China Powder Metallurgy 2026: The Dual-Track Breakthrough of MIM and High-End Metal Soft Magnetic Powder Cores

Chapter 1: Industry Overview — An Underestimated "Forming from Powder" Business

If we view the entire metalworking industry as a tree, powder metallurgy has long been the root buried in the soil. It does not have the striking exterior of a CNC machine tool, nor does it command the per-unit price of an aero-engine turbine blade that often runs to over a million RMB. Yet almost all gears, sprockets, bearing sleeves, solenoid valve cores, mobile phone hinges, smartwatch cases, EV inductor cores, and gas-turbine blades trace back to powder metallurgy in the end. This business grinds metal into powder and presses it back into parts; it sounds like a step backward, but behind it lies an entire independent body of process engineering, equipment science, and materials science.

In 2025, the global powder metallurgy market was about USD 34 billion, with a five-year CAGR of around 6 percent. China is the largest and fastest-growing single market within it. According to the China General Machinery Components Industry Association and its Powder Metallurgy Branch, total industry sales in China reached approximately RMB 48 billion in 2025, up 18 percent from RMB 40.8 billion in 2023. In their reports issued in the first half of 2026, securities companies pushed the full-year estimate to around RMB 53 billion, with a further step up forecasted for 2027.

The industry itself is not a single product category — it contains four sub-tracks with vastly different growth rates. The first is traditional press-and-sinter structural parts, mainly automotive transmission gears, oil pump rotors, and air-conditioning compressor components — the largest segment, accounting for about 45 percent of total powder metallurgy output, but with the lowest growth, just 3 to 5 percent annually, having entered the era of mature volumes. The second is soft magnetic powder cores — alloys such as iron-silicon, iron-silicon-aluminum, and amorphous nanocrystalline materials made into powder and pressed into magnetic components used in EV on-board chargers, photovoltaic inverters, energy-storage converters, 5G base-station power supplies, and server power supplies. This segment was about RMB 12 billion in 2025, with annual growth above 25 percent, the fastest within powder metallurgy. The third is metal injection molding, commonly called MIM in the industry, which mixes metal powder with a binder, injects it like a plastic injection molding process into a mold, and after debinding and sintering produces small complex metal parts close to net shape. This segment was about RMB 8.5 billion in 2025, with main customers including mobile phone hinges, smartwatches, smart glasses, TWS earbud skeletons, surgical instruments, and turbocharger actuators. The fourth is 3D printing metal powder, including titanium, superalloy, aluminum, and stainless steel powders specifically for selective laser melting and electron beam melting equipment — about RMB 6 billion in 2025, growing more than 30 percent annually, supplying aero-engine additive manufacturing, COMAC C919/C929 titanium structural parts, medical orthopedic implants, and aerospace structural parts.

These four sub-tracks share one upstream commonality: an intermediate product called "powder." Powder itself is an independent business. Press-and-sinter structural parts use water-atomized iron powder, with major suppliers including GRIPM Advanced Materials, Laiwu New Powder, in-house Dongmu powder, and Höganäs' Chinese joint-venture plant; soft magnetic powder cores use alloy soft-magnetic powder, with major suppliers including Poco Holding, Dongmu's metal soft-magnetic powder core subsidiary, and Yitong New Materials; MIM uses fine carbonyl iron powder and gas-atomized alloy powder, supplied by Yuean Advanced Materials, Yitong New Materials, and Germany's BASF; 3D printing uses spherical superalloy powder, spherical titanium powder, and spherical aluminum powder, supplied by Gaona Aero Material, Yitong New Materials, GRIPM Advanced Materials, the UK's LPW, Sweden's Sandvik Materials, and Canada's AP&C (acquired by France's GE Additive). One ton of water-atomized iron powder sells for about RMB 12,000 to 15,000 ex-works, one ton of fine MIM-grade gas-atomized powder sells for about RMB 80,000 to 120,000, one ton of 3D-printing-grade spherical superalloy powder sells for about RMB 600,000 to 1,000,000, and one ton of high-purity spherical titanium powder sells for about RMB 800,000 to 1,500,000. From ordinary iron powder to 3D-printing titanium powder, the unit price differs by nearly a hundred fold, and the gross-margin structure differs entirely as well. This is why the same industry contains both massive but low-margin water-atomized iron powder plants (gross margin in the low teens) and modest but high-margin spherical titanium powder plants (gross margin of 40 to 50 percent).

From the customer-structure perspective, the downstream of powder metallurgy is extremely fragmented yet flows intensively into a few industries. Automobiles account for 65 percent of total powder metallurgy consumption, with traditional internal combustion vehicles contributing transmission gears, oil pump rotors, connecting rods, and bearing sleeves, while EVs contribute inductor cores, OBC soft-magnetic powder cores, motor end parts, and thermal management actuators. Consumer electronics account for 15 percent, mainly MIM parts; between 2019 and 2025, fueled by the volume ramp of foldable phone hinges, this segment grew from a little over RMB 1 billion to nearly RMB 5 billion. Appliances account for 7 percent, including AC compressors, washing machine dampers, and range-hood impellers. Energy equipment accounts for 6 percent, including gas-turbine blades, wind turbine gearbox parts, and small nuclear components. Other medical, aerospace, hardware, and electrical applications together account for 7 percent.

From a global competitive standpoint, this business divides into two camps. One is the established Europe-Japan-US giants: GKN Powder Metallurgy of the UK (under Melrose Industries of the US), Sweden's Höganäs, Sweden's Sandvik Materials, Sumitomo Electric of Japan, Hitachi Metals (now merged into Proterial), and Nippon Steel of Japan. These companies have decades of accumulation in water-atomized iron powder, traditional press-and-sinter structural parts, and 3D printing spherical powder, with 60 to 80 percent share in the European, American, and Japanese OEM supply chains. The other is the Chinese local camp, led by Dongmu Co., Poco Holding, Yuean Advanced Materials, Yitong New Materials, Gaona Aero Material, GRIPM Advanced Materials, FRT, Jingyan Technology, and Tonglian Precision. The Chinese camp has reached the global front in metal soft-magnetic powder cores, MIM parts, and parts of 3D printing titanium powder; in sub-tracks like EV OBC soft-magnetic powder cores, foldable hinge MIM parts, and medical orthopedic titanium powder, Chinese makers already hold 30 to 60 percent global share.

Understanding the structure of this business gives every company, every figure, and every story in what follows a place to stand. Powder metallurgy is not a single industry but four sub-tracks layered on top of one another, each with very different growth rates, gross margin structures, and customer mixes. To understand it, the four tracks must be examined separately.

The industry is long underestimated by outsiders mainly because of the hidden nature of its products. A user folding a foldable phone has no clue that the hinge holds more than a dozen MIM parts; an EV driver pressing the accelerator does not know the electric control box contains kilograms of metal soft-magnetic powder cores in the OBC module; a passenger watching a commercial flight take off would not imagine the fuel nozzle of the engine was 3D-printed layer by layer from superalloy powder. All powder metallurgy products are "parts hidden inside systems," not items end consumers can directly perceive. This invisibility leads to industry exposure far below its real scale, and capital markets have long lagged actual operating performance in recognizing it. It was only between 2023 and 2025, when foldable phone hinges, EV OBCs, and AI server power supplies pushed powder metallurgy output to a new level, that the industry entered the field of view of buy-side researchers and the media.

Between 2020 and 2025, China's powder metallurgy industry posted a compound annual growth rate of about 12 percent, significantly higher than the industry-wide 5 percent for machinery manufacturing. The growth-rate gap stems from sub-track structural shifts: the share of traditional press-and-sinter structural parts fell from 60 percent in 2020 to 45 percent in 2025; soft-magnetic powder cores rose from 8 percent to 25 percent; MIM rose from 15 percent to 18 percent; 3D printing metal powder rose from 3 percent to 12 percent. This structural shift lifted the industry-wide average gross margin from 16 percent in 2020 to 24 percent in 2025.

In the second half of 2025 and the first half of 2026, another phenomenon emerged: the gross-margin gap between sub-categories widened further. Press-and-sinter structural parts held a gross margin of 15 to 20 percent; metal soft-magnetic powder cores climbed to 32 to 38 percent thanks to strong downstream demand; MIM fell back to 22 to 26 percent because of price competition in consumer electronics; 3D printing metal powder held at 40 to 50 percent on strong demand from aerospace and medical. This margin divergence reflects the bargaining power differential each sub-track has in the value chain — the closer to high-end processes, aerospace and medical, or AI compute demand, the stronger the pricing power; the closer to traditional automotive structural parts or mass consumer electronics, the weaker.

Another key feature of the industry is the geographic distribution of industrial clusters. China's powder metallurgy has four core clusters: first, the Yangtze River Delta (Nantong Haimen, Taixing, Yangzhou in Jiangsu; Hefei in Anhui; Shanghai periphery), home to Dongmu Haimen, Yitong Taixing, Tonglian Precision, and Yuean Maanshan, focused on metal soft-magnetic powder cores, MIM, and gas-atomized alloy powder; second, the Pearl River Delta (Heyuan, Huizhou, Shenzhen periphery in Guangdong), home to Poco Holding and FRT, focused on metal soft-magnetic powder cores and MIM; third, the Bohai Rim (Yantai, Laiwu in Shandong; Beijing; Zhuozhou in Hebei), home to GRIPM Advanced Materials, Gaona Aero Material, and Laiwu New Powder, focused on water-atomized iron powder and superalloy powder; fourth, Central and Western China (Zhuzhou in Hunan; Deyang in Sichuan; Xi'an in Shaanxi), home to Xi'an Sailong, Zhuzhou Diamond, and the Deyang aero engine-related powder plants, focused on spherical titanium powder and aero-engine supporting powder. The geographic distribution of these four clusters reflects the real structure of the value chain: the Yangtze River Delta and the Pearl River Delta are close to consumer electronics and EV OEMs; the Bohai Rim is close to steel raw materials; Central and Western China are close to aerospace assembly plants.

Viewed against the global map, China's share of the global powder metallurgy market has risen from about 18 percent in 2020 to about 26 percent in 2025, and is projected to reach about 35 percent by 2030. The drivers are threefold: the rising global share of China's EV, consumer electronics, and AI server downstream industries pulling powder metallurgy demand along; technology breakthroughs by Chinese local makers in high-value-added sub-categories such as metal soft-magnetic powder cores, MIM parts, and 3D printing metal powder; and the cost competitiveness and supply chain completeness of Chinese manufacturing overall. North America, Europe, Japan, and Korea have seen their global share decline correspondingly, though they still hold important positions in mature categories such as traditional press-and-sinter structural parts. Southeast Asia, India, and Brazil have inched up in share, but absolute scale remains small, mainly absorbing low-end capacity transfers.

