How long can China play the “rare earths card”?

Arnaud Bertrand

This is probably the most important geopolitical question in the world right now: for how long can China play the “rare earths card”? It’s now well established this gives China considerable leverage. For one thing the frantic state of panic of US Treasury Secretary Bessent is a pretty big tell: he publicly insulted senior Chinese officials over the move, lobbied for “emergency powers” and said this was a Chinese attack on the “world” that would meet a “fulsome group response” from the US and its allies. If that’s not Washington being rattled, I don’t know what is.What seems to be the consensus view, because I’ve seen it mentioned over and over again, is that one of the main bottlenecks to break this rare earths stranglehold is environmental regulations. As the narrative goes, the West essentially regulated itself out of the rare earths business by imposing environmental standards that China simply ignored. And so, by implication, all it would take is the right regulatory changes and government subsidies and the problem is solvable within a few years or so; it’s mainly a question of political will to accept environmental trade-offs.

There is some degree of truth in that – rare earths processing can be very polluting – but this is otherwise very much magic bullet thinking. The difficulty of breaking the rare earths stranglehold is far more intense than mere regulatory adjustments. China’s dominance has much more to do with the scale of their manufacturing and the vertical integration of their supply chains, and as such breaking the stranglehold at this stage requires upgrading the West’s industrialisation level comprehensively. We’re talking about something requiring a complete makeover of the West’s socioeconomic structure, involving trillions in capital in investment – with profitability perhaps two decades away – as well as a profound upending of its education system. In short, a generational-level undertaking on an almost unprecedented scale.You might be tempted to compare the efforts needed to the Manhattan Project or the Apollo Program – that’s mighty enough, right? – but that would actually be vastly understating it. The amount of effort needed is more comparable to something like the Industrial Revolution itself than to any individual megaproject.

You don’t believe me, right? Surely I must be exaggerating! No way it can be that dramatic! That’s why I wrote this article. To show you in detail the absolutely titanic efforts that would be needed to break the stranglehold for just one of the elements on China’s list of export controls: gallium. And bear in mind when you read the article that it’s just ONE chemical element out of 21 under export controls, and that China’s export controls don’t only include chemical elements but also downstream products (lithium-ion batteries, superhard materials, etc.).

After finishing this article, I bet Bessent’s panic will feel almost understated to you.

What is gallium?

Gallium is not actually a rare earth: it’s a soft, silvery metal that would literally melt in your hand on a warm day. Yet it’s one of the most strategically important materials in the world today, as it’s – among other applications – foundational to the latest generation of GaN semiconductors, as well as modern AESA military radars that can detect targets at nearly double the previous range. A top Raytheon executive noted in 2023: “GaN is foundational to nearly all the cutting-edge defense technology that we produce.” China has cornered a staggering 98 percent of the worldwide primary low-purity gallium production, meaning it has near-total control over the material.

Let’s ask ourselves a simple question: what would it take to produce 100 tons of gallium? It’s not a huge amount. China produces 600 tons of it, with a production capacity of 750 tons so we’re talking less than 17 percent of China’s current production.

Many people imagine gallium extraction works like mining any other metal: find a deposit, dig it up, add some chemicals, extract the metal. But gallium is fundamentally different – it’s not found as an independent ore but is recovered as a byproduct of aluminum production. Think of it like juicing oranges: gallium is like the small amount of essential oil that clings to the orange peel. Without the juice factory processing massive quantities of oranges, you have no practical way to obtain that essential oil separately. You can’t just ‘mine gallium’ – you need an entire aluminum industry running at scale to capture the trace amounts that emerge.

To understand the scale involved, consider China Aluminum Corporation (Chalco), the world’s largest aluminum producer: in 2022, they processed 17.64 million tons of alumina from which they refined 6.88 million tons of primary aluminum and finally extracted 146 tons of gallium – a ratio of approximately 1:47,000 for gallium-to-aluminum, or 1:120,000 for gallium-to-alumina.

Building alumina refineries and aluminum smelters

The ratios we just saw mean that, to produce 100 tons of gallium, you would first require a proportionate aluminum industry capable of producing 12 million tons of alumina and 4.7 million tons of actual aluminum annually. That’s your first step. For reference China today has 60 percent market share of global aluminum production, India is a very distant second with only 3.5 million tons of aluminum (refined from alumina) produced in 2022-2023 (meaning the whole country produced only half the amount produced by Chalco, a single Chinese company) and the US produced less than 0.8 million in 2023.

So if the US wanted to become a big player in gallium, it’d first need to increase its aluminum production capacity almost six fold, from the current 0.8 million tons to the 4.7 million tons needed to produce 100 tons of gallium, which again would only make its gallium production less than a fifth that of China.This involves building two types of factories: alumina refineries (which process bauxite ore into alumina) and aluminum smelters (which convert alumina into metallic aluminum through electrolysis – the stage where gallium is extracted).

Outside China, aluminum smelters cost about $4 billion per million tons of annual production, meaning we’re speaking about a $20 billion investment just for the smelters. Alumina refineries would add another $10 billion. So we’re looking at $30 billion in factory construction costs just to ramp up alumina production to the level required.

