从一台烧结炉开始的溃败：印度稀土磁体"突围"与中国护城河观察

Defeat Starting from a Sintering Furnace: Observations on India's Rare-Earth Magnet "Breakthrough" and China's Moat

一、728亿卢比的雄心，与三次延期的招标

I. The ₹72.8-Billion Ambition and a Tender Extended Three Times

2025年11月，莫迪内阁批准"烧结稀土永磁体制造促进计划"（REPM），7年总盘子728亿卢比（约7.65亿美元），遴选最多5家企业、单厂上限1200吨/年、合计6000吨/年一体化产能，覆盖从NdPr氧化物到成品磁体的全链条，其中645亿按销量发销售挂钩激励、75亿作资本补贴。

In November 2025, the Modi cabinet approved the Scheme to Promote Manufacturing of Sintered Rare Earth Permanent Magnets (REPM): a 7-year envelope of ₹72.8 billion​ (~US$765 million), up to 5 beneficiaries, 1,200 t/yr per plant, 6,000 t/yr​ integrated capacity from NdPr oxide to finished magnets—₹64.5 billion as sales-linked incentives, ₹7.5 billion as capital subsidy.

计划的直接外力，是2025年4月中国对钐、钆、铽、镝、镥、钪、钇等中重稀土及含铽/镝钕铁硼永磁体实施出口许可证管理，印度车企库存一度只够4–6周。

The direct trigger was China's April 2025 export-license regime on Sm, Gd, Tb, Dy, Lu, Sc, Y and Tb/Dy-bearing NdFeB magnets, which once cut Indian automakers' inventory to 4–6 weeks.

但招标启动后连续延期：5月28日→6月29日→7月29日→8月12日，技术标8月13日开启。早期报道称仅交4份标书、名额空缺过半；截至8月13日重工业部最新披露，全球投标收到20份（L&T、Coal India等在列），表面"回暖"，但5个名额、6000吨目标、2年建设期不变，企业普遍质疑"小而全+限时补贴"的商业生存性。

Yet the tender slid: May 28 → Jun 29 → Jul 29 → Aug 12, technical bids opened Aug 13. Early leaks said only 4 bids; MHI's latest disclosure on Aug 13 puts global bids at 20​ (L&T, Coal India among them)—a face-saving "recovery," but 5 slots, 6,000 t and a 2-year build window are unchanged, and firms still question whether "small-integrated + time-boxed subsidy" can survive commercially.

钱不是问题，链条才是。印度这一轮不是"失败于没钱"，而是失败于把"有矿"误认为"有产业"。

Money was never the bottleneck; the chain was. This round failed not for lack of funds, but for mistaking "having ore" for "having an industry."

二、四道护城河：印度被卡在哪一环

II. Four Moats: Where India Actually Got Stuck

1. 原料天花板：723万吨储量，500吨氧化物

1. Feedstock ceiling: 7.23 Mt reserves, 500 t oxide

印度稀土储量全球第三（约723万吨），但唯一具备商业规模氧化物的国企IREL，NdPr氧化物年产能仅400–500吨，不到6000吨磁体目标所需1500–1700吨的1/4；重稀土镝、铽国内近乎空白，车规/风电/军工磁体绕不开。

India holds the world's third-largest reserves (~7.23 Mt REO), yet the only commercial oxide producer, state-run IREL, makes just 400–500 t/yr​ of NdPr oxide—under a quarter of the 1,500–1,700 t needed for 6,000 t of magnets. Heavy rare earths Dy and Tb are virtually absent at home, which high-end motor and defence magnets cannot do without.

2. 规模经济：中国单厂4万吨 vs 印度5厂共6000吨

2. Scale economics: China 40 kt/plant vs India 6 kt/5 plants

Lohum高管直言：中国一家厂年产4万吨磁粉，印度全国规划才6000吨且强制每厂"氧化物→金属→合金→磁体"全自干，违背分工摊薄成本的规律。

Lohum's Tarun Singh put it plainly: a single Chinese plant makes 40 kt of magnetic powder a year; India plans 6 kt total, and forces each plant to span oxide→metal→alloy→magnet in-house, throwing away the division-of-labor cost curve.

3. 设备墙：一台真空烧结炉，交付排到2028

3. Equipment wall: one vacuum sintering furnace, delivered 2028

全球70%+真空烧结炉产能在日本ULVAC等手中。印度拒买中国炉（交付6–7个月、价为日系1/3）而转向日本，交货期拉到15个月，且2025–2026年全球有约20个稀土自主化项目抢设备，印度小单排不到优先档，2年建设期从开局就逾期。

Over 70% of global vacuum sintering-furnace capacity sits with Japan's ULVAC and peers. India shunned Chinese furnaces (6–7 month lead, ~1/3 the price) for Japanese ones—15-month​ lead—while ~20 global de-China projects booked the same queue in 2025–26; India's small order lost priority, so the 2-year build window was behind schedule from day one.

4. 牌号与认证：补贴按重量发，买家无换供动机

4. Grade and certification: subsidy by weight, no switch incentive

PLI按公斤而非磁能积/矫顽力牌号发钱，低牌号中国磁体仍可自由出口，Ather、Tata AutoComp等换国产要重做电机设计与车规认证，风险自负。

The PLI pays by kilogram, not by grade (BHmax, coercivity). Low-grade Chinese magnets remain freely exportable; for Ather, Tata AutoComp et al., switching means re-designing motors and re-running automotive certification at their own risk.