Another important dimension is "upstream mineral raw materials." The most upstream of powder metallurgy's raw-material chain is iron ore, nickel ore, copper ore, titanium ore, and rare earth ore. China is the world's largest iron ore importer (mainly from Australia and Brazil) and one of the world's largest rare earth resource countries and titanium ore producers. This "upstream raw material + midstream powder + downstream application" integrated chain gives China a full-chain advantage that other countries find hard to replicate. But it should also be noted that some critical alloying elements (rhenium, tantalum, hafnium, platinum group metals) remain heavily import-dependent, representing latent supply-chain fragility.

In 2025 and 2026, another notable trend emerged: deep integration of powder metallurgy with the three macro themes of "green manufacturing," "digital manufacturing," and "smart manufacturing." On the green side, head firms launched "green powder" (lower carbon footprint), "green magnetic cores" (recyclable design), and "green MIM parts" (using renewable feedstock) catering to high-end downstream ESG demands. On the digital side, leaders increased investment in industrial internet, AI quality prediction, digital twin process simulation, and other new technologies, improving production efficiency and product quality. On the smart side, leaders accelerated upgrades like "lights-out factories," "intelligent production lines," and "robot collaboration," reducing labor costs and improving product consistency. The convergence of these three themes is reshaping the production model and competitive landscape of powder metallurgy.

Chapter 2: Process Routes — Press-and-Sinter, Injection Molding, Hot Isostatic Pressing, Laser Sintering

Powder metallurgy is essentially a "forming from powder" process route, parallel to the three traditional metal-processing routes of casting, forging, and machining. The difference is that casting melts metal into liquid then solidifies it into shape, forging deforms metal blocks in solid state, machining removes excess from a block to leave the finished piece, while powder metallurgy grinds metal into powder, presses it into a "green compact" near net shape, then through sintering enables the powder particles to diffuse together at high temperature, finally producing a dense metal part. The route's main advantages are material savings (net-shape yield exceeds 95 percent, often only 40 to 60 percent in machining), capability for complex shapes (gears, sprockets, internal grooves in valve cores that tools cannot reach can be formed in one shot), and the ability to make special materials (tungsten-copper composites, iron-silicon-aluminum alloys, nanocrystalline alloys that cannot be cast or forged can only follow the powder route). Drawbacks: lower single-piece strength than forgings, difficulty achieving full density, high mold-development costs, and uneconomical for small batches.

The four mainstream process routes within the industry differ far more than meets the eye.

The first is press-and-sinter (PS), the oldest and most mature powder metallurgy process, scaled up in Europe by the 1920s. The flow is to mix metal powder, place it in a tungsten-carbide mold, apply 400 to 800 megapascals of pressure with a mechanical or hydraulic press to form a green compact, then send it to a mesh-belt or pusher furnace and sinter at 1,100 to 1,250 degrees Celsius in a reducing atmosphere for 30 minutes to 2 hours, achieving densities of 85 to 92 percent. This route suits relatively simple shapes, parts weighing several grams to several hundred grams, in large volumes where extreme density is not required. The most typical products are automotive transmission gears, sprockets, oil pump rotors, connecting rods, bearing seats, eccentric wheels in AC compressors, and valve plates. The largest Chinese producer is Dongmu, with 2025 press-and-sinter structural-parts revenue of about RMB 2.5 billion, supplying nearly every OEM including FAW-VW, SAIC-GM, BYD, Chery, Great Wall, Honda China, and Toyota China. The largest overseas player is GKN Powder Metallurgy of the UK, with 2025 global revenue of about EUR 1.4 billion and 30 to 40 percent share in North America and Europe.

The second is Metal Injection Molding (MIM), matured in the US in the 1980s, moved to Europe and Japan in the 1990s, and to China in the 2000s. The flow is to mix fine metal powder (typically 5 to 20 micrometers) with thermoplastic binder (mostly POM or PE with a little wax) at about 60:40 by volume into a "feedstock," inject it into a mold like plastic injection molding to form a green compact, debind with solvent or catalysis to obtain a brown compact, then sinter in a vacuum or hydrogen furnace at 1,200 to 1,380 degrees Celsius for 4 to 8 hours to reach densities of 96 to 99 percent. MIM's greatest advantage is the ability to form any complex shape of small parts (typically 0.1 to 200 grams), with density and strength close to forgings, mass-producible at low single-piece cost like injection molding — sometimes only a few cents to a few yuan per piece. Typical products are phone hinges, smartwatch cases, AirPods charging case hinges, dental orthodontic brackets, surgical instruments, nozzle rings and blades in turbochargers. The largest Chinese MIM makers are Jingyan Technology and Tonglian Precision: 2025 Jingyan revenue of about RMB 2.8 billion (MIM about RMB 2.0 billion), Tonglian revenue of about RMB 1.3 billion (almost all MIM). The largest overseas players are Germany's ARC Group (Danaher subsidiary) and Indo-MIM of the US (an Indian company listed in the US).

The third is Hot Isostatic Pressing (HIP), developed by the US Battelle Memorial Institute in the 1950s and matured in the 1980s. The flow is to load metal powder (usually spherical) into a metal or glass capsule, vacuum and seal, then HIP at 1,000 to 1,300 degrees Celsius and 100 to 200 megapascals of argon pressure for 2 to 4 hours, diffusing powder particles into a fully dense part. HIP's main advantage is fully dense high-end alloy parts with uniform isotropic microstructure (no directional anisotropy). Typical products are aero-engine high-pressure compressor disks, gas-turbine turbine disks, nuclear-reactor pressure vessel internals, and deep-sea drilling valve bodies. The largest HIP equipment maker in China is AVIC Mete; the largest HIP service provider is Gaona Aero Material. The largest overseas players are Sweden's Quintus Technologies (a former ASEA Brown Boveri subsidiary) and the US's PSM. HIP is characterized by high unit price, small batches, and customers concentrated in extreme-condition fields: aerospace, gas turbines, nuclear, and deep sea.

The fourth is Selective Laser Melting (SLM) and Electron Beam Melting (EBM). These are the two forming routes for 3D-printed metal powder. SLM melts powder layer by layer with a laser on a powder bed to form 3D parts; EBM uses an electron beam instead of a laser to melt powder in a vacuum. Both are the ultimate "forming from powder": each powder layer is just 20 to 60 micrometers thick, the laser or electron beam scans along the designed path, building complex parts impossible by traditional means. Advantages: extreme design freedom (topology-optimized hollow structures, internal cooling channels, conformal thin walls); drawbacks: expensive equipment, expensive powder, low speed, poor surface finish requiring post-processing. Typical products: aero-engine fuel nozzles (GE's LEAP nozzle, originally an assembly of 20 parts, printed as one piece), medical orthopedic implants (titanium hip joints, intervertebral cages), tool inserts (conformal-cooling injection-mold inserts), aerospace satellite lightweight brackets. The largest SLM equipment makers in China are Bright Laser Technologies (BLT) and Farsoon Tech; the largest powder makers are Gaona Aero Material, Yitong New Materials, and AVIC Mete.

The differences across the four routes determine completely different downstream customer mixes, gross-margin levels, and technical barriers. Press-and-sinter is mature with fragmented customers, gross margin 15 to 20 percent, typical listed company PE in the low teens; MIM is medium-barrier, customers tilted to consumer electronics and medical, gross margin 20 to 30 percent, PE 20 to 30; HIP is high-barrier, customers in aerospace and gas turbines, gross margin 30 to 40 percent, PE 30+; SLM/EBM is equipment-and-powder integrated, customers in aerospace and medical, gross margin 30 to 50 percent, PE 40+. Understanding the essential differences across these four routes is the key to understanding the valuation structure of the entire powder metallurgy industry.

The four routes are not isolated; they interpenetrate and substitute each other. Press-and-sinter and MIM compete on mid-complexity structural parts: the same automotive turbocharger actuator can be made either way. Press-and-sinter wins on large volumes and low cost; MIM wins on complex shape and high density. The choice depends on geometric complexity, annual volume, and customer density requirements. MIM and HIP compete in high-end medical implants: a titanium hip joint can be MIM+HIP densified or directly EBM-printed in one shot. MIM+HIP suits standard high-volume products; EBM suits customized products. HIP and SLM/EBM compete in high-end aero-engine parts: a turbine disk can be HIP-formed from powder superalloy or SLM-printed then HIP-densified. HIP suits relatively simple disk-shaped parts; SLM suits complex nozzles and combustor liners.

The interpenetration is also reflected in the rise of "hybrid processes." Starting in 2025, leading Chinese firms began exploring MIM+HIP, SLM+HIP, and press-and-sinter+HIP hybrid processes, combining the strengths of different routes. For example, Tonglian Precision developed a "MIM+HIP" process in 2025, MIM-sintering titanium alloy parts then HIP-densifying to push density from 97 percent to 99.5 percent, mainly for medical orthopedic implants and high-end foldable phone hinges. Gaona Aero Material developed an "SLM+HIP" process in 2025, HIP post-treating SLM-printed superalloy parts to remove internal microcracks and pores, mainly for aero-engine high-pressure compressor disks. The rise of hybrid processes signals a new stage of innovation in powder metallurgy: no longer optimization of a single route, but combinational innovation across multiple routes.

The four routes also differ markedly in equipment investment intensity. Press-and-sinter's core equipment is a mechanical or hydraulic press (RMB hundreds of thousands to a few million per machine) plus mesh-belt or pusher furnace (several million to tens of millions), with whole-line investment of tens of millions to over a hundred million yuan. MIM's core equipment is the injection molding machine (a few hundred thousand yuan), debinding furnace (a few hundred thousand), and vacuum sintering furnace (a few million), with whole-line investment from tens of millions to several tens of millions. HIP's core equipment is the HIP furnace (RMB 20 to 50 million per machine), whole-line investment one to several hundred million. SLM's core equipment is the SLM printer (RMB 5 to 20 million each), whole-line investment from tens to hundreds of millions. EBM's core equipment is the EBM printer (RMB 10 to 30 million each), whole-line investment from tens to hundreds of millions. The higher the equipment investment intensity, the lower single-line capacity and the higher fixed cost share, meaning each route imposes very different demands on a company's balance-sheet structure, cash-flow model, and expansion pace.