The energy challenge

There’s an issue however: converting alumina into metallic aluminum through electrolysis is extremely energy-intensive. Industry data shows that producing one ton of electrolytic aluminum consumes approximately 13,000-15,000 kWh of electricity.The US currently produces 0.8 million tons of aluminum, so it would need to add another 3.9 million tons of capacity. How much electricity does that require? Using the lowball figure of 13,000 kWh per ton, it translates to roughly 51 billion kWh of additional electricity – flowing continuously, 24/7, 365 days a year. Aluminum smelters can’t simply shut down when power is unavailable; the molten metal would solidify in the electrolytic cells, destroying them.

What does 51 billion kWh mean? To put it in perspective, let’s look at America’s most recent nuclear project: Vogtle Units 3 and 4 in Georgia. These two reactors have a combined capacity of 2.2 GW and can produce approximately 17-18 billion kWh annually at full capacity. The US would need to replicate the entire Vogtle 3 & 4 project three times to meet the 51 billion kWh requirement – essentially building six new reactors in three separate construction projects. Cost wise, Vogtle 3 & 4 reached a final price tag of $36.8 billion after massive overruns from an initial $14 billion estimate. Three such projects would cost approximately $110 billion – and that’s before the $30 billion needed for the aluminum refineries and smelters themselves. Total infrastructure investment: $140 billion.

Timeline wise, construction on Vogtle 3 & 4 began in 2013, with Unit 4 finally entering commercial operation in April 2024 – nearly 11 years. Even with lessons learned and parallel construction (itself questionable given the shortage of qualified nuclear contractors and specialised equipment), a realistic timeline for three new Vogtle-scale projects extends to 2035-2036 at the very earliest. And remember, again, that this $140 billion investment and 12-year timeline would yield just 100 tons of gallium annually – representing only 17 percent of China’s current production and less than 14 percent of their production capacity, which again is only ONE of the 21 chemical elements China applied export controls on.

The human challenge

Building the facilities is only half the battle, the greater challenge lies in finding the people to run them. US manufacturing employment peaked at 19.6 million in 1979 but has declined to approximately 12.9 million by late 2024 – a loss of nearly seven million jobs over 45 years. This isn’t merely about numbers, it also represents a fundamental erosion of the skilled industrial workforce. And the challenge is that aluminum processing is a very worker-intensive industry. The reason is because aluminum cells are dynamic systems where conditions vary cell-to-cell and hour-to-hour, with operators making hundreds of small adjustments daily based on visual inspection, sound, and instrument readings – the kind of complex judgment calls that remain difficult to automate.

You just need to check the numbers in China, the country with the most advanced facilities and access to the latest automation technology: it still employ tens of thousands of workers for aluminum production. Chalco, which we spoke about earlier, employs 58,009 people to produce their 6.88 million tons of aluminum. China Hongqiao, the second-largest aluminum producer in the country (after Chalco), employs 49,774 people and produces approximately six million tons of aluminum annually.

So we’re talking ratios of about 8,500 people per annual ton of aluminum, in the most advanced facilities in the world, with Chinese working hours and efficiency. Meaning that to add another 3.9 million tons of capacity, the US would need to find at the very least 33,000 additional workers just for aluminum production. With all that entails: training skilled aluminum operators requires years of hands-on experience with high-temperature industrial processes, metallurgy, and complex equipment – not skills acquired through short courses.

And I’m not even speaking about the workers needed for the energy part: 800 permanent jobs were created specifically for the new Units 3 & 4 at the Vogtle nuclear station. Three Vogtle-scale projects would require approximately 2,400 additional nuclear operations workers – engineers, control room operators, maintenance technicians, and security personnel. Exceedingly difficult to do in a country where the manufacturing sector already faces 1.9 million unfilled jobs by 2033, and where a significant chunk of the existing nuclear workforce is likely to retire over the next decade. The US would need to spend years to train 35,500 specialised industrial workers for this single gallium project – representing 17 percent of China’s production capacity for one element – while simultaneously backfilling retirements.

The industrial ecosystem challenge

It’s not just factories, energy and people – you need a complete industrial ecosystem. Even if you have money to build factories, technology to build power plants, and the ability to find tens of thousands of workers, there’s an even more difficult problem: supporting facilities. Industrial production is not an island; it requires a complete ecosystem. For example, producing alumina requires bauxite, lime, and soda ash. The US doesn’t lack lime and soda ash, but bauxite mainly needs to be imported. You need stable bauxite supply channels and ports for transportation. Producing electrolytic aluminum requires auxiliary materials like fluoride salts and carbon anodes – factories must produce these too. You also need highways and railways to transport them to the factory area. Once products are made, they need to be transported to ports for export or to downstream chip factories and radar factories – this requires a very mature logistics network. These supporting facilities aren’t as simple as building a few bridges or paving a few roads. They represent a nation’s industrialisation level.

China spent 40 years building the world’s most complete industrial system from scratch. From bauxite mining to alumina and electrolytic aluminum production, to gallium extraction and purification, even downstream chip manufacturing – every link has mature enterprises and supporting infrastructure. This gap in the industrial ecosystem can’t be filled just by throwing money at it. It requires time, it requires accumulation over generations, it requires the entire nation to highly value manufacturing.

