On 1 July, China’s Ministry of Commerce quietly turned rare earth export control from paperwork into something closer to a criminal exposure. MOFCOM Announcement No. 26 created a public reporting system for anyone who exports a controlled mineral without a licence, disguises it as something else, routes it through a third country, or simply fails to flag a risk they had reason to suspect. Weeks earlier, Beijing had added ten more American firms to its own control list — including MP Materials and USA Rare Earth, the two companies Washington is counting on to break Chinese dominance of the sector. Two Japanese nationals were already sitting in detention in Dalian on smuggling allegations. None of this was a one-off. China’s 2025 licensing regime has become permanent infrastructure, and in 2026 it is tightening.
In April 2025, Beijing began requiring a Ministry of Commerce licence for exports of seven of the seventeen rare earth elements — including dysprosium and terbium, the two used in the permanent magnets inside electric-vehicle motors and precision-guided weapons — plus any oxide, alloy, compound or mixture containing them. The restriction bites regardless of who mines the ore, because China does not just dominate extraction: it controls roughly 90% of the world’s rare earth refining and separation capacity, the specialised chemical step that turns raw ore into usable metal. A country can own a mountain of rare earth-bearing rock and still depend entirely on Beijing to make it useful. That single fact is why “shadow” refining capacity — processing plants assembled quietly, country by country, without any pretence of a coordinated bloc — has become the insurance policy of choice for governments that do not want their industrial policy held hostage to a Chinese export licence.
Four mid-sized economies are placing that bet in parallel, and the details matter more than the headlines, because each one now depends on somebody else in a way that looks a lot like the arrangement it is trying to escape.
Kazakhstan has the biggest numbers and the least to show for them. Abu Dhabi’s sovereign fund ADQ has joined a $1.8 billion mining consortium; Washington’s Export-Import Bank and Development Finance Corporation have committed $1.6 billion toward Kazakh tungsten; Riyadh has signed its own mining deal; and a 2025 discovery near Astana claims more than 20 million tonnes of rare earth-bearing material. But Kazakhstan has no domestic separation facilities and relies on equipment its own officials call outdated, and whatever it mines has to leave on the Middle Corridor, a rail route of more than 4,000 kilometres whose cost and reliability Astana does not fully control. Its insurance policy against Beijing is underwritten almost entirely by American development-finance agencies and Gulf sovereign capital.
Stay ahead of the geopolitical week.
MD Briefing delivers expert analysis across five global fronts — the Indo-Pacific, energy, geoeconomics, European security, and the Middle East — every Monday morning. Free.
Vietnam adds corruption to the mix. It is often cited as holding the world’s second-largest rare earth reserves, though the US Geological Survey quietly cut that estimate this year from 22 million to 3.5 million tonnes — a reduction that says as much about how thin the original claim was as about the deposit itself. In March, Australia’s Lynas and South Korea’s LS Eco Energy agreed to build a metal-conversion plant, starting with samarium before moving to heavier elements. Meanwhile Vietnam’s flagship project, the Dong Pao mine, sits inside a graft investigation: the chairman of Vietnam Rare Earth JSC, the firm meant to refine Dong Pao’s ore, was arrested with five others over roughly $18 million in illegally mined material and $25.8 million in misappropriated funds. Blackstone Minerals, the Australian miner still pursuing a concession auction there, says it has no connection to the case and intends to proceed anyway. Vietnam’s refining future now rests on an Australian junior and a Korean conglomerate cleaning up after a scandal involving the firm that was meant to be its national champion.
Morocco is the most credible of the four, and the most tied to a single patron. Rather than mining new ore, state phosphate giant OCP is extracting rare earths from phosphogypsum, fertiliser-production waste that testing confirms contains all four elements used in high-strength magnets. The programme, run with Rainbow Rare Earths and a Moroccan university, is still pilot-scale, with commercial validation years off. But Rabat moved fast on diplomacy, signing a critical minerals memorandum with Washington on 4 February. Morocco’s hedge against China is, in practice, co-signed by the United States.
Indonesia is the cautionary case: despite years of speculation, it has no verified commercial rare earth mining, separation, or exports as of mid-2026. Its most plausible path — recovering monazite, a rare earth-bearing mineral, from Bangka-Belitung tin waste — depends on an industry where an estimated 80% of production already happens outside the law. A new state miner was created in February to map prospective blocks, which is a plan, not a supply chain.
The pattern is the same across all four. None of these governments is building refining capacity with its own money or technology. Every serious project has a foreign anchor — American development-finance vehicles, Gulf sovereign wealth, or Australian and South Korean firms with separation know-how nobody else has. That is not multipolarity, which requires several independent centres of power bargaining on roughly equal terms. It is a hub-and-spoke system with a different hub: fewer chokepoints, but chokepoints still, now denominated in American offtake agreements and allied capital rather than a Chinese export licence.
The fair objection: isn’t four partial alternatives still better than one absolute chokepoint? Yes. Beijing’s leverage over any single industry is no longer total, the way it was in 2025 when yttrium exports to the US fell 95% almost overnight. Producer governments also gain real bargaining power from being courted by Washington, Seoul, Canberra and the Gulf at once — Kazakhstan is already playing financiers off one another. But leverage over financing terms is not an independent processing base. Swapping “Beijing decides” for “a queue of bilateral partners, each tied to one foreign offtake deal, decides” is a change of supplier, not of structure.
Base case (roughly 55%): Diversification continues but stays slow and partial. By 2028–2029, Kazakhstan exports some processed tungsten and light rare earths via the Middle Corridor; Vietnam’s Lynas–LS Eco Energy plant produces modest volumes of samarium metal; Morocco’s phosphogypsum programme remains pre-commercial; Indonesia stays aspirational. China still controls the large majority of global refining capacity throughout, but its export licence becomes one constraint among several rather than the only one. This depends on US and Gulf financing continuing at its current pace and none of the four host governments rupturing with its anchor investor.
Downside case: One of the weakest links breaks the whole chain. If Vietnam’s corruption probe widens to freeze the Dong Pao concession auction, or a change in Washington’s priorities pulls back Kazakhstan’s development-finance commitments, the pipeline for the two elements in tightest supply — dysprosium and terbium — takes the hit, because Vietnam and Kazakhstan are disproportionately where non-Chinese heavy rare earth supply is supposed to come from. Given how thin this pipeline already is, a stall in one project delays the insurance policy for the entire group.
Upside case: Beijing’s own aggressiveness backfires. If the whistleblower mechanism and expanding entity list squeeze legitimate industrial buyers hard enough, fast enough, it could push Washington, Seoul, Canberra and Gulf financiers to stop competing for access to these four countries and start pooling capital and technical standards instead — an alliance of buyers rather than four separate bilateral bets. Nothing currently in motion suggests this coordination is happening, which is exactly why it would be underpriced if it started.
The rush to build rare earth refining capacity in Kazakhstan, Vietnam, Morocco and Indonesia is real money and real diplomacy, not vapourware. But it is not — yet, and may never become — a multipolar alternative to Chinese control. It is four separate insurance policies, each underwritten by the same narrow set of external financiers and technical partners, each one political shock away from stalling. Watch the Dong Pao concession auction in Vietnam, which Blackstone says it intends to pursue despite its partner’s arrest. If that auction closes cleanly and the mine gets built, it will be the first real evidence that one of these four projects can survive contact with local politics rather than merely surviving contact with a press release. If it stalls instead, treat every other announcement in this piece as inventory, not infrastructure.
