The Rare Earth Supply Gap Investors Keep Mispricing

The consensus reading on rare earths right now is that softening electric vehicle sentiment has punctured the investment thesis. The maths says otherwise.

To meet projected magnet demand in 2035, the world would need to build entirely new refining and separation capacity that simply does not exist today, on a timeline that leaves almost no margin for delay.

As of September 2026, the landscape outside China remains startlingly thin. Dozens of mining announcements have crossed the wires over the past two years, yet only two non-Chinese companies produce rare earth oxide at commercial scale. Between them and the projected demand sits a supply gap that institutional forecasters put at 60,000 to 70,000 tonnes or more over the next decade.

That gap is the entire story, and it is also where investors get the story wrong.

This piece lays out a framework for evaluating Western rare earth equities: how to separate upstream mining noise from the downstream refining realities that actually determine viability, and how sovereign price protections change the risk calculation on the equities worth owning.

The refining bottleneck making mining announcements irrelevant

A new mine does not equal new supply. That is the single most expensive misunderstanding in this sector, and it explains why so many junior explorers trade on hope rather than cash flow.

The refining vs. mining value chain distinction is the core capital allocation error in this sector: junior miners with in-ground resource estimates but no credible separation pathway are priced as if downstream processing costs and timelines do not exist.

Rare earth ore coming out of the ground is close to worthless in its raw state. The value sits downstream, in the separation and refining that converts mixed concentrate into neodymium-praseodymium (NdPr) oxide, the specific material that goes into the high-strength permanent magnets used in electric vehicle motors and wind turbines. Separation is chemically brutal, capital-intensive, and slow to commission. It is a different business entirely from digging.

Here is where the concentration becomes stark. China controls roughly 60% of global rare earth mining, but its grip on the midstream is far tighter: an estimated 90% to 94% of the world’s separation and refining capacity sits inside China. It took the country roughly three decades to build that position.

The Congressional Research Service rare earth analysis published in mid-2026 puts China’s share of separation and processing capacity at 90%, a figure consistent with the three-decade consolidation of midstream infrastructure that Western policy is now scrambling to replicate.

Outside China, the operational picture reduces to two names. MP Materials, running Mountain Pass in the United States, exited 2025 with an effective run-rate near 4,000 tonnes per year of NdPr oxide after doubling output year-on-year. Lynas Rare Earths, spanning Australia and Malaysia, reported FY26 NdPr production of 7,260 tonnes, up 11% on the prior year. That is the entire Western oxide production base.

The gap widens further downstream. According to International Energy Agency analysis, announced refining capacity outside China is on track to reach only about two-thirds of expected mine output by 2035, and announced magnet production capacity only around one-third.

What this tells you is direct: an early-stage explorer without a funded, credible pathway to refined oxide is not a rare earth supply story, it is a mining lottery ticket. Value the midstream refining capability, not the raw tonnes in the ground.

How electric vehicles and robotics compound the 2035 deficit

The demand side is where the numbers stop being incremental and start being structural. This is not a market forecast to grow a few points a year and strain gently at the edges.

Adamas Intelligence projects that global demand for NdFeB magnets will grow at a compound annual rate of 8.6% through to 2035. The resulting shortfall in NdPr oxide equivalents is forecast to widen from 21,000 tonnes per year by 2030 to 68,000 tonnes per year by 2035.

EV permanent magnet demand is the dominant load driver in every major scenario model, and the physics of why NdFeB magnets are difficult to displace in traction motors, not just the economics, is what gives the structural deficit its durability even as substitution research accelerates.

The value figure is more arresting still. Consumption of NdPr oxide for energy-transition applications is forecast to rise 11-fold by 2035, implying a roughly 1,000% volume increase relative to current levels.

Where does that demand come from? Ranked by weight in scenario modelling from SparkCo:

1. Electric vehicles, accounting for roughly 60% of modelled demand. EV magnet consumption alone is forecast to triple by 2035 as adoption resumes its upward path.
2. Wind turbines, at around 20%, driven by direct-drive turbine designs that rely heavily on permanent magnets.
3. Defence applications, at approximately 10%, spanning guidance systems, actuators, and other high-reliability hardware.
4. Humanoid robotics, the wild card. Institutional forecasts do not yet disaggregate robotics as a quantified segment, but each humanoid unit is expected to consume a rare earth load comparable to an EV. Initial production is anticipated to begin next summer from a converted Tesla facility, scaling the following year.

The robotics layer sits on top of an automotive baseline that is already stretched. Even if robotics disappoints entirely, the EV and wind numbers alone describe a deficit.

A 1,000% demand increase is not a call for marginal supply improvement. It tells you the opportunity, and the requirement, is entire new Western supply chains that currently do not exist. Size the thesis to that reality, not to a quarterly production beat.

Sovereign floor prices and the escape from Chinese benchmarks

Supply volume is only half the equation. The other half is price, and price in this market has historically been set in a room the West does not sit in.

The dominant reference has long been the Asian Metals Index, quoted ex-works China and heavily shaped by Chinese domestic pricing. As of September 2026, that spot price sits in a wide band: the Shanghai Metals Market assessed NdPr near US$96 per kg on 8 September 2026, while the Association of China Rare Earth Industry reported a comparable domestic level of US$107 to US$109 per kg in mid-August 2026.

