The Mining Stock Where the Pentagon Is the Largest Shareholder: MP Materials and the Geopolitics of Magnets

TL;DR Every high-torque motor in a robot runs on a permanent magnet, and China controls close to 90% of the world's magnet processing. The Department of Defense became MP Materials' (MP) largest shareholder, locked in a 10-year price floor on its key material, and agreed to buy the entire output of its new plant. Revenue climbed from under $70 million to about $275 million, but the company still loses money and has burned more than $300 million in cash. You're not buying earnings here. You're buying a factory targeted for 2028.

The simplest part in the robot turned out to be the most dangerous one

Every robotics conversation starts with the brain. But when you take the machine apart by cost, the riskiest component isn't a chip. It's a magnet.

Every high-torque motor in a humanoid, every joint, and every electric vehicle depends on permanent magnets. It's the most ordinary-looking line item on the bill of materials, and without it the machine simply does not move.

China controls roughly 90% of the processing for that ordinary part. As I argued in The 7 Bottlenecks of the AI Supply Chain, pricing power always shows up at the narrow points.

The point where cost triples

One number explains the whole problem. Build a humanoid without Chinese parts and the cost goes from roughly $46,000 to $131,000. Close to three times.

That isn't just an expense issue. At triple the cost, commercial mass adoption doesn't happen, and defense robots or autonomous systems can't be built on a domestic supply chain at all. It stops being a problem with one machine and becomes a problem with an entire industry.

Which is why the United States stopped waiting for the market to solve it and stepped in directly.

The day the Pentagon bought a stake

MP Materials (MP) is the only American company that owns the entire rare earth magnet chain under one roof, from a mine in California all the way to a finished magnet.

The Department of Defense came in, and not with a small grant. It bought a stake big enough to become MP's largest shareholder. On top of that, it guaranteed the company a floor price on its key material for 10 years and agreed to buy the entire output of its new plant.

That's the moment I changed the frame I use for this company. A mining stock normally lives or dies on commodity price cycles. But a company with a decade-long price floor underneath it and its demand locked in by contract looks less like a commodity bet and more like an infrastructure project. When the Pentagon co-signs your business plan, you're not really just a mining stock anymore.

Commercial demand showed up too. Apple has already signed a deal worth around $500 million for American-made magnets.

The income statement still hurts

Time to be honest. This is not a stock you buy on earnings, because there really aren't any yet.

Revenue has climbed from under $70 million to about $275 million. As a growth rate, that's impressive. But the company still loses money running the business and has burned through more than $300 million in cash building out infrastructure.

I don't read that as failure. Building a chain from mine to finished magnet from scratch swallows cash for years by nature. It just means the question an investor should be asking changes from "how much is it earning now" to "when does the plant run, and at what scale."

From 1,000 tons to 10,000

So the number that matters isn't on the income statement. It's capacity.

The existing Texas plant makes about 1,000 tons of magnets a year. The one going up right beside it is built for 10,000. That's a tenfold jump, with production targeted for 2028.

Item	Today	New plant (2028 target)
Magnet capacity	~1,000 tons/year	~10,000 tons/year
Selling price	Market	Floor guaranteed for 10 years
Demand	Individual contracts	Government agreed to buy all of it

Laid out like that, the nature of the investment gets clear. This isn't a bet on whether rare earth prices go up. It's a bet on whether that plant runs on schedule in 2028. A much narrower question, and a much more verifiable one.

After the magnets comes assembly

Once materials open up, the next bottleneck moves to who is willing to assemble millions of these machines. The early leaders are already lining up.

Jabil has become the worldwide production partner for Apptronik's Apollo humanoid. The car world is pushing in too, with auto parts giant Magna taking a stake in robot maker Sanctuary AI and opening its factories.

Companies that already know how to build complex hardware by the million tend to win the mass production phase, and these two are planting flags early. It's structurally the same picks-and-shovels logic I applied to drones in Picks and Shovels in the Drone Supply Chain.

My take: this isn't an earnings stock

I file MP under speculative swings, and I mean that both as praise and as a warning.

The good part: a materials company with both its price and its demand underwritten by the government is rare. America bet on this company, and this company bet on America.

The warning: execution risk runs all the way to 2028. Plants get delayed, budgets get overrun, yields come in below plan. That's the base case, not the bear case. And through all of it the company keeps burning cash. This is why you shouldn't hold it like stable dividend infrastructure.

Small size, long horizon. And each quarter, the thing to check isn't EPS, it's construction progress on the new plant.

FAQ

Q: Isn't MP Materials just a mining stock? A: It looks like one, but the economics are different. A typical miner's revenue tracks the spot market. MP has a 10-year price floor and an agreement covering the entire output of its new plant. That makes it closer to a build-and-commission bet than a commodity price bet.

Q: Why is the stock up if the company loses money? A: The market is pricing 2028 capacity, not current profit. The tenfold jump from 1,000 to 10,000 tons plus the government and Apple contracts are the basis. The flip side is that schedule slippage is exactly what hits the stock hardest.

Q: Why do magnets matter so much in robotics? A: Every joint in a humanoid carries a high-torque motor, and every one of those motors contains a permanent magnet. If magnet supply gets cut, it doesn't matter how good the brain is. The machine doesn't move.

Q: Is China really at 90%? A: On processing, yes, not on mining. Rare earths aren't actually rare in the ground, but separation, refining, and magnet-making are extremely concentrated. That's precisely why the US put money into the processing chain rather than into a mine.

This is educational analysis, not financial advice.