日本のレアアース危機 単一供給国からの脱却へ、なぜ東京は海底採掘に賭けるのか

On January 6, 2026, China’s Ministry of Commerce added seven categories of medium and heavy rare earth elements to a tightened export-control list bound for Japan: samarium, gadolinium, terbium, dysprosium, lutetium and others, folded into a broader list of roughly 1,100 dual-use items (CNBC). Beijing framed it as a response to Prime Minister Sanae Takaichi’s remarks on a possible Taiwan contingency, and aimed at Japan’s military supply chain (Malay Mail). But a dysprosium-stabilized magnet does not know whether it is going into a missile guidance system or an EV drive motor. Japan’s Ministry of Foreign Affairs called the controls “unacceptable and deeply regrettable.”

The exposure was already on the table before January. China supplies about 80% of Japan’s rare earth imports and a third of its rare earth magnet imports, and refines more than 99% of the world’s heavy rare earths — the dysprosium and terbium grades that keep a magnet strong at high temperature (Benchmark Minerals). Japan had already seen this coming in a smaller way: after an earlier round of Chinese restrictions in April 2025, magnet exports to Japan dropped sharply and bottomed at 80 tonnes in May before rebounding once those controls eased. The January 2026 list turned that pattern into policy that Japanese manufacturers now have to price into every EV motor, robot and wind turbine order running through 2030.

2010: the first cutoff, and what changed since

Japan is not short on rare earths as such — the elements sit in ordinary crustal rock almost everywhere. What Japan lacks is separation capacity: the plants that take mixed ore and split it into pure oxides at industrial scale. China built that capacity over three decades while producers elsewhere shut down as uncompetitive, absorbing environmental costs few other countries wanted to carry. Even where raw material comes from somewhere else, the refining step still routes back through Chinese processors. Benchmark Minerals’ trade figures show China’s exports of rare earth oxides and compounds to Japan actually grew, from 4,262 to 5,380 tonnes year-on-year, even as the two governments traded formal protests. There was nowhere else for most buyers to go.

Japan has lived this problem before. China cut rare earth exports to Japan once already, in 2010, over a maritime dispute near the Senkaku Islands. Fifteen years and a declared diversification strategy later, Japan is still 80% dependent on the same supplier for the same materials. Eighty percent is where the 2026 controls found Japan.

One earlier bet, and how long it took

Japan’s diversification record is not empty. On June 2, 2011, JOGMEC and trading house Sojitz formed Japan Australia Rare Earths B.V. and put roughly ¥20 billion (about $180 million at the time) into Lynas Rare Earths, an Australian miner then close to running out of cash, in exchange for a long-term supply relationship (Global Trade Alert). In March 2023, JOGMEC added a further AUD 200 million to fund a heavy rare earth separation line at Lynas’s Mt. Weld site in Western Australia, with Sojitz and JOGMEC securing rights to up to 65% of the dysprosium and terbium it produces for the Japanese market (JOGMEC). On October 30, 2025, that arrangement produced its first cargo: Sojitz imported heavy rare earths separated in Malaysia from Mt. Weld ore into Japan for the first time (Sojitz). Fourteen years passed between the first investment and the first shipment.

Four tracks, run at once

Diplomacy. On October 27, 2025, President Trump and Prime Minister Takaichi signed a non-binding critical minerals framework at Akasaka Palace, setting up a Critical Minerals Supply Security Rapid Response Group co-led by the US Secretary of Energy and Japan’s Minister of Economy, Trade and Industry, with a six-month window to identify and start funding joint mining, separation and recycling projects (Manufacturing Dive). A follow-on US-Japan Critical Minerals Investment Ministerial met in Tokyo on March 14, 2026, adding the Department of the Interior and the EPA to the table. The framework produces no refined material by itself. It is a channel for allied capital and permitting support to reach projects that still have to be built.

Stockpiles. Japan has run a national rare metal stockpile since 1983, jointly managed by JOGMEC (the Japan Organization for Metals and Energy Security) and industry, releasing material during past shortages (JOGMEC). Under the 2020 International Resource Strategy, Japan’s Agency for Natural Resources and Energy raised its target to 60 days of consumption for standard-risk minerals and up to 180 days for those at elevated geopolitical risk — the tier rare earths sit in (per Japan’s 2020 International Resource Strategy, catalogued by the IEA). 180 days does not close an 80% dependency. It buys time for the other three tracks.

The seabed. In February 2026, Japan’s deep-sea drilling vessel Chikyu spent a month roughly 1,900 km southeast of Tokyo, near Minamitorishima, and pulled up a continuous core of rare-earth-bearing mud from around 5,600–6,000 metres down — the first continuous lift of its kind at that depth. The Japan Agency for Marine-Earth Science and Technology (JAMSTEC) analyzed about 50 tonnes of the mud and found medium and heavy rare earths made up 54% of the total rare earth content: yttrium at 29.9%, neodymium at 18.4%, gadolinium at 4.9%, dysprosium at 4.6% (Nation Thailand). A large-scale trial targeting 350 tonnes of mud a day is set for February 2027. A feasibility report on domestic production is due by March 2028. Realistic output, if any, is still years out.

Recycling. The fourth track skips mining altogether. Japan’s Environment Ministry plans to start subsidizing recycling infrastructure — equipment for transport, storage and testing of scrap containing rare earths — from fiscal 2026, aimed at pulling neodymium and other elements back out of discarded motors rather than importing them again (Nikkei Asia). Daikin, Shin-Etsu Chemical, Hitachi and recycler Tokyo Eco Recycle are building automated recovery equipment through 2026, with full operations targeted for 2027; Daikin alone expects to collect around 10,000 compressors a year for magnet recovery.

Why a climate-tech agenda should care

Neodymium-iron-boron magnets, stabilized with dysprosium and terbium, are what let an EV traction motor or a direct-drive offshore wind turbine generator be built smaller and lighter than the copper-wound alternative. Japan’s own EV manufacturing targets and offshore wind auctions carry this dependency baked into their numbers without saying so. When the input runs short, the effect does not show up as a magnet headline. A magnet shortage in 2026 becomes a late turbine delivery in 2027, and the delay gets logged under a different cause.

Some manufacturers are changing the magnet recipe rather than the supplier. Proterial, the magnet maker that used to be Hitachi Metals, announced in 2025 that it was developing dysprosium- and terbium-free magnet formulations for EV motors, aiming to design the dependency out of the product rather than out of the supply chain (Benchmark Minerals).

Anyone building or financing climate-tech ventures with an Asia-Pacific supply chain should track that lag directly, not wait for the magnet story to make the news. Japan is running four responses to the same 80% number at once: a diplomatic channel that moves capital but produces nothing directly, a stockpile that buys time but not independence, a recycling push that starts drawing subsidies in fiscal 2026, and a seabed extraction bet that reports its first feasibility verdict in March 2028.

Tech for Impact Summit’s 2027 program is building out an energy and industrial-resilience track around exactly this kind of story. If your work sits at the intersection of climate policy, deep-tech investment and supply-chain security, search “Tech for Impact Summit” to learn more about the invitation-only gathering in Tokyo, May 2027, and reach out through the site about a seat.