In 2025, China's powder metallurgy R&D spending totaled about RMB 2 billion (4.2 percent of industry revenue). Listed leaders (Dongmu, Poco, Yuean, Gaona, GRIPM, BLT, Farsoon, Jingyan, Tonglian, FRT) accounted for about RMB 1.2 billion. R&D spending is allocated across three buckets: powder formulation, sintering process, and equipment process-pack development. Compared with overseas giants, Chinese leaders' R&D-to-revenue ratio and absolute value have approached GKN Powder Metallurgy and Sumitomo Electric's powder metallurgy units. This is why Chinese local players have caught up with — and in some cases overtaken — overseas giants over the past decade.

The four routes correspond to different "capital return cycles." Press-and-sinter line investment recovers in three to five years (relatively low equipment cost, stable utilization). MIM recovers in four to six years (high mold cost, slow initial capacity ramp). HIP recovers in six to ten years (single-unit investment huge, utilization sensitive to order swings). SLM/EBM recovers slowest, eight to fifteen years (both equipment and powder costs high while product prices constrained by downstream bargaining). The differing payback cycles dictate which kinds of investors suit each route: press-and-sinter for traditional manufacturing capital; MIM for consumer-electronics-chain capital; HIP for military and aerospace capital; SLM/EBM for long-cycle industrial and government-backed funds.

Cost-structure comparisons across the four routes are also instructive. In press-and-sinter, raw powder accounts for 50 to 60 percent, labor and manufacturing overhead 20 to 30 percent, depreciation 10 to 20 percent. In MIM, raw powder accounts for 40 to 50 percent (fine powder is more expensive), labor 20 percent, mold depreciation 10 to 15 percent, other manufacturing overhead 15 percent. In HIP, raw powder (spherical) accounts for 40 to 50 percent, energy and argon 15 to 20 percent, depreciation 15 to 20 percent, other 10 to 15 percent. In SLM/EBM, raw powder accounts for 30 to 40 percent (because product prices are high), equipment depreciation 20 to 30 percent, post-processing 20 to 30 percent, other 10 to 15 percent. The differing cost structures determine where cost optimization should focus: in press-and-sinter, raw powder procurement cost; in MIM, mold-development capability; in HIP, capacity utilization; in SLM/EBM, equipment-powder synergistic optimization.

Chapter 3: Critical Powders — Water-Atomized Iron, Gas-Atomized Alloy, Spherical Superalloy, Iron-Silicon-Aluminum Soft Magnetic

"Powder" itself is the very upstream of powder metallurgy, and the segment of the business with the widest margin gap. Both called "metal powder," water-atomized iron powder sells for RMB 12,000 per ton, while 3D-printing spherical superalloy powder sells for RMB 1,000,000. The difference is nearly a hundredfold. The difference lies in purity, particle-size distribution, sphericity, oxygen content, and flowability.

Water-atomized iron powder is the largest by volume. The flow: iron melt flows from a tundish, high-pressure water jets at 300 meters per second break the molten stream into fine droplets, which solidify into irregular powder particles in water, then dried, reduction-annealed, and screened. Water-atomized powder is characterized by irregular shape (good for particle interlocking and green-compact strength during pressing), wide particle-size distribution (20 to 200 micrometers), relatively high oxygen content (0.15 to 0.25 percent after reduction annealing), and low unit price (RMB 12,000 to 15,000 per ton). Downstream is mainly press-and-sinter structural parts. Typical players: GRIPM Advanced Materials with 2025 water-atomized iron powder revenue of about RMB 1.5 billion and capacity of about 100,000 tons; the Höganäs Chinese JV plant Xinghua Hengrui at about 60,000 tons; Shandong Laiwu New Powder; and in-house Dongmu powder. This is a scale business with margins of 12 to 15 percent and stable cash flow.

Gas-atomized alloy powder sits at mid-tier. The flow: alloy melt (iron-, nickel-, or cobalt-based) flows out of a crucible, high-pressure nitrogen or argon atomizes the stream into fine droplets that solidify into near-spherical powder particles. Features: near-spherical (sphericity above 0.8, good flowability suitable for automation), controllable particle size (5 to 45 micrometers tight distribution), low oxygen content (below 100 ppm with argon atomization), mid-range unit price (RMB 80,000 to 120,000 per ton). Downstream: mainly MIM feedstock, HIP processes, mold weld build-up. Major Chinese players: Yuean Advanced Materials (carbonyl iron is the mainstay, but gas-atomized alloy powder share is rising), Yitong New Materials, and Dongmu's Nantong Dongmu Tech subsidiary. Gross margins 20 to 30 percent.

Spherical superalloy powder is high-end powder, the key raw material for 3D printing aero-engine parts. The flow is more delicate than gas atomization, typically using Electrode Induction Melting Gas Atomization (EIGA) or Plasma Rotating Electrode Process (PREP). EIGA uses superalloy rod as electrode and inductive coils to melt crucible-free, with droplets falling into an argon atomization chamber. PREP rotates an alloy rod at high speed; a plasma torch melts the rod's end to form a liquid film, ejected by centrifugal force into droplets. Both yield highly spherical (sphericity 0.95+), tightly sized (15 to 53 micrometers), low-oxygen (below 50 ppm) spherical powder. This powder sells at RMB 600,000 to 1,000,000 per ton. The largest Chinese producer is Gaona Aero Material, with 2025 superalloy powder revenue of about RMB 600 million, supplying AECC main engine makers, COMAC, CASIC, and BLT; Yitong also has PREP spherical titanium capacity with 2025 spherical powder revenue of about RMB 300 million. The largest overseas players: AP&C (acquired by GE Additive), LPW (acquired by Carpenter), and Germany's Nanoval. Margins are 40 to 50 percent.

Spherical titanium is another high-end powder, mainly supplying medical orthopedic implants, COMAC C919 titanium structural parts, and aerospace satellite brackets. Production is also PREP or EIGA, but with stricter purity requirements (titanium readily absorbs O, N, H). Largest Chinese producers of spherical titanium: Gaona Aero Material, Xi'an Sailong, and Yitong. One ton of high-purity spherical titanium can sell for RMB 800,000 to 1,500,000, with margins of 40 to 60 percent.

Iron-silicon-aluminum (Sendust) soft magnetic powder is another special powder, dedicated to soft-magnetic powder cores. The flow water- or gas-atomizes the iron-silicon-aluminum alloy into powder, passivates it (coats particle surface with an insulating oxide layer), mixes with organic binder and presses into magnetic cores, and finally sinters into product. The key is not particle size or shape, but insulating-film thickness and uniformity, plus magnetic permeability and loss. The largest Chinese producer is Poco Holding, with 2025 alloy soft-magnetic core revenue of about RMB 2.8 billion (powder is in-house), customers including BYD, CATL, Sungrow, Huawei, and Inovance; Dongmu's subsidiary Dongmu Magneto had 2025 metal soft-magnetic core revenue of about RMB 1.8 billion; Yitong also has some soft-magnetic capacity. Largest overseas: Sumitomo Electric of Japan, Magnetics (Spang) of the US, and Höganäs of Sweden (Somaloy brand). Margins are 30 to 40 percent.

Carbonyl iron powder is another special powder, mainly supplying ultrafine iron powder for MIM and soft-magnetic cores. The route is not atomization but chemical: iron and CO react under high pressure to form iron pentacarbonyl liquid, then heated to decompose into ultrafine spherical iron powder. Features: extremely fine particles (1 to 5 micrometers), ultra-high purity (99.9 percent+), good sphericity. Price: RMB 80,000 to 150,000 per ton. The largest Chinese producer is Yuean Advanced Materials, with 2025 carbonyl iron revenue of about RMB 600 million, having almost fully replaced BASF's Asian market share. Globally, only BASF in Europe still produces.

Once the structure of powders is clear, all downstream firms' margins, valuations, and customer mixes have explanatory power. Powder sets the ceiling for parts; parts set customer acceptance; customers set the maker's cash flow.

The "devil" of powder process is in the details. Although water-atomized iron powders from different plants may look alike, their in-customer performance varies completely. The deciding parameters include over a dozen: chemical composition (pure iron or pre-alloyed; uniformity of alloying elements), particle-size distribution (D10/D50/D90), apparent density (g/cm³), flow rate (s/50g), compressibility (density at given pressure), green-compact strength (MPa), sintering shrinkage (%), oxygen content (ppm), carbon content (ppm), sulfur content (ppm), grain morphology (under SEM), internal defects (pores, inclusions). Each affects customer yield, cycle time, and final part performance. To switch powder sources, customers must re-tune pressing, sintering curves, and post-processing — a three- to six-month process. This is the natural moat of the powder business.

Spherical superalloy powder process details are stricter. In PREP, electrode rod rotation speed (10,000 to 30,000 rpm), plasma torch power (40 to 100 kW), argon purity in the chamber (99.999 percent+), chamber pressure (1 to 5 atm), and rod chemical uniformity — any micro-fluctuation affects final sphericity, particle size distribution, oxygen content, and secondary particle content. One ton of high-end spherical superalloy powder takes three to five days to produce, with dozens of process parameters continuously monitored. Such capital and technical barriers underpin the high margin. Among Chinese spherical superalloy producers, only Gaona and Yitong stably supply aerospace-grade product (meeting AMS 5662, AMS 5664, AMS 5708, etc.); others sell only into civil 3D printing or non-aerospace markets.

Carbonyl iron process is also intricate. Reaction temperature (180 to 200 degrees Celsius), pressure (150 to 200 atm), CO purity (99.99 percent+), iron feedstock purity and morphology, thermal decomposition temperature (300 to 400 degrees Celsius), decomposition atmosphere (hydrogen or nitrogen) — all critical. Carbonyl iron splits into two core grades: CIP (Carbonyl Iron Powder, standard, 4 to 10 micrometers) and CIP-SQ (Super Fine, 1 to 3 micrometers). CIP-SQ is mainly used in MIM feedstock, EMI shielding materials, and ultra-high frequency soft-magnetic cores, selling at two to three times CIP. Yuean broke through stable CIP-SQ mass production in 2025, raising the domestic share from under 30 percent in 2023 to nearly 100 percent in 2025, directly replacing BASF's Asia-Pacific share.

Iron-silicon-aluminum powder passivation is another core difficulty. The particle surface needs a uniform, dense, insulating oxide layer (phosphate, silicate, chromate, or epoxy), tens to hundreds of nanometers thick, that directly governs resistivity, permeability, and loss. Two mainstream routes: chemical passivation (immerse powder in phosphoric or chromic acid and bake) and gas-phase passivation (form an insulating layer on particle surface in a reaction furnace). Poco developed third- and fourth-generation passivation between 2018 and 2022, raising magnetic core resistivity by two orders of magnitude and cutting loss by 30 percent — the key to its leadership in high-power density. A fifth generation (nano ceramic coating) completed pilot in H1 2026, with mass production targeted for 2027, aiming at 800V on-board chargers and AI server power supplies.

