Mountain Pass tells an unusual story.

The rare-earth mine in the California desert opened in 1952 and supplied much of the world's demand through part of the Cold War.

By 2002, it was defunct.

Then the strategic environment changed.

Investors formed Molycorp in 2008 to revive it. China restricted rare-earth exports to Japan in 2010. Prices surged.

The timing appeared almost perfect.

Molycorp raised capital, expanded and acquired additional processing assets.

Then revenues collapsed.

The company filed for bankruptcy in 2015.

Two years later, the mine sold for $20.5 million to a consortium that included a Chinese-owned company.

For a period, material extracted from the California desert was shipped to China for processing.

Across those decades, something important never changed.

The rock.

The minerals remained in the same mountain.

What changed was the price.

And that gives us the organising idea for Article 5 of Series II.

China accounted for roughly 61% of global rare-earth mining in 2024.

That is substantial concentration.

But its share of refining and processing was around 91%.

Those are not two versions of the same statistic.

They describe two different systems.

And the second is much more difficult to replace.

Rare earths are not particularly rare.

Gallium can be recovered as a by-product of aluminium refining.

Graphite is mined in multiple countries.

The geological resource exists elsewhere.

The concentration lies increasingly downstream:

separation;

refining;

specialised graphite products;

magnet manufacturing.

These industrial processes were allowed to concentrate over several decades because moving them elsewhere often made economic sense at the company level.

They are capital intensive.

Some are chemically difficult.

They create environmental costs.

Margins can be thin.

Buying the processed material from somewhere cheaper looked rational.

Repeated thousands of times across three decades, those rational decisions produced a strategic dependency.

A mine can exist outside China while the supply chain remains dependent on China.

Mountain Pass demonstrated this directly.

Material extracted in the United States still required processing elsewhere.

That is why counting mines produces a misleading picture of resilience.

The relevant question isn't:

Where is the ore?

It is:

Where is the industrial process that turns the ore into something the next layer can use?

And then comes the more difficult question:

Can that process survive commercially?

A new separation facility can take roughly three to seven years to build.

For the first three articles of Series II, a number like that would have been our substitution time.

Not here.

Imagine an investor evaluating a new separation plant.

The current ex-China price is attractive.

Demand exists.

Governments want diversification.

The technology is available.

So the project gets financed.

Several years later, it begins operating.

But the incumbent still controls the overwhelming majority of global capacity.

If prices fall sharply, the new facility's economics deteriorate.

The incumbent doesn't even have to cut prices during construction.

The possibility that it could do so afterwards already affects the project's financing.

This is the unusual feature of the fifth row:

the alternative's future revenue is influenced by the party it is intended to replace.

So the true substitution time isn't three to seven years.

Three to seven years tells us when the plant can exist physically.

It doesn't tell us whether the plant will still exist economically ten years later.

The split is already visible.

In September 2026, dysprosium oxide was around $212–218/kg inside China.

North American assessments reached as high as roughly $3,250/kg.

Terbium oxide showed another very large divergence.

The ex-China market is thin, and those prices should therefore be read as estimates and orders of magnitude rather than perfectly comparable clearing prices.

But the underlying message remains.

The market is separating into two.

Inside one system, processing capacity is abundant.

Outside it, scarcity commands a substantial premium.

Normally, a price difference that large would solve itself.

Capital would enter.

Capacity would expand.

Competition would reduce the premium.

But that process assumes the new entrant can survive until the market normalises.

Mountain Pass shows why that assumption cannot simply be made.

This is why the latest US intervention around Mountain Pass deserves attention beyond the individual company involved.

The Department of Defense structure includes a ten-year $110/kg price floor for NdPr, a contract-for-difference mechanism in which the government participates in upside, $400 million in preferred equity, a warrant and guaranteed offtake for expanded magnet capacity.

It is not simply paying to construct another factory.

It is attempting to guarantee that the factory survives.

That distinction is fundamental.

Molycorp managed to raise money.

Capital alone was not enough.

The new mechanism targets revenue certainty.

If market prices fall below the floor, the project can continue operating.

If demand is uncertain, guaranteed offtake removes part of that risk as well.

Whether this specific arrangement works remains an open question.

But the mechanism identifies the constraint more precisely than a conventional construction subsidy.

And if it works, it is unlikely to remain confined to critical minerals.

There is another dimension to the dependency.

Export controls are normally discussed in terms of quantity.

Material is available.

Then a restriction is introduced.

Less material crosses the border.

But licensing systems can also require applicants to provide detailed information about the intended use of controlled materials.

What product?

Which customer?

What quantity?

What schedule?

Viewed purely as trade administration, this is compliance.

