply. A Chinese decision to flood the market and crater prices would squeeze a still-ramping Western producer — the backstop stack mitigates this but does not eliminate it.

3. The demand mix within rare earths is itself evolving — per MP's own management. Bloomberg's May 7, 2026 reporting has MP signaling that heavy-rare-earth demand could fall as motor technology advances. When the company itself flags that the scarcity narrative may soften, the magnet-demand thesis deserves a haircut.

4. The valuation already prices a successful ramp. A ~48.6x forward P/E and a ~$10.6 billion cap on a company with a consolidated GAAP net loss requires the Independence and 10X build-outs to execute and the magnet revenue to scale on schedule. Execution slippage on the H2 2026 magnet-revenue guidance would hit the multiple.

Optionality

The trade

MP is a Bucket A name for the robotics theme — and the standout entry setup in this compute-and-materials batch — because it is the only name here that combines a genuine physical bottleneck position, an exceptionally deep government-and-customer funding backstop, a real financial inflection, and a calm, non-extended tape. Initiate in a $57-63 entry zone (current $59.75 ± 5%; the calm RSI 44.3 / +0.9%-above-50MA tape means this is a true current-price entry, not a wishful pullback target), size at 1.5-2.5% of risk capital — the margin of safety the tape provides supports a fuller position than the extended names in the batch warrant — with a stop near $51 (below the 50-day moving average and the recent consolidation floor). The defining near-term binary is Q2 2026 earnings around August 6, 2026, with the H2 2026 initial-magnet-revenue guidance the milestone the whole magnet thesis hangs on. The honest caveat is that MP's robotics exposure is the longest-dated in this batch — zero revenue today, a structural demand thesis rather than a contracted book — so MP is owned for the bottleneck position and the inflection, with robotics as the long-horizon demand tailwind rather than the near-term driver. There is no cleaner listed pure-play for the Western rare-earth-magnet bottleneck — MP is the pure-play — so the only pivot is sizing: this is a core position in a robotics book precisely because, alone in this group, you are not chasing it. Conviction: 7 / 10.

Sources referenced inline throughout. Reference v1 of this template format: _Watchlist/hanmi-photoncap-style.md.

\newpage