The shared takeaway from these process details: the real moat of powder metallurgy is not "can you make powder" but "can you stably make powder with consistent parameters and uniform performance." The former just needs equipment and a recipe; the latter needs over a decade of process accumulation, hundreds of process engineers' experience, and thousands of experimental data points. That is why all global powder metallurgy leaders have existed for many decades (Höganäs founded 1797, GKN 1759, Sumitomo Electric 1897), and Chinese local leaders also have 20–30+ years of accumulation (Dongmu 1958, Poco 2009, Yuean 2011, Gaona 1956, GRIPM 2005 restructured). Powder metallurgy is not an industry "pulled up by capital" — it is an industry "earned by time."

Another important powder classification is by alloy system. Powders break down roughly into ten categories: pure iron powder (water-atomized pure iron and reduced iron, mainly for low-end press-and-sinter parts, EMI shielding, and electrolytic powders); pre-alloyed iron-based powder (iron-copper, iron-nickel-copper, iron-molybdenum, iron-copper-molybdenum, mainly for mid-to-high-end press-and-sinter parts, lifting mechanical performance and wear resistance via pre-alloying); stainless steel powder (304, 316L, 17-4PH, mainly for MIM, 3D printing, HIP, etc., to make stainless precision parts); iron-silicon-aluminum soft-magnetic powder (Sendust, Mo-Permalloy, High Flux, etc., dedicated to soft-magnetic cores); titanium powder (pure titanium, TC4, TA15, for medical orthopedic implants, aerospace structural parts, COMAC structural parts); aluminum powder (pure Al, AlSi10Mg, AlMgScZr, for 3D-printed aerospace satellite structural parts, consumer-electronics high-end structural parts); superalloy powder (Inconel 718, Inconel 625, Hastelloy X, CM247LC, for aero engines, gas turbines, nuclear reactors); copper, copper-alloy, silver, and gold precious-metal powders (for electronics, decoration, electrode materials); cemented carbide powder (mainly tungsten carbide, for carbide tools — outside the strict scope of this report but process-related); rare-earth permanent magnet powder (NdFeB, SmCo powder, for high-performance permanent magnets — also outside scope but process-related).

Different alloy systems differ entirely in process, application, and margin structure. A powder firm typically specializes in one or two systems rather than going "broad and shallow." For example, GRIPM specializes in pure iron and pre-alloyed iron powders; Yuean specializes in carbonyl iron; Gaona in superalloy and titanium powders; Yitong in spherical powders (various alloys but distinguished by sphericity); Poco in iron-silicon-aluminum soft-magnetic powders (in-house). This division of labor is typical of powder metallurgy and the root reason that leaders can maintain deep technical advantage.

"Batch-to-batch consistency" is another key quality indicator worth unpacking. Customers typically require three to five consecutive batches of samples for comparison during powder qualification; any noticeable batch difference fails qualification. Consistency comes from rigorous process control: tightly controlled melt composition each heat, stable atomization parameters each run, identical screening and post-processing each batch. Over the past decade, Chinese leaders introduced industrial automation, smart manufacturing, and statistical process control (SPC) management practices, improving batch consistency from the early significant fluctuations to a level comparable to overseas giants. This is the foundation of domestic substitution.

"Failure mode" analysis of powders is another deep technical topic. When downstream products experience quality issues, powder failure modes must be traced. Common failures include: hollow particles (gas entrapment during gas atomization); excessive surface oxide layer (poor sintering bonding); abnormal size distribution (poor flowability); foreign particles (defects); composition deviation (off-spec performance). A leading powder maker maintains a complete failure-mode-analysis system using SEM, XRD, EDS, and chemical analysis to reverse-engineer process optimization. This capability is the key to leaders' continuous quality improvement.

"In-line inspection" technology is also progressing rapidly. Traditional powder inspection samples offline — low efficiency, lagging. New in-line inspection uses laser diffraction, X-ray, and machine vision for real-time monitoring during production, catching anomalies instantly. Chinese leaders began heavy adoption of in-line inspection in 2025 and 2026, markedly improving process controllability. Widespread in-line inspection will lift quality across the industry to a new level.

Powder application also involves "powder pre-processing" and "powder post-processing" segments. Pre-processing includes screening (by particle size), mixing (homogenizing alloys and sizes by ratio), adding lubricants (for particle flow during pressing), and adding binders (to preserve green-compact strength). Post-processing includes passivation (insulating coating), surface activation (improving interparticle diffusion), and recycling (handling 3D printing unmelted powder). These pre- and post-processing steps look insignificant but directly determine final performance. Leaders generally integrate them in-house, while second- and third-tier firms outsource.

Chapter 4: Major Players — Six Core Domestic Leaders Plus Four Overseas Giants

China's powder metallurgy core firms include six listed companies plus several private niche leaders. Each has its own dedicated track and moat.

Dongmu Co. (600114.SH) is the largest comprehensive powder metallurgy company. 2025 revenue about RMB 5.6 billion, net profit attributable to parent about RMB 600 million, gross margin about 26 percent. Three business segments: traditional press-and-sinter structural parts mainly for automotive OEMs (2025 revenue about RMB 2.5 billion); metal soft-magnetic powder cores for EV OBC and PV-storage (2025 revenue about RMB 1.8 billion); MIM parts mainly for consumer-electronics foldable hinges, accessed via acquisition of Japan's IMECO MIM plant (2025 revenue about RMB 1.0 billion). Since 2020 Dongmu has actively restructured, cutting press-and-sinter share from 60 percent to 45 percent, lifting metal soft-magnetic share from 10 percent to 32 percent, and growing MIM from zero to 18 percent. The restructuring lifted gross margin from 18 percent in 2020 to 26 percent in 2025; H1 2026 guidance points to about 28 percent. Dongmu Haimen's metal soft-magnetic phase II went into production in H2 2025, adding 20,000 tons, primarily serving BYD and CATL on-board chargers.

Poco Holding (300811.SZ) is a pure metal soft-magnetic core leader. 2025 revenue about RMB 3.2 billion, net profit about RMB 600 million, gross margin about 36 percent. Over 90 percent of business is alloy soft-magnetic cores, covering iron-silicon, Sendust, Fe-Si-Mg, Fe-Ni-Mo, and amorphous nanocrystalline series, with customers including BYD, CATL, Huawei Digital Power, Sungrow, Inovance, Tesla (via ODMs), Europe's SMA, and Sungrow. The moat is in-house alloy powder: powder-to-core integration drives powder cost down to about 60 percent of external sourcing, and powder formula can be tuned in reverse based on core performance. This is what others cannot match. H1 2026 guidance shows AI server power, 800V on-board chargers, and 800V storage PCS combined raised to about 35 percent of orders, lifting gross margin to about 38 percent.

Yuean Advanced Materials (688786.SH) is the hidden champion of carbonyl iron powder. 2025 revenue about RMB 900 million, net profit about RMB 180 million, gross margin about 38 percent. Carbonyl iron powder about RMB 600 million, gas-atomized alloy powder about RMB 200 million, other specialty powders about RMB 100 million. Carbonyl iron domestic share above 70 percent; in Asia-Pacific it has displaced BASF. Downstream: mainly MIM (Jingyan, Tonglian, FRT are core customers), soft-magnetic cores (supplying Poco and Dongmu), EMI shielding materials, and 3D-printing fines. The Maanshan phase II added 5,000 tons of carbonyl iron in 2025, mainly for consumer-electronics foldable hinge MIM parts and AI server soft-magnetic cores.

Yitong New Materials (871642.NQ, NEEQ-listed and in IPO queue) is the spherical-powder niche leader. 2025 revenue about RMB 800 million, net profit about RMB 120 million, gross margin about 32 percent. Water-atomized iron about RMB 300 million, gas-atomized alloy about RMB 200 million, spherical superalloy and titanium about RMB 300 million. Yitong has PREP spherical powder capacity in Taixing, Jiangsu; a new PREP line in 2025 mainly supplies COMAC C919/C929 titanium structural parts and AECC main engine makers' superalloy powder.

Gaona Aero Material (300034.SZ) is principally in superalloy forgings, but its subsidiary Gaona Powder is the largest domestic spherical superalloy powder supplier. 2025 consolidated revenue about RMB 6.8 billion, with powder business revenue about RMB 600 million and 42 percent gross margin. Gaona's moat is deep binding with the AECC system: integrated supply of powder, wrought superalloy, cast superalloy, and master alloy, customers including AECC main engine makers, COMAC, CASIC, BLT, and Farsoon. The Gaona Powder phase II commissioned in 2025 added 1,000 tons of spherical superalloy capacity, mainly for the COMAC C929 program's domestic Changjiang engine.

GRIPM Advanced Materials (688456.SH) is the scale leader in water-atomized iron. 2025 revenue about RMB 2.8 billion, net profit about RMB 200 million, gross margin about 15 percent. Water-atomized iron about RMB 1.5 billion, battery anode material precursor about RMB 800 million, other specialty powders about RMB 500 million. Cost control and stable downstream binding: three bases in Beijing, Yantai, and Hefei totaling about 100,000 tons of water-atomized iron capacity, customers including Dongmu, FRT, Daika, and Benteler — almost all domestic press-and-sinter makers.

Among the four overseas giants, GKN Powder Metallurgy of the UK is the world's largest in press-and-sinter structural parts: 2025 global revenue about EUR 1.4 billion, 30 to 40 percent share in European and US OEM supply chains. Sweden's Höganäs is the global water-atomized iron leader: 2025 global revenue about EUR 1.2 billion, water-atomized iron capacity about 500,000 tons, China JV Xinghua Hengrui about 60,000 tons. Japan's Sumitomo Electric powder metallurgy division supplies both metal soft-magnetic cores and press-and-sinter structural parts: 2025 segment revenue about USD 1 billion. Japan's Proterial (former Hitachi Metals) is the global leader in amorphous nanocrystalline soft magnetic strip and metal soft-magnetic cores: 2025 magnetic materials segment revenue about USD 1.5 billion.

Looking at the six domestic core firms and four overseas giants together, China's local camp has caught up with and surpassed overseas giants in niches like metal soft-magnetic cores, MIM, and 3D printing powders; but in traditional press-and-sinter structural parts, Dongmu remains only half the size of GKN, and truly globalized supply-chain relationships are still being built.

