America's Most Vulnerable Metal Costs $3

Over the past two weeks I've been building a list.

It has every company I can find outside China that's building a plant to separate heavy rare earths. That's the only way yttrium, the metal I've been writing about all week, ever gets made. There are about a dozen names on it, spread across five continents.

Most of them have never sold a kilogram of separated rare earths. Several don't have the money to build what they've announced, and a couple don't have the ore yet. One of them is already in our portfolio, down 74%.

I'll walk you through all of it in a minute, but first, a number that puts the whole group in perspective.

The $3 Metal

The Colorado School of Mines runs a research shop called the Payne Institute.

In late September it put out a paper with a title that tells you everything: the most vulnerable mineral in America's arsenal costs three dollars a kilogram.

The mineral is samarium.

By the U.S. Geological Survey's count, samarium oxide cost about $2.82 a kilo last year. A USGS model ranked it first out of 84 minerals for the economic damage a one-year supply cutoff would do.

It goes into the heat-resistant magnets inside guided missiles and fighter jets, and China put it under export controls on the same day in April 2025 that it put yttrium under them.

Or take gallium. When two American radar systems in the Gulf were destroyed in the opening days of the Iran war, researchers calculated it would take about 77 kilograms of gallium to rebuild them.

One of those radars costs about $1.1 billion and five to eight years to replace. The gallium is a rounding error on that bill.

That's how it goes with these metals. You need a tiny amount, it costs next to nothing, and the machine doesn't work without it.

A government that needs its jets to fly doesn't haggle over price. It'll pay $9 or $1,150 for the same kilo of yttrium oxide, and it'll pay whoever can deliver.

Which is why I think the money here is in the plants that separate these metals, and in whoever owns one when a government shows up with a checkbook.

The Map

This is what the list looks like today.

In the United States, MP Materials finished building its heavy-rare-earth circuit in California in May and said it would be running later this year.

As of its last update in October it wasn't yet, and it hasn't put a date or a volume on yttrium.

Energy Fuels, the uranium company in Utah, is adding heavy-rare-earth separation at its mill, with commercial output penciled in for late 2027 or early 2028.

No yttrium circuit there either.

It has sold roughly $300 million of stock through its at-the-market program and announced a fresh $500 million one last month.

There's a deposit in West Texas that's about 70% heavy rare earths by content, with a feasibility study due this quarter and yttrium penciled in for around 2029. And there are two separation plants planned for Louisiana.

One redesigned itself on October 6 specifically to add yttrium (smart move). But its owner has about $43 million in the bank, roughly $1.5 billion of projects on the drawing board, and a letter of interest from the Export-Import Bank that isn't a commitment.

The other still hasn't locked up the ore it plans to process.

In Europe, Solvay's plant in La Rochelle, France, can separate all 17 rare earths and is adding the heavy ones this fall.

A new plant in Lacq, in the south of France, is due to start up by the end of this year, and Japan's government helped pay for it in exchange for a long-term supply deal.

And a Canadian company runs a small heavy-rare-earth line in Estonia that was just picked for a €126 million European grant to expand.

Then there are two companies that have both a date and the money.

One of them is already selling separated heavy rare earths at a profit, which nobody else on this list can say.

More on those two in a minute.

Three Questions I Asked Before Picking One

Last month, in my Why Critical Minerals Got Sold Like AI essay, I gave you three questions for any critical-minerals stock: does it produce today, is it funded, does it have customers under contract.

For this list I had to get more specific.

1. Did it sell anything last quarter? How many tons of separated rare earths, and is that number in the financial statements or on a slide? If it only lives on a slide, it's a plan.

2. Who's paying for the plant? Its own cash flow, a government loan with a signed date, or the next stock sale? The companies on my list that fell hardest this year are the ones in that third group.

3. Who's buying what comes out? A government buyer with a floor price, or a signed long-term supply contract (what the industry calls an offtake), is worth more than any press release. It means somebody has already decided this plant needs to exist.

Run the names above through those three and there aren't many left.

The Position That Changed the Next Pick

We own two companies on that list.

Our first rare-earth pick went out to Premium Members in August 2025. Seven weeks later it was up 140%, and we sold enough to pull our entire stake out. It has fallen a long way since, but every share we still hold is house money.

The second went out in October 2025, and it's down 74%. By my own test, it's a plan.

The first machine isn't scheduled to be commissioned until the first half of next year, and there's no date on the full plant.

The ore isn't locked up, and the company says its balance sheet will need topping up within two years.

That position is why the next pick in this group is a different kind of company.

One Move: Decide the Size Before You Know the Name

Next week, Premium Members get a name. Before you know what it is, decide how much a story like this deserves in your account.

Here's how I think about it. A company that sells metal today, with cash in the bank, in a sector that's down 40% to 70% from its highs, is a core position in this theme.

For me that means 1% to 2% of the money you've set aside for speculative ideas, bought in two pieces. A developer with a plant still on the drawing board is a lottery ticket, so half that, and you don't add to it on hope.

Write your number down now. It's a lot easier to size a position before you've fallen for the story.

By the way, next Tuesday China's one-year pause on its broader rare-earth export controls expires on paper. It doesn't touch yttrium, which was never paused.

So don't buy anything because of a date on a calendar. If you buy, buy because of what the company sells and what you're paying for it.

What's Coming Next Week

The essays are free and they'll stay free. Next week's alert isn't, and here's the record behind it.

14 closed trades. 14 winners. Average gain 83.5%.

Over the same stretches, the S&P 500 did 37.75%.

Here's what that record doesn't show. Four of our open small-cap metals positions are down more than half. That's the price of buying plans, and it's why we keep them small enough that being wrong hurts without doing real damage.

Premium Members see those in the same portfolio as the winners.

Next week's alert is a different kind of pick. I can tell you this much without giving it away:

• It already sells separated heavy rare earths, and has since last year.

• It has more cash than debt, and no convertible notes hanging over the stock.

• Governments on two continents are on its customer list.

• It's down almost half from its high.

• It isn't cheap on this year's earnings. I'll show you exactly what you're paying, and where I'd set the watch price.

The same alert closes one position, and Premium Members get the name and the reason. It also names a second company I'm deliberately not buying yet, along with the one event that would change that.

If you're reading this for free, here's how next week goes… 

The alert lands with Premium Members. Sometime after that, I'll write about the decision in a free essay, and by then the price will be whatever it is. That's how free readers heard about our shipping trade this week, after we'd sold it at $10.10.

Premium is $35 a month. Join before the alert goes out and you'll have it the morning it does.

Double D

🔓 Premium Content Begins Here 🔒