Viewed as an information system, it is something more significant.

Over time, those applications can produce a detailed map of dependency across foreign industrial systems.

Who needs which material?

For which product?

In what quantity?

How regularly?

That information potentially enables a much more precise form of restriction.

A blanket embargo affects everyone.

That creates a large coalition with an incentive to respond.

Selective denial can affect one company, programme or sector while leaving everyone else supplied.

The wider market then sees a company-specific problem rather than a systemic one.

That distinction matters.

And it means restored export volumes do not necessarily mean restored resilience.

A valve that reopens is still a valve.

The architecture remains.

And so does the information collected through it.

For businesses manufacturing physical products, the first correction is straightforward:

trace critical materials to the refining step, not just to the supplier.

A procurement system may show two independent component suppliers.

But if both ultimately depend on the same separation facility, the underlying supply chain still has one source.

The second question concerns information.

Companies should understand what end-use information has already been disclosed through licence applications, either directly or through suppliers.

That information is usually treated as a procurement or compliance matter.

It should also be reviewed as a map of industrial exposure.

Third, engineering substitution should move onto the roadmap before the shortage becomes acute.

Reduced-heavy-rare-earth magnet chemistries, alternative motor architectures and recycling from end-of-life products become more attractive when ex-China prices carry a large structural premium.

Procurement can buy another batch.

Engineering can sometimes remove part of the dependency entirely.

Those are different levels of resilience.

For most households, this is not primarily a shortage story.

Rare-earth magnets and other specialty materials often represent a relatively small fraction of the total cost of a finished product.

A severe price increase in the material does not automatically mean cars, appliances or electronics disappear from shops.

The effect is more likely to arrive through:

slightly higher prices;

delayed product launches;

changes in engineering;

or production problems affecting particular models.

The consumer impact can therefore remain modest even while the industrial significance is very high.

That distinction is important.

A substrate can be strategically critical without creating an obvious consumer crisis.

The most important distinction is between capital availability and revenue durability.

A project can be fully financed and still fail.

For new non-Chinese refining capacity, examine the structure supporting the project's revenues.

Is there a price floor?

For how long?

Is it contractual, legislative or administrative?

Does it survive a change in government?

Is offtake guaranteed?

How much output is covered?

Those questions may tell us more about whether a facility will still be operating in ten years than the quality of the mineral deposit underneath it.

This is an unusual industry in which the boring contract may matter more than the spectacular resource estimate.

Our one-year forecast, resolving at the end of 2027:

70% — ex-China dysprosium or terbium oxide remains at least three times the corresponding Chinese domestic price.

Our three-year forecast, resolving at the end of 2029:

65% — at least one additional G7 government establishes a price floor, contract-for-difference or guaranteed-offtake mechanism for a domestically refined critical mineral.

Our five-year forecast:

70% — China's share of global rare-earth separation capacity remains above 75% at the end of 2031.

And the deliberately uncomfortable ten-year forecast:

60% — at least one non-Chinese rare-earth separation or magnet facility commissioned between 2025 and 2031 ceases operations or enters insolvency by the end of 2036.

The forecasts are designed to test the mechanism, not simply produce a directional narrative.

If alternative capacity survives without revenue guarantees, the central thesis weakens.

If design substitution reduces dependency faster than new supply can be built, the importance of the refining chokepoint itself declines.

Those are outcomes we should want the framework to recognise. TIC_S2A05_Website_Full

Series II now contains five physical layers:

Electricity.
Frontier computing.
Maritime chokepoints.
Water.
Refined materials.

And something interesting has happened.

Twice, substitution time has refused to become a simple engineering number.

With water, the alternative may exist physically but already belong to another user.

Access therefore depends partly on allocation.

With critical materials, the alternative can be built physically but its economic survival may depend on prices influenced by the incumbent.

In both cases, the same limitation appears from a different direction.

Concentration is a property of the physical world.

Criticality is a property of the physical world.

But substitution time isn't always.

Sometimes it is a property of the relationship between actors.

That means our analytical instrument has to evolve.

Knowing how many tonnes exist is insufficient.

Knowing how many factories exist is insufficient.

Knowing construction time is insufficient.

We also need to know:

Who controls access to the alternative?

Who controls its economics?

And can they prevent it from becoming durable?

For critical materials, those questions explain more than the geology ever could.

The material is abundant.

The viable alternative is not.

Next in Series II:

Connectivity — the sixth row.

THRIVE IN CHAOS
Decision Intelligence for an Uncertain World

Analysis → Forecast → Recommendations
Signal → Meaning → Action → Stability

Forecasts are probability-based analytical assessments, not certainties. This material supports independent judgment and does not constitute investment, legal or tax advice.