IEA 2024口径：中国占全球稀土矿产量~60%、精炼分离91%、烧结NdFeB磁体94%。越下游，印度落差越大。

Per IEA 2024: China ~60% of mined output, 91%​ of refined separation, 94%​ of sintered NdFeB. The further downstream, the wider India's gap.

三、中国的牌：不是禁运，是许可证

III. China's Hand: Not an Embargo, but Licenses

2025年4月4日商务部、海关总署18号公告：钐、钆、铽、镝、镥、钪、钇相关金属/合金/氧化物/化合物及含铽、含镝钕铁硼永磁体、钐钴永磁体列入管制，出口须向商务部申领《两用物项出口许可证》，报关备注管制编码。

On April 4, 2025, MOFCOM & GACC Announcement 18 put Sm/Gd/Tb/Dy/Lu/Sc/Y items—and Tb/Dy-bearing NdFeB, SmCo magnets—under licensing; exporters must obtain a dual-use export license and declare the control code at customs.

2025年10月61号、62号公告把边界再推一层：

• 境外制造含中国原产管制物项价值占比≥0.1%的成品，反向纳入许可；

• 稀土开采、冶炼分离、金属冶炼、磁材制造、二次资源回收技术及产线调试维护技术，一律不准未许可外流。

10月后部分中重稀土公告曾随经贸磋商暂停至2026年11月，但18号公告（中重稀土物项本身）始终有效，且2025年12月起对大型磁材企业发放"通用许可证"，逐笔审批转向终端用户证明（人形机器人、汽车等民用用途可放，军工转售嫌疑卡死）。

A late-2025 pause touched some October measures through Nov 2026, but Announcement 18 itself stayed in force; from Dec 2025, big magnet makers got "general licenses," shifting from per-shipment review to end-user proof—civilian EV/robot use passes, defence diversion stalls.

这意味着：中国没有切断贸易，而是把流向控制权拿回手里。印度车企慌的是4–6周断供，但更长期的信号是——只要北京常态发证，印度自建产能的经济动机就被持续稀释。

The point: China didn't cut trade, it reclaimed flow-control. Indian OEMs panicked over a 4–6 week gap, but the deeper signal is that as long as Beijing licenses routinely, the economic case for India's own capacity keeps eroding.

四、印度"第二战场"为什么补不上中上游

IV. Why India's "Second Front" Can't Fill Midstream

新德里同步在做的：奥里萨/喀拉拉/安得拉建"稀土走廊"、修MMDR法放私企采矿、去澳/越/智/秘鲁锁矿、海外买50处资产、15亿卢比推回收。

New Delhi's parallel moves: "rare earth corridors" in Odisha/Kerala/Andhra/Tamil Nadu, MMDR amendment to let private mining, offtake deals in AU/VN/CL/PE, 50 overseas assets, ₹1.5 billion for recycling.

但分析师共识：矿山不是绑定约束，分离线+金属化+重稀土才是。没有分离，海外矿石只是转口原料；没有重稀土分离，军工与高性能电机磁体仍回头找中国或第三方转口。

But the analyst consensus holds: mines aren't the binding constraint—separation + metalization + heavy rare earths​ are. No separation, overseas ore is just transit feedstock; no heavy-RE separation, defence and premium-motor magnets still bounce back to China or third-party re-export.

Ashvini（浦那）用BARC技术开出的NdPr金属线月产15吨，价格比中国高15–20%，且印度至今无商业级烧结磁体出货；2030年5000吨目标目前是纸面。

Ashvini (Pune) runs a BARC-tech NdPr metal line at 15 t/month, 15–20% above Chinese price, and India has yet to ship a single commercial-grade sintered magnet; the 5 kt-by-2030 claim is still on paper.

五、结论：补贴买得到起步，买不到三十年

V. Bottom Line: Subsidy Buys a Start, Not Thirty Years

• 印度REPM不是"彻底归零"，8月13日20标让5厂6000吨在2029–2030有概率形成千吨级底盘，优先供军工与个别车企试点；

• 但"替代85%中国进口、成本正面脱钩"——未来5–7年做不到；

• 真正难的是跨过中国用三十年垒起的"分离—金属—磁体—专利—设备—认证"六道墙，不是内阁批一笔728亿卢比。

India's REPM isn't "dead"—20 bids on Aug 13 mean a kiloton-class floor​ by 2029–30 for defence and pilot OEM use is plausible.

But "replace 85% of Chinese imports and decouple on cost" is off the table for 5–7 years.

The real barrier is crossing the six walls—separation, metal, magnet, patent, equipment, certification—that China spent three decades​ stacking, not a cabinet vote of ₹72.8 billion.

稀土博弈的真相：矿是地质，磁体是文明。印度有前者，缺后者。

The truth of the rare-earth game: ore is geology, magnets are civilization. India has the former, not the latter.

关键时间线 / Key Timeline

• 2025-04-04 中国18号公告，中重稀土及Tb/Dy磁体许可证化

• 2025-11-26 印度内阁批REPM，728亿卢比/6000吨/7年

• 2026-03-20 全球RFP发出；4月7日标前会26家到场

• 2026-05→08-12 投标截止四度延期；8-13技术标开，20份投标入池

• 2025-12起 中国对龙头磁材企业发通用许可证，终用户证明制落地

继续阅读：印度   稀土   磁体   中国   合金

此文由 怡心湖 编辑，若您觉得有益，欢迎分享转发！：首页 > 观·世界 » 从一台烧结炉开始的溃败：印度稀土磁体"突围"与中国护城河观察Observations on India's Rare-Earth Magnet "Breakth