The market challenge

The last, and perhaps most critical, challenge is the question of the market. Assuming the US somehow managed to overcome all the other issues: it has built the three Vogtle-scale energy projects, the two factories, found tens of thousands of workers and developed the ecosystem around it all, it still needs to sell the stuff – both the aluminum and the gallium. Total US aluminum consumption runs at approximately four million tons annually, yet as we saw, producing just 100 tons of gallium requires 4.7 million tons of aluminum as an unavoidable byproduct. The entire domestic market couldn’t absorb this production: even capturing every aluminum customer in the US would leave 700,000 tons of surplus metal.

International markets offer no solution. Global aluminum markets already face structural overcapacity and US aluminum produced at market rates with higher costs and wages couldn’t compete with China on price. So should the US sell at a loss? What sustains the operation then? Would the US government subsidise the operations year after year, keeping the project running at a loss?

This all creates an economically irrational situation where producing a strategic material (gallium) requires maintaining permanently unprofitable industrial capacity (aluminum smelting). No market-based enterprise would undertake this voluntarily. All the more since, as we just saw, this requires an initial investment of $140 billion.

What about substitutes?

You’ve certainly thought about it: “If producing gallium ourselves is such a massive effort, surely we can substitute it for something else?”The problem is that material properties aren’t negotiable. Gallium Nitride semiconductors aren’t used because they’re trendy, they’re used because silicon physically cannot do what GaN does. GaN can handle 10x the voltage, operate at frequencies where silicon fails, and tolerate temperatures that would destroy silicon chips.

Think about it, if substitutes were viable, the Pentagon would already be doing it. The US military has known about rare earths vulnerability since at least China’s 2010 embargo against Japan. That’s 15 years to find alternatives. And yet here we are, with – again – a Raytheon executive stating: “GaN is foundational to nearly all the cutting-edge defense technology that we produce.”

And even if you could substitute gallium, you’d probably find yourself in the exact same place. A substitute that’s been mentioned is Silicon Carbide (SiC) but…China controls most SiC production too, and it doesn’t match GaN for the applications that matter most. And even if perfect substitutes existed for gallium – which they don’t – you’d still face the same problem for the other 20 elements on China’s export control list. The strategy of “substitute everything” eventually crosses into absurdity. At a certain scale, “find alternatives for 21 strategically critical materials” becomes functionally equivalent to disputing the results of the Big Bang – you’re demanding that nature provide you with different fundamental building blocks than the ones that exist.

Conclusion

So how long can the “rare earths card” be played? We just saw the titanic efforts that would be needed to simply produce less than a fifth of the amount of gallium that China produces: an initial investment of $140 billion, building two giga factories and three large-scale nuclear plants, finding and training over 35,000 highly specialised workers, building the entire industrial ecosystem around it. All for an operation that will never be able to compete with Chinese prices in global markets, and as such probably needs to be permanently subsidised by US taxpayers.

Take that and multiply it by 21, the total number of chemical elements on China’s export controls list (which again, is not the extent of it because they also have export controls on downstream products), and you start to grasp the strength of the “rare earths card”.

Another very similar element to gallium, also dominated by China and also on China’s export controls list, is indium, a byproduct of copper. Much like gallium, to break the indium stranglehold you’d have to rebuild a complete copper industry chain – mines, smelting, chemical processing, electricity, transportation. Do you start to understand Bessent’s panic? This isn’t something that a mere Manhattan Project or Apollo Program can solve, this is something far more intractable: China’s advantage isn’t technological, it’s systemic.

We’re not speaking of discrete projects here, we’re speaking about something that’d require a complete societal stack – from how children are educated to how capital is deployed. Consider what it takes to produce just one skilled aluminum smelter operator: first, a middle school student needs to see industrial work as a viable, respectable path – not failure to get into college. Then, they need access to a vocational school with up-to-date equipment and industry connections – schools the West mostly closed in the 1980s. Then, they need 2-3 years of training plus 3-5 years of on-the-job experience to become truly proficient. That’s 8-10 years from the decision point to competent operator. Now multiply that across 35,000 workers for this one element – then multiply by 21 elements, and multiply again all of this by all the supporting roles needed to build the facilities and staff the vocational schools.

China has this. In 2023, they had a total of 11,000 vocational schools nationwide with nearly 35 million students studying at these educational institutions. It’s normalised, systematic, continuous. The West doesn’t just lack the programmes – it lacks the entire cultural and institutional framework that feeds students into those programmes. You’d need to rebuild that framework before you could rebuild the workforce.

Or look at capital allocation: building rare earth capacity requires accepting decade-long losses and 20-year payback periods, extremely patient capital. Patient capital requires investors willing to accept long horizons. Long horizons require regulatory and political stability. Stability requires societal consensus that manufacturing is strategic. Consensus requires…we’re back to education, media, and culture.

So how long can China play the rare earths card? Looks like the realistic answer is: that one is here to stay for a very, very long time.

Courtesy:theconversation.com