The Seaborne Index reality check

The response has been an attempt to build an independent, ex-China pricing mechanism, most notably the Seaborne Index developed by Benchmark Mineral Intelligence, intended to price rare earths free of Chinese domestic influence.

The ambition is sound. The traction is not there yet. Emerging ex-China indices are referenced in some term sheets, Arafura Rare Earths has cited “global seaborne index prices”, but market observers still regard these benchmarks as nascent. No accessible public source quantifies a live, authoritative spread between them and the Asian Metals Index, which remains the de facto reference embedded in major US strategic contracts.

So allied governments have taken a more direct route. Rather than waiting for an independent index to mature, they have bypassed index volatility altogether by writing price floors directly into offtake agreements.

The Japan-Lynas JARE agreement, backed by JOGMEC and Sojitz and extended through 2038, sets a US$110 per kg floor on 5,000 tonnes per annum of NdPr, with a clause handing Japan 30% of revenue above US$150 per kg. The US Department of Defense struck a comparable structure with MP Materials: a US$110 per kg NdPr floor across a 10-year offtake covering the full magnet output of a planned US facility, again with a 30% upside-share. A separate 15-year DoD agreement with Serra Verde embeds the same US$110 per kg floor for its Phase I magnet rare earths.

Government-backed floor prices convert a project’s most dangerous unknown, future selling price, into a defined revenue baseline. For an equity, that is the difference between a speculative bet on commodity direction and a de-risked cash-flow profile that banks can actually finance against.

The read for your holdings is concrete. A US$110 per kg sovereign floor gives strategically aligned Western producers a revenue backstop that structurally insulates them against aggressive Chinese price undercutting. That protection is a real, quantifiable risk mitigant, though analysts at the Columbia Center on Global Energy Policy caution that scaling it beyond a handful of state-backed deals would demand significant public spending and difficult allied coordination.

Market flooding, cyclical gluts, and motor substitution risks

Every structural scarcity thesis attracts believers who forget that scarcity can be engineered away or flooded out. This one is no exception, and the bear case has teeth.

Three downside catalysts deserve genuine attention:

• Market flooding. China has demonstrated its willingness to let prices fall hard. NdPr oxide slid to roughly US$63 per kg in March 2025 amid a severe supply glut, and S&P Global and Benchmark commentary flagged ex-China oversupply running into 2026, with prices slipping below the US$110 per kg floor level. Cyclical surpluses can appear even inside a long-term structural deficit.
• Technology substitution. Automotive OEMs are already deploying and assessing motor designs that reduce or eliminate rare earth content, including permanent-magnet-assisted synchronous reluctance motors using low-cost ferrite, and electrically-excited synchronous machines. Older thinking held that alternatives only made economic sense above roughly US$1,000 per kg; OEMs are moving far earlier to sidestep geopolitical risk.
• Demand destruction. Battery supply-chain bottlenecks, slower EV uptake, policy rollbacks, or delayed robotics scale-up could each pull demand below aggressive baselines.

There is a counterweight worth holding. The cost premium for non-Chinese NdPr is estimated near US$50 per kg, a rounding error against an average EV sale price of roughly US$46,000. That thin premium argues against demand fleeing on cost alone.

But the substitution push is the signal that matters most. It tells you demand is not perfectly inelastic. Watch automotive motor architecture decisions as closely as you watch mine output, because a design choice made in a manufacturer’s engineering department can erode the thesis faster than any new mine can supply it. It is also worth separating NdPr from the heavy rare earths dysprosium and terbium, which command far larger ex-China premiums and face fewer substitution routes.

Heavy rare earth premiums for dysprosium and terbium reflect a structurally tighter supply picture than NdPr faces: both elements are critical for high-temperature magnet performance in EV and defence applications, both are produced in even more geographically concentrated deposits, and both have fewer credible substitution routes than the NdPr literature typically acknowledges.

Valuing Western NdPr producers in a structurally constrained decade

The core tension is unresolved by design. A genuine 68,000-tonne structural deficit by 2035 sits alongside genuine risks of Chinese flooding, cyclical gluts, and engineered substitution. Both are true at once, and any thesis that ignores either half is incomplete.

The framework that survives this tension is consistent throughout. Raw mining assets hold little standalone value without a funded, credible route to oxide separation and refining. That is the first filter, and it eliminates most of the field.

Over the next 12 to 18 months, the milestones that will arbitrate equity value are Western refining capacity coming online and the evolution of sovereign offtake structures. Those are the events to track, not exploration drill results.

For capital allocation, the read is clear: prioritise producers with state-backed price protections, because a US$110 per kg floor is what converts a commodity gamble into a financeable business through a volatile decade.

For readers wanting a structured assessment of how export controls, processing chokepoints, and allied policy failures interact, our dedicated guide to rare earth supply chain security maps the specific vulnerability categories that institutional investors and strategic buyers use to stress-test Western supply narratives.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions.

Past performance does not guarantee future results. Financial projections are subject to market conditions and various risk factors, and forward-looking statements are speculative and subject to change based on market developments and company performance.