From an operating-tempo perspective, the six domestic core firms each have distinct growth trajectories. Dongmu is the case of "the established leader pivoting" — between 2018 and 2020 the company faced growth pressure as traditional press-and-sinter slowed, its share price bottomed in March 2020, and management made a decisive strategic shift, increasing investment in metal soft-magnetic cores and MIM. The new growth cycle began in 2021, and the share price has since quadrupled from the 2020 low. This shows that powder metallurgy incumbents that catch downstream pivot windows can reinvent themselves from "old guard" to "new force." Poco represents the "niche leader scaling up" — RMB 400 million revenue at IPO in 2018, RMB 3.2 billion by 2025, seven-fold growth in seven years and a five-fold stock gain. This shows that in a clear-growth niche, pure-play specialists can ride a steep growth curve. Yuean is the "hidden champion" — small but a global top three / Asia top one in carbonyl iron, with margins above 35 percent and extremely sticky customers, largely insulated from cyclicality. Yitong is the "private-equity new force" — focused on the high-barrier niche of spherical powders, with revenue of just RMB 800 million but excellent margins and growth; its 2026 IPO is in the inquiry stage with NEEQ listing expected H1 2027. Gaona is the "SOE-backed platform" — under China Iron & Steel Research Institutes Group, with a comprehensive layout across superalloy powder, wrought superalloy, cast superalloy, and master alloy, customers tightly bound to AECC, less explosive growth but very stable. GRIPM is the "old powder raw material supplier" — under Grinm Advanced Materials Group, the largest supplier of water-atomized iron and battery anode precursors but with lower margins, the classic "scale for profit" model.

The relationships among the six domestic firms are also worth understanding. They are not a simple competitive set but a network of "specialized cooperation + local competition." GRIPM supplies water-atomized iron to Dongmu; Yuean supplies carbonyl iron to Poco; Yitong supplies spherical powders to Gaona and BLT; Gaona supplies superalloy powder to AECC; Dongmu and Poco are competitors in metal soft-magnetic cores while jointly opposing overseas giants. This networked value-chain forms a unique "industry ecosystem" where any one firm's capacity bottleneck or technology breakthrough affects the entire system.

The state of the four overseas giants also deserves elaboration. GKN, acquired by Melrose in 2023, went through internal restructuring, closing several aging plants in Europe and North America and consolidating capacity to core bases in Bamberg (Germany), Pune (India), and the Shanghai periphery (China). 2025 global revenue about EUR 1.4 billion, gross margin about 15 percent, net margin about 7 percent — a typical mature, low-growth company. Höganäs has 2025 global revenue of about EUR 1.2 billion, water-atomized iron capacity of 500,000 tons, and 40+ percent share in Europe and the US. Over the past three years, Höganäs has focused on high-end powders (spherical superalloy, 3D-printing aluminum), with one PREP line and two gas-atomization 3D powder lines already built. Sumitomo Electric's powder metallurgy unit had about USD 1 billion in 2025 revenue, fully laid out across metal soft-magnetic cores and press-and-sinter structural parts, mainly serving Toyota, Honda, Nissan, and Mazda. It also has solid presence in Europe and Southeast Asia. Proterial's 2025 magnetic materials revenue is about USD 1.5 billion — global outright leader in amorphous nanocrystalline strip, with full presence in metal soft-magnetic cores.

In late 2025, overseas giants began to respond visibly to Chinese leaders' rapid rise. GKN announced Q4 2025 expansion of its Suzhou MIM plant in China, mainly serving consumer electronics and medical. Höganäs announced Q1 2026 expansion of gas-atomized alloy powder capacity in Yangzhou, China, primarily serving China EV OBC soft-magnetic cores. Sumitomo Electric announced Q3 2025 collaboration with Chinese local firms on next-generation 800V on-board charger metal soft-magnetic cores. Proterial announced Q2 2026 increased R&D in amorphous nanocrystalline strips. Overseas giants now recognize that the China market has shifted from "follower" to "definer"; without active adaptation to its tempo, they will lose global share.

The HR structure of the six domestic core firms is also worth touching on. Dongmu has six production bases nationwide (Datong in Shanxi; Yantai in Shandong; Shenzhen in Guangdong; Haimen in Jiangsu; Xuchang in Henan; Shenyang in Liaoning) and about 6,000 employees, including about 600 in R&D. Poco has two bases (Heyuan and Huizhou in Guangdong) and about 1,800 employees, with about 200 in R&D. Yuean has two bases (Liyang in Jiangsu and Maanshan in Anhui) and about 1,000 employees, including about 100 in R&D. Yitong has bases in Taixing (Jiangsu) and Hefei, totaling about 800 employees. Gaona is headquartered in Beijing with bases in Beijing, Zhuozhou (Hebei), and Deyang (Sichuan), totaling about 2,500 employees and about 400 in R&D. GRIPM has bases in Beijing, Yantai, and Hefei, totaling about 2,000 employees. By staff structure, Chinese leaders have approached overseas giants in scale and efficiency.

The six domestic core firms' performance in H1 2026 broadly exceeded market expectations. Dongmu's H1 guidance shows revenue up 31 percent YoY, net profit up 52 percent YoY, driven by rapid ramp in metal soft-magnetic cores. Poco's H1 guidance shows revenue up 38 percent YoY, net profit up 62 percent YoY, driven by AI server power supplies and 800V OBC. Yuean's H1 guidance shows revenue up 35 percent YoY, net profit up 45 percent YoY, driven by MIM fine powder and AI server soft-magnetic cores. Yitong, not yet listed, communicates revenue up about 40 percent YoY, driven by spherical titanium and superalloy powder ramps on COMAC C919 localization. Gaona's H1 revenue up 18 percent YoY, driven by steady superalloy forgings, with the powder business growing about 28 percent. GRIPM's H1 revenue up 12 percent YoY, driven by steady water-atomized iron and some volatility in battery anode materials.

Taken together, China's powder metallurgy leaders showed accelerating growth in H1 2026, with industry sentiment near a five-year high. Whether momentum carries into H2 depends on EV sales, AI server shipments, consumer-electronics foldable shipments, and COMAC C919 deliveries. If these downstream indicators sustain current trajectories, full-year results should broadly beat early-year expectations.

Beyond the six listed leaders and four overseas giants, China's powder metallurgy industry has roughly 300 mid-sized plants distributed across the Yangtze River Delta, Pearl River Delta, and Bohai Rim. Their revenue ranges from RMB 50 million to RMB 500 million, each specializing in some niche (a specific press-and-sinter part, a narrow MIM category, or a specific soft-magnetic core size). They are the "waist" of the chain — downstream customers of leaders, localization-target suppliers for overseas giants, and potential acquisition targets for upstart makers. Their combined capacity exceeds that of the listed leaders, but because of fragmentation, they rarely enter capital-market view. Only B2B factory-data platforms can comprehensively map their real state.

Stringing together leaders, second-tier players, and waist plants, China's powder metallurgy forms a relatively stable "pyramid." The tip is six listed companies plus several unlisted niche leaders, accounting for about 40 percent of industry revenue; the middle is some 20 niche-focused firms with revenue of RMB 300 million to RMB 2 billion, also about 30 percent; the base is some 300 plants with revenue of RMB 30 million to RMB 300 million, also about 30 percent. This pyramid matches China's broader manufacturing structure: leaders have brand, technology, and scale; mid-tier players have specialization, customers, and personality; base plants have capacity, flexibility, and local color. The three layers depend on each other and together support the industry's development.

M&A and consolidation will visibly accelerate between 2026 and 2030. Tip firms have incentives to acquire mid-tier firms (to expand niche categories) and base plants (to add capacity); mid-tier firms have incentives to acquire base plants (to scale); base plants have incentives to consolidate among themselves (to scale jointly). Capital-market participation will accelerate the process: industry-investment funds, local government guidance funds, and overseas PE funds have all placed powder metallurgy on their watch lists. The industry should see 10 to 20 significant M&A transactions per year over the next five, driving concentration steadily up.

Overseas giants' China strategy also visibly shifted in 2025 and 2026. The past approach was "export first, localization second"; the new approach is "localization first, export second." The fundamental reason is that the China market is large enough, growing fast enough, and idiosyncratic enough that overseas players must localize deeply to keep participating. GKN ramped Suzhou MIM, Höganäs ramped Yangzhou gas-atomized powder, Sumitomo Electric partners with Chinese firms on new products, and Proterial doubled down on China for amorphous nanocrystalline. This "China strategy pivot" by overseas giants in turn poses a new challenge for Chinese locals: they must keep a competitive edge against both domestic peers and locally entrenched overseas players.

Chapter 5: MIM Metal Injection Molding — A Small-Parts Revolution in Consumer Electronics, Autos, and Medical

MIM has been the fastest-growing sub-track in Chinese powder metallurgy over the past decade, rising from under RMB 2 billion in 2015 to RMB 8.5 billion by 2025. The real engine behind it is consumer electronics, especially foldable phone hinges.

The hinge of a foldable phone is the product that pushed MIM from a "small track" into a "big track." A mainstream foldable phone hinge holds 15 to 20 MIM parts: main-axis gears, cams, links, sliders, limiters, and damping components. Each part sells at a few yuan to a few tens of yuan; total hinge MIM-part value per phone is RMB 300 to 600. After Huawei launched its first-generation Mate X in 2019, the supply chain matured rapidly; 2025 global foldable phone shipments approach 40 million units, corresponding to a hinge MIM market of about RMB 12 billion. Chinese makers dominate the supply chain: Jingyan, Tonglian, FRT, and Dongmu's MIM subsidiary together account for over 70 percent.

The technical evolution of foldable hinges directly pulled MIM forward. The first generation (2019 to 2021) used 17-4PH stainless; total hinge weight including MIM parts was about 30 grams, with folded-phone thickness about 17 millimeters. The second generation (2022 to 2023) shifted en masse to titanium-alloy MIM (mainly TC4 / Ti-6Al-4V), cutting weight 40 percent (hinge MIM parts about 18 grams) and folded thickness to about 14 millimeters. The third generation (2024 to 2025) adopted TA15 high-strength titanium and "hinge-frame integration," forming hinge and middle frame as a single MIM piece, further reducing weight and thickness to about 10 millimeters folded. The fourth generation (2026 to 2027) is exploring "water-drop hinges" and "titanium MIM + HIP" hybrid processes targeted at tri-fold and rolling-screen formats. This biennial iteration pushed foldable thickness from 17 mm to under 10 mm, driven by continuous MIM breakthroughs in titanium alloys, complex geometry, and hybrid processes.

Apple's foldable roadmap is another industry weathervane. Apple has not launched a foldable yet, but consensus expects a first iPhone Fold in 2026 or 2027. Once Apple enters, foldable shipments will leap from 40 million in 2025 toward 100 million by 2028. Apple's supply-chain bar is famously strict; if Chinese head MIM firms enter Apple, MIM standards across the industry will rise further. Jingyan, Tonglian, and FRT began Apple-chain qualification samples in 2025, with progress closely held.

In smartwatches, Apple Watch's influence is most far-reaching. Since launch in 2015, Apple Watch has developed a mature MIM-part design specification and supply chain. Its case comes in aluminum, stainless, titanium, and ceramic versions, with stainless and titanium cases using MIM. The digital crown, side button, and band lugs are MIM parts. A titanium Apple Watch Ultra has about RMB 300 in MIM-part value, far above ordinary smartwatches. Apple's MIM qualification is extremely strict; current suppliers concentrate among Foxconn, Luxshare, and AAC EMS-tier players, with Chinese local MIM houses entering Apple indirectly via these EMS partners.

Domestic smartwatch brands' (Huawei, Xiaomi, OPPO, Honor, vivo) MIM supply chains are more open. Huawei Watch GT, Xiaomi Watch S, and OPPO Watch series cases, crowns, and buttons are mainly supplied by Jingyan, Tonglian, FRT, and Dongmu MIM. Rapid growth of domestic smartwatches (about 60 million units combined in 2025) provides steady downstream demand for domestic MIM houses.

TWS earbud charging-case hinges are another MIM segment. AirPods and domestic TWS charging cases use magnetic-snap lid designs; the hinge between lid and body needs to be both flexible and durable. A pair of TWS charging-case hinge MIM parts costs RMB 10 to 30. Global TWS shipments in 2025 are about 300 million pairs, corresponding to RMB 6 billion in MIM. The niche is competitive, with limited unit pricing and lower margins.

AR/VR headset frame connectors, hinges, and temples are also MIM applications. Meta Quest, Apple Vision Pro, PICO, and Huawei VR Glass precision metal components added up to about RMB 2 billion in 2025. AI glasses (Meta Ray-Ban, Rayneo X3, Mi Glasses, Huawei AI glasses) hinge MIM parts are the hottest new scene in 2025–2026: AI glasses are small, light, and worn for long durations, demanding precision, durability, and lightness from hinges. AI-glasses temple-hinge MIM parts cost RMB 20 to 50, with strong margins. Tonglian and Jingyan entered AI glasses MIM supply in 2025, with about RMB 500 million in revenue contribution expected in 2026.

Inside automotive turbochargers, MIM has the earliest and most mature application of the automotive segment. A turbocharger has a turbine end (hot side, up to 900°C) and a compressor end (cold side). On the turbine end, variable-geometry blades, nozzle rings, and shaft bushings — needing heat resistance, corrosion resistance, and complex shape — heavily adopt MIM in Inconel and other superalloys. Per-turbo MIM-part value is RMB 30 to 100, supplying Honeywell (now spun off as Garrett Motion), BorgWarner, Bosch, MHI, and IHI. Chinese makers' share in turbo MIM exceeds 50 percent, led by Tonglian and FRT.

Inside transmissions, MIM is also applied. Synchronizer rings, shift forks, oil-pump impellers, and hydraulic control valves — complex shapes in big batches with wear-resistance requirements — fit MIM. Per-transmission MIM-part value is RMB 20 to 80. Chinese makers mainly supply CNHTC, Weichai, Yuchai, Great Wall, GAC, and SAIC.

In safety systems, MIM parts include seat-belt retractor pawls, ABS valve cores, ESP valve cores, and airbag igniters. Safety parts demand extreme reliability and long qualification cycles, yielding higher margins. Chinese makers mainly supply Autoliv, ZF TRW, Continental, and Visteon.

In automotive electronics, MIM parts are growing fast: BMS connectors, internal contacts of high-voltage relays, and precision parts of electronic control modules — demanding high precision, high reliability, and complex shape — fit MIM. As EV penetration rises, this segment will grow. A mainstream EV's electronic-system MIM parts are worth RMB 50 to 150.

Stepping back to the MIM industry overall, the core drivers between 2026 and 2030 are four: consumer electronics (foldable hinges + smartwatches + AI glasses + AR/VR), robotics (humanoids + collaborative + service), medical (surgical-robot accessories + interventional devices + orthopedic implants), and automotive (EV-specific electronic and battery systems + traditional auto turbos and transmissions). Combined, MIM's share of Chinese powder metallurgy will rise from 18 percent in 2025 to about 25 percent by 2030, with industry size growing from RMB 8.5 billion in 2025 to over RMB 20 billion by 2030.

Medical MIM is the third-largest application of MIM and the highest-margin one. Dental orthodontic brackets, surgical instruments, endoscope end-effectors, and orthopedic implant accessories are all MIM parts. MIM's medical edge: ability to handle biocompatible alloys (316L, 17-4PH, titanium alloys, cobalt-chrome) and shape complex small parts. A dental bracket sells at RMB 20 to 50 with margins above 50 percent. Tonglian, Jingyan, and FRT have medical MIM businesses with combined 2025 revenue of about RMB 800 million. Globally, Indo-MIM holds the largest medical MIM share.

Medical MIM has extremely high qualification thresholds. Beyond ISO 13485, products need FDA 510(k) (US), CE MDR (EU), and NMPA registration (China) per category, taking 6 months to 2 years and costing from a few hundred thousand to several million yuan. Once qualified, products supply stably for 10+ years; switching cost is prohibitively high. This barrier underpins medical MIM's 50+ percent margins.

Medical MIM also features slow iteration and long lifecycles. A dental bracket design, surgical tool, or implant accessory, once qualified, can stay on market 10 to 20 years. Long lifecycles mean MIM makers' medical revenue is extremely persistent with virtually no "product obsolescence" risk — another reason medical MIM warrants a valuation premium.

Medical MIM growth between 2025 and 2030 comes mainly from three places. First, surgical robot accessories. Intuitive Surgical's da Vinci plus domestic MicroPort, Konda, Sizhe surgical robots all use MIM precision parts; per-robot MIM is RMB 3,000 to 5,000. Second, interventional devices. Stent delivery systems, TAVR systems, and neuro-interventional devices increasingly need small precision metal parts. Third, assisted reproduction and diabetes management devices. ICSI needles, insulin-pump precision valves, and CGM sensor parts are all MIM applications.

MIM process innovation deserves attention. First, "Two-Component MIM" injects two feedstocks into different mold positions, forming a composite part in one shot. A phone hinge can be partly titanium and partly stainless, eliminating later assembly. This entered small-volume production in 2025. Second, "Micro MIM" — single parts at 0.01 to 0.1 grams for micro medical devices, micro electronic connectors, and miniature watch movements. Third, "metal 3D printing + MIM" hybrid — using metal 3D printing for complex preforms then MIM-densifying. Still lab stage but with great potential. Fourth, "High Entropy Alloy MIM" — making HEAs into MIM parts to achieve strength, toughness, wear resistance, and corrosion resistance beyond conventional alloys. Targeted at aerospace and deep-sea extremes.

MIM also has overlooked details. First, mold "flash" and "parting lines." Because MIM is injection-based, flash forms at parting lines and needs deburring. Parting-line placement directly affects appearance and performance. Second, debinding's "skin" issue. Binder removes from outside in; too-fast debinding forms a "skin" hampering sintering, hence the typical 24- to 72-hour gentle debinding cycle. Third, sintering "warp." Green compacts shrink 15 to 20 percent during sintering; if furnace temperature is uneven or compact placement is uneven, parts warp. Controlling furnace uniformity and placement is a core MIM engineering skill. Fourth, post-processing dimensional precision. Sintered MIM parts have dimensional precision of 0.5 to 1 percent; high-precision customers (medical, semiconductor equipment) need under 0.1 percent, requiring precision machining or grinding after sintering. These details determine a real MIM maker's competitiveness.

MIM also has the "customer co-design" feature. Unlike "supplier-makes-to-drawing," MIM requires deep supplier-customer collaboration at the design stage because of geometric complexity and process specifics. Customers propose functional needs; suppliers propose optimal geometry given MIM constraints (debinding/sintering shrinkage, minimum wall thickness, draft angles). The two together set final design. This "co-design" tightly binds head MIM suppliers to large customers, with prohibitive switching cost. This is why head MIM firms (Jingyan, Tonglian, FRT) supply foldable hinges and smartwatches with long stability.

MIM also has a unique "trial-error cycle." A new product typically needs dozens to hundreds of experiments — each tuning mold, feedstock ratio, debinding, or sintering parameters — to observe effects on final performance. Cycles take 3 to 6 months and consume serious R&D resources. That is why leading MIM houses keep "R&D workshops" with small injection machines, debinding furnaces, sintering furnaces, and inspection gear for fast trial-error. Workshop scale and efficiency directly determine a MIM firm's product-development capability.

Chapter 6: Soft Magnetic Powder Cores Explode — EV OBC, PV-Storage, AI Servers

If MIM is the small-parts revolution sparked by consumer electronics, metal soft-magnetic powder cores are the magnetic-core revolution sparked by EVs.

Soft-magnetic materials are foundational components in any power-electronics circuit. An inductor, a transformer, a common-mode choke, a PFC inductor, an LLC resonant inductor — all need soft-magnetic cores. The traditional soft-magnetic material is ferrite (MnZn or NiZn), mainstream since the 1950s in consumer-electronics, industrial, and power-grid transformers. But ferrite has an inherent flaw: saturation flux density of only 0.4 to 0.5 tesla, prone to saturation in high-current, high-power-density scenarios. Metal soft-magnetic powder cores (iron-silicon, Sendust, Fe-Si-Mg, Fe-Ni-Mo, amorphous nanocrystalline) have saturation flux density of 1.0 to 1.8 tesla, two to four times ferrite. This means at the same power density, metal powder cores can be a third to half the volume of ferrite.

EV on-board chargers (OBC) are the first large-scale application of metal soft-magnetic cores. A BEV's OBC needs 5 to 8 magnetic cores (PFC, LLC resonant, common-mode, differential-mode), about 1 to 2 kilograms in total weight and RMB 200 to 500 in value. 2025 global EV production is about 15 million units, equating to RMB 6 billion of OBC powder cores. From 3.3 kW OBC in 2020 to 6.6 kW in 2022 to mainstream 11 kW and high-end 22 kW bidirectional OBC in 2025, every doubling of power density doubles powder-core demand. In 2025, BYD Han, the China-built Tesla Model S refresh, and the Huawei AITO M9 standardized 22 kW bidirectional OBC; this is the main axis of metal soft-magnetic core demand growth for the next three years.

Photovoltaic inverters are the second-largest application. Centralized PV inverters (Huawei, Sungrow, TBEA) and string inverters (Huawei, Sungrow, Ginlong, Goodwe) use large amounts of metal soft-magnetic cores in boost and filter inductors. A string inverter (commonly 100 to 350 kW) holds 5 to 20 kilograms of metal soft-magnetic cores worth RMB 1,000 to 5,000. 2025 global PV new installations are about 580 GW, equating to about RMB 3 billion in PV inverter powder cores.

Energy storage converters (PCS) are the third-largest application. Storage PCS power ratings exceed PV inverters (typically 500 kW to 3 MW), driving even larger per-unit powder-core consumption. 2025 global storage new installations are about 220 GWh, equating to about RMB 2 billion in PCS powder cores. This will become the fastest-growing incremental source for powder cores between 2026 and 2030.

AI server power is the surprise breakout of 2025. The GPU compute boom drives single AI server power from a traditional 1 kW to 10 kW+ (NVIDIA GB200 per-rack power is 120 kW). High-power-density server power (800V DC bus; 3 to 5.5 kW per module) consumes 5 to 10× the powder cores of traditional servers. A 5.5 kW power module holds RMB 100 to 200 in metal soft-magnetic cores. 2025 global AI server shipments are about 1.2 million units, equating to about RMB 2.5 billion in powder cores. Poco's H1 2026 guidance specifically calls out AI-server-power orders, signaling this new scene has moved from sample qualification to volume ramp.

Wind power converters and industrial UPS are the fifth application. Onshore wind is now mostly 6 to 10 MW; offshore exceeds 15 MW; their converters use large amounts of metal soft-magnetic cores for grid filter inductors. A 6 MW wind converter holds 20 to 30 kilograms of soft-magnetic cores worth RMB 3,000 to 5,000. 2025 global new wind installations are about 130 GW, equating to about RMB 1 billion in wind converter powder cores. Industrial UPS in 2025 also began switching from ferrite to metal soft-magnetic cores as data centers and precision manufacturing demanded higher reliability and power density that ferrite no longer satisfied.

5G base-station power and 6G pre-research base-station power are also applications. One 5G base-station power module holds RMB 50 to 100 in metal soft-magnetic cores. By 2025 China's installed 5G base stations are about 4 million with annual additions about 600,000, equating to about RMB 500 million. 6G pre-research base stations demand higher power density; if mass deployment begins in 2028, powder-core demand could grow 3 to 5× over 5G.

Data centers' backup power (UPS) and distributed power (DPS) also face AI-era upgrades. Traditional data center UPS uses lead-acid + inverter with ferrite. AI data centers — single-rack power rising from 10 kW to 100+ kW — need lithium battery + high-power-density inverter architecture using metal soft-magnetic cores extensively. An AI data center's backup-power system holds RMB 500,000 to 2 million in cores. 2025 global new AI data centers number about 100; 2026 to 2030 likely 100 to 200 per year, equating to RMB 2 to 4 billion in powder cores annually. This is a fast-growing new scene.

Medical equipment is another growth scene. MRI, CT, X-ray, ultrasound diagnostic, and radiation therapy equipment all need high-performance power supplies. They demand extreme reliability, accuracy, and low noise — heavy use of metal soft-magnetic cores. A high-end MRI's power system holds RMB 200,000 to 500,000 in cores. 2025 global medical equipment power supply is about RMB 10 billion, of which metal soft-magnetic cores are about RMB 1 billion, with 15 percent CAGR expected over the next five years.

Industrial automation and robotics also use soft-magnetic cores. Variable-frequency drives, servo drives, robot controllers, and industrial UPS all consume large amounts. 2025 China industrial automation equipment is about RMB 20 billion in annual sales (corresponding RMB 2 billion of cores); global about RMB 60 billion (about RMB 6 billion). As industrial automation spreads and robot numbers grow, this niche will keep expanding.

EV charging piles are another growth scene. Fast chargers (60 to 360 kW) and superchargers (480 kW+) demand power density that drives extensive use of metal soft-magnetic cores. A supercharger holds RMB 1,000 to 3,000 in cores. 2025 China new charging-pile installations are about 5 million (mainly slow), equating to RMB 1.5 billion. If the fast and supercharger share rises from 30 percent in 2025 to 60 percent in 2030, the market doubles.

PV+storage integration also pulls powder cores. Integrated systems combine PV inverter and storage PCS, demanding even more cores. 2025 global new PV+storage installations are about 50 GWh, equating to about RMB 1 billion in cores. This niche will sustain growth from 2026 to 2030.

Adding up all these niches, metal soft-magnetic core downstream demand is evolving toward "diversification, higher power, higher frequency." Diversification means the industry no longer depends on a single niche (auto OBC); higher power means per-application core consumption rises; higher frequency means core technology bar rises. Layered together, metal soft-magnetic cores are among the highest-certainty high-growth sub-tracks in powder metallurgy.

Amorphous nanocrystalline soft-magnetic strip is a close cousin of metal soft-magnetic cores but on a completely different process route. Amorphous nanocrystalline rapid-cools (cooling rate over 1 million degrees Celsius per second) iron- or cobalt-based alloy melt into amorphous or nanocrystalline-structured strip, then winds it into cores. Strengths: extremely high permeability, very low loss; suited to high-power transformers and high-frequency inductors. Largest Chinese makers: Advanced Technology & Materials Co., Ningbo Yunsheng, Qingdao Yunlu, Foshan Zhongyan Amorphous. Largest overseas: Proterial (former Hitachi Metals). Amorphous nanocrystalline competes with metal soft-magnetic cores in many scenarios; because the process is harder and the unit price higher, it goes to high-end scenarios (VFDs, servo drives, high-power welders) while metal soft-magnetic cores serve mid-range (OBC, PV inverters, storage PCS). The two are both competitive and complementary.

The design and manufacturing of soft-magnetic cores has several engineering details often overlooked by outsiders. First, fine-grained alloy composition. The same "iron-silicon" with different Si content (6.5%, 9.5%, 10.5%) has different permeability, loss, and thermal characteristics. Downstream power-supply OEMs reverse-specify the precise alloy composition based on working frequency, temperature range, current amplitude, and waveform. This requires powder-core firms to provide "custom-formulation" services, not just a few standard products. The capability to tailor recipes is the key separator of head and second-tier firms.

Second, precise powder-binder mixing. In a finished powder core, metal accounts for 75 to 85 percent by volume and binder (epoxy, phenolic, silicone) accounts for 10 to 20 percent. Binder type, content, and curing process directly affect mechanical strength, thermal stability, and insulation. As downstream operating temperature requirements rise from 100°C historically to 155°C now (some new scenes demand 200°C), binder systems must keep up.

Third, shape and size optimization. Common core shapes include toroidal, E-type, UU-type, pot-type, and planar. Each suits a different inductor design, cooling strategy, and assembly mode. Power-supply OEMs often demand non-standard custom shapes. Such customization involves mold development, pressing adjustment, and sintering optimization — typical cycles of 2 to 6 months. Leaders, with strong mold capability and rich process experience, can compress cycles to 2 to 3 months, the key to winning premium orders from large customers.

Fourth, secondary processing. Sintered cores need secondary processing per downstream needs: precision cutting (toroidal cut into halves for E-type), surface coating (better insulation or corrosion resistance), winding (direct production of finished inductors). Secondary-processing quality directly affects core reliability in finished products. Leaders integrate secondary processing in-house; second-tier firms may outsource.

Fifth, end-to-end quality control. A qualified high-end soft-magnetic powder core needs strict process control across every step — alloy powder production, passivation, mixing, pressing, sintering, inspection. A leading firm's inspection equipment list typically includes particle-size analyzer, XRD, SEM, magnetic performance tester, loss tester, thermal cycling chamber, vibration tester, life tester. Combined investment is RMB 10 to 20 million. Inspection capability directly determines product stability and customer trust.

In 2025 and 2026, soft-magnetic cores saw several notable new technology developments. First, R&D of "amorphous powder cores" and "nanocrystalline powder cores." Grinding amorphous nanocrystalline strip into powder then pressing into cores combines amorphous nanocrystalline's low loss with powder cores' flexible shapes. The route is in pilot stage at AT&M and Poco, with mass production expected 2027. Second, rapid development of "Soft Magnetic Composites" (SMC). SMC, promoted by Höganäs since the 2000s, coats iron powder particles with an insulating layer then presses into 3D-magnetic-circuit cores. SMC offers unique advantages for motor stators, rotors, and automotive transformers. Höganäs' Somaloy is the global leader; Chinese local firms are catching up. Third, development of "high Bs" (high saturation flux density) alloys. Next-gen auto OBC and AI server power demand higher Bs from cores (rising from current 1.6 tesla toward 1.8 or even 2.0 tesla), requiring new alloy systems. Fourth, "high-frequency low-loss" alloy development. AI server power frequency is rising from 100 kHz toward 500 kHz or even MHz, requiring soft-magnetic cores to keep loss extremely low at high frequency. These developments mark leaders' technology reserves for 2026 to 2030.

Chapter 7: Platform Perspective — Reverse-Indexing Powder Metallurgy Downstream Plants by Process

Spreading the full powder metallurgy value chain — from powder to parts, from parts to downstream OEMs — there are hundreds of mid-sized plants in between. Some only do press-and-sinter, some only MIM, some only soft-magnetic cores, some only 3D printing metal powder; some specialize in one specific alloy powder passivation, some in MIM secondary precision machining. Looking only at the top listed companies, this business has just a dozen players; broken down by process and niche, it has 400 to 500 fragmented plants.

This "concentrated at the top, fragmented at the bottom" structure is exactly where platform-type tools can add value. Tianxia Gongchang is a B2B platform of 4.8 million in-production factories. Unlike business-registry tools, which focus on corporate-level legal-entity relations, equity penetration, and risk alerts, this platform indexes factories' real production capability by process, material, product, customer qualification, capacity, and region — letting upstream material providers and downstream OEMs start from concrete questions like "I need a factory that does titanium MIM," "I need a factory with PREP spherical-powder capacity," "I need a factory making EV OBC soft-magnetic cores" — directly landing on candidate lists.

For powder metallurgy, this "process-reverse-indexing" capability has several concrete uses.

First, EV OEMs' second-tier supplier sourcing. An OBC module is integrated by a Tier 1 (Inovance, Sungrow, Infineon, Bosch); the Tier 1 doesn't make soft-magnetic cores itself and sources from second-tier suppliers. When a Tier 1 wants domestic substitution or backup, it needs a quick list of all factories that can make iron-silicon soft-magnetic cores, Sendust cores, or amorphous nanocrystalline cores, then filter by capacity, customer experience, and region to three to five candidates for sampling.

Second, foldable phone brands locking in backup capacity. Phone brands depend heavily on a few hinge MIM suppliers (Jingyan, Tonglian, FRT); when leads run short, they need fast access to backup plants for titanium MIM, stainless MIM, MIM hinges, or phone hinges.

Third, COMAC, AECC, and CASIC sourcing qualified 3D printing metal powder and printed-part plants. Qualification thresholds are extremely high — needing both aerospace-grade powder capability and SLM/EBM printing capability. On the platform you can filter by 3D printing titanium powder, superalloy powder, spherical titanium powder, and selective laser melting keywords.

Fourth, overseas brands entering China seeking contract manufacturers. A German power brand wanting local PCS soft-magnetic inductors needs soft-magnetic core inductors, common-mode inductors, or PFC inductors plants in China, filterable by coastal proximity, export experience, ISO and UL certifications.

Fifth, powder material providers expanding downstream. Powder makers (GRIPM, Yuean) need rapid lists of water-atomized iron powder-consuming press-and-sinter plants, carbonyl iron powder-consuming MIM plants, and spherical powder-consuming 3D printing plants. This is what material providers most practically need.

Sixth, investors' and brokers' chain research. A buy-side analyst writing a powder metallurgy deep-dive needs to inventory all powder metallurgy, press-and-sinter, metal injection molding, and metal soft-magnetic plants nationwide by region, scale, and customer mix. What used to take a team of ten analysts two weeks of plant visits can now be done in hours, then visited on the ground.

The platform does not replace professional judgment but compresses "finding factories" from weeks to hours and pushes research granularity from "top public companies" down to "all candidate plants." This is the value of data infrastructure.

Unpacking the actual operational flow of "process reverse-indexing" reveals several details. Step one is identifying process keywords. Process naming varies by company, region, and niche; multiple synonymous keywords need to be searched together. For example, "metal injection molding," "MIM," and "precision MIM parts" cover non-identical sets of factories; the union must be taken. "Metal soft-magnetic core," "alloy soft-magnetic core," and "soft-magnetic inductor" also intersect partially. Linking these synonyms yields a comprehensive candidate list.

Step two is secondary filtering by region, scale, customer qualification, and capacity. Powder metallurgy qualification systems are complex: ISO 9001 (general quality), IATF 16949 (automotive), ISO 13485 (medical), AS 9100 (aerospace), CQI-9 (heat treatment for autos), Nadcap (aerospace special processes), UL (power safety). Each downstream OEM has different requirements; candidate plants must be filtered by target customer's qualification needs. On the platform, certifications are standard fields, directly filterable.

Step three is RFQ and sample evaluation. After narrowing to three to five candidates, send RFQs (Request for Quotation) and sample requests to each. Powder metallurgy quotes typically include mold cost (MIM needs dedicated tooling, RMB 100k to 1M), unit price, MOQ, and lead time. Sample evaluation covers dimensional accuracy, performance, and life. This is the core of sourcing; the platform doesn't make the technical call but narrows the candidate range and establishes first-hand contact, sharply reducing trial-error.

Step four is regular supplier evaluation and dynamic updates. Once selected, suppliers' capacity, customer mix, technical capability, and compliance must be updated annually. The platform updates this dynamically; buy-side procurement teams can periodically revisit supplier pages to track changes. If a major customer is lost, capacity swings, or compliance issues emerge, the strategy can be adjusted promptly.

Step five is building an industry-wide global view. Beyond specific sourcing, the platform also helps establish a global view of the industry. Want to see "national distribution of all titanium MIM factories," "national distribution of all metal soft-magnetic core factories," "national distribution of all spherical superalloy powder factories"? Keyword + geographic visualization shows it at a glance. Such a global view is valuable for buy-side researchers, industry capital, M&A buyers, and government industrial planners alike.

Data infrastructure's core value isn't the data itself but data's "searchability, filterability, and connectability." A "concentrated at top, fragmented at bottom, process-divided" industry like powder metallurgy is exactly where data infrastructure adds the most value. Transforming the industry from "by relationship" to "by data," lowering sourcing granularity from "top public companies" down to "hundreds of mid-sized plants" — this is a concrete embodiment of industry digitalization.

Beyond downstream OEM sourcing, data infrastructure has several extended applications in powder metallurgy. One is financial institutions' supply-chain finance. Commercial banks and supply-chain finance firms need quick reads on a plant's capacity, customer mix, orders, and compliance when financing small powder metallurgy plants. The traditional way is self-reporting plus on-site due diligence — slow and prone to inflation. Plant data on the platform enables banks to make initial calls quickly, sharply cutting loan-approval cycles.

Two is industry investment funds' target sourcing. PE and industry funds searching for M&A or investment targets need fast lists of plants in a revenue range (say RMB 300 million to RMB 2 billion), in a specific niche (e.g., spherical titanium powder), with a specific customer base (e.g., already in COMAC or AECC). What once took an investment-bank chain team weeks now does initial filtering in hours.

Three is government agencies' industrial policy formulation. Local-government investment promotion and MIIT industrial policymakers need national-scale distribution, capacity, technology, and employment figures for a given niche. Previously gleaned through plant visits or sampling surveys — limited precision and slow. With the platform's plant data, governments can do near-real-time inventories, providing data backbone for precise policy.

Four is industry associations' and media's industry research. The CGMCIA Powder Metallurgy Branch, the China Additive Manufacturing Industry Alliance, the China Power Supply Society, plus financial and industry media all need industry-wide data support. The platform's plant data provides a more comprehensive and accurate base.

These five extended uses together form the complete value picture of "data infrastructure" for powder metallurgy. It is not a single tool but a digital infrastructure empowering the entire industry — connecting supply and demand, lowering transaction cost, and improving matching efficiency. This is the greatest value B2B platform tools can create in vertical industries.

Looking further, powder metallurgy data infrastructure has a unique value: it aggregates real production-capability data scattered across hundreds of plants into analyzable, comparable, decision-ready "industry-wide data." Previously industry-wide data could only be obtained indirectly through association statistics, sample surveys, and expert interviews — lagging, of limited precision, and incomplete coverage. Via a plant data platform, industry-wide data can be presented near-real-time, full-sample, multi-dimensional. This presentation capability provides a new data base for industry research, policymaking, and investment decisions. Over the next five to ten years, this infrastructure will become an indispensable "industry OS" for powder metallurgy, empowering value-chain operations.

Another point worth noting: data infrastructure and the industry's existing professional services (associations, broker research, consultancies, third-party testing labs) are not substitutes but complements. Professional services offer deep judgment and bespoke analysis; data infrastructure offers comprehensive data and efficient retrieval. Together they deliver complete value to industry users. This "data + professional services" dual-engine model is the correct path of industry digital upgrade.

Chapter 8: 3D Printing Metal Powder — Superalloy + Titanium + Aluminum

3D printing metal powder is the highest-margin and deepest-barrier sub-track in powder metallurgy. The chain comprises "powder, equipment, printed parts, post-processing." Chinese firms are laid out in all four segments, but real domestic substitution mainly occurs in "powder" and "equipment."

Superalloy powder is the largest 3D-printing category. Downstream: aero-engine additive manufacturing. Typical products: fuel nozzles, turbine blade cooling channels, combustor liners. GE's LEAP engine fuel nozzle is the classic case: a complex nozzle originally assembled from 20 parts is now 3D-printed as one piece, with 25 percent weight reduction and 5× life. COMAC's C919 also has 3D-printed parts in its CFM LEAP-1C engine. The CJ-1000A and CJ-2000 Changjiang domestic engines plan over 15 percent 3D-printed parts in pre-research. A Changjiang engine uses 200 to 300 kilograms of superalloy powder at RMB 600,000 to 1,000,000 per ton — powder cost alone runs into hundreds of thousands.

Chinese makers of aerospace-grade superalloy powder are Gaona, Yitong, and AVIC Mete. Gaona is most deeply tied to AECC, integrating powder, wrought superalloy, cast superalloy, and master alloy. Yitong enters via PREP, mainly supplying COMAC C929 and CASIC solid rocket motors. AVIC Mete sits within the AVIC system, mainly serving internal demand.

Titanium powder is the second-largest 3D-printing category. Downstream splits two ways: aerospace and medical. Aerospace mainly covers COMAC C919/C929 fuselage titanium structures, landing gear, and load-bearing beams. C919 titanium content is 9.5 percent of fuselage weight, with some critical parts under 3D-printing evaluation. Medical mainly covers orthopedic implants: titanium hip joints, intervertebral cages, maxillofacial restorations. 3D-printed titanium orthopedic implants' advantage is creating bone-trabecula-like porous structures promoting bone in-growth, shortening recovery. 2025 global 3D-printed titanium powder market is about RMB 3 billion, with China at about 30 percent. Chinese makers: Gaona, Xi'an Sailong, and Yitong. Xi'an Sailong, incubated from Northwestern Polytechnical University, has the highest medical share.

Aluminum powder is the fastest-growing 3D-printing category. Downstream: aerospace satellite structures, high-end consumer electronics (Huawei Mate back covers and middle frames), and inserts in EV gigacasting. The challenge: aluminum is reactive and oxidizes easily, requiring inert atmospheres throughout production and printing. AlSi10Mg and AlMgScZr are mainstream alloys. Chinese makers: BLT (in-house), Farsoon (in-house), Yitong, and GRIPM. 2025 global 3D-printing aluminum powder market is about RMB 1.5 billion, growing 35 percent.

Stainless steel powder and tool-steel powder are large-volume but lower-margin 3D-printing categories, mainly used in tool inserts (conformal cooling), medical surgical instruments, and consumer-goods molds. One ton of 3D-printing 316L stainless powder is about RMB 200,000, with 20 to 30 percent margins.

Equipment is the downstream complement to 3D-printing metal powder. Largest Chinese SLM makers: BLT (688333.SH) and Farsoon (688433.SH), with about 1,500 SLM units shipped in 2025, about 35 percent global share. Largest overseas: Germany's EOS, SLM Solutions (acquired by Nikon), 3D Systems of the US, and Renishaw of the UK. EBM globally has only Sweden's Arcam (acquired by GE Additive) and Tsinghua-backed Qingyan Zