South Korea refines 16.4% of indium. Rare-earth mine output is unreported.

Korea's trade ministry plans a hotline and joint committee with Chinese authorities to speed critical-mineral imports during disruptions. 3 At the same time, Seoul joined the US-led Forum on Resource Geostrategic Engagement (FORGE) partnership and was selected to chair it through June 2026. The two tracks recognise a timing problem: diplomacy can protect this year's production while new plants chase the next decade.

Seoul plans to monitor 17 critical minerals more closely. It also allocated 250 billion won, about US$172 million at the reported exchange rate, to help Korean companies develop overseas mines. 3 That support shows the government is acting directly, although the wider mineral envelope should not be relabelled as rare-earth project finance.

The diversification list names the United States and Vietnam, with Laos appearing alongside them. 3 Those choices already match the emerging corporate map unusually well. A government response centre created in late 2025 gives firms a place to report supply stress, but its value will show up in response times rather than plant capacity. 9

Korea is keeping Chinese supply moving while its companies build slower alternatives around it.

What the numbers say

POSCO International says its Malaysia and Laos projects could secure about 4,500 tonnes of rare-earth separation and refining products each year. 4 The company wants to expand the base beyond 10,000 tonnes later. Both remain plans without a reported current domestic mine total behind them.

The group is also investing through a Korean corporate venture-capital fund. Its first rare-earth target received 8 billion won through that vehicle. The investee claims the ability to make heavy oxides at 99.5% purity. 4 A purity claim shows the intended product, although recurring output accepted by independent customers still needs to be demonstrated.

POSCO and ReElement announced a US$200 million US venture in May 2026. Stage one targets 3,000 tonnes of annual separation capacity. 5 The design can later double, with pilot output targeted for the fourth quarter of 2027 before commercial production in 2028. POSCO would control the venture while ReElement supplies its separation process.

Lynas and LS Eco Energy signed a preliminary arrangement for Australian or Malaysian oxides to feed a planned Vietnam plant. The first stage focuses on samarium, with operations targeted for the fourth quarter of 2026. 6 Preliminary is the controlling word because definitive supply and investment agreements were still to be negotiated.

Each company proposed buying about A$30 million of the other's convertible instruments. 6 The cross-investment would align feed and processing economics, but it had not closed at the Reuters report date. Before LS can use it as a dependable bridge to US magnet customers, the Vietnam plant must still produce metal accepted by customers.

At the later factory stage, the JS Link agreement is more specific. Its Kuantan plant is designed for 3,000 tonnes a year of sintered NdFeB magnets. 7 Lynas will invest about A$50 million and supply both Malaysian and Korean JS Link facilities through January 2038. A long feed term reduces one risk; construction and customer approval remain.

POSCO's supply base sits in Malaysia and Laos, while its largest separation proposal sits in the United States. LS is using Vietnam, and JS Link is expanding in Malaysia. 4 5 6 7 Adding those announced capacities as Korean production would erase local permitting and count corporate plans before they operate.

Domestic demand provides the commercial pull, and POSCO describes STAR Group as Korea's only rare-earth permanent-magnet producer. The company has also reported long automaker orders into North America and Europe. 8 Treat the exclusivity and order values as company claims, but the customer geography explains why a distributed chain can still serve Korean industrial policy.

The IEA expects announced non-Chinese later processing and factory capacity to remain well below projected demand in 2035. 2 Korea is trying to claim part of that scarce capacity through ownership and long supply arrangements. The honest 2030 scorecard should count tonnes accepted by customers and delivered through Korean-controlled chains, then show country of production in a separate field.

Korea's strategy succeeds when motors receive customer-approved magnets on time, even if the ore and processing never cross a Korean port.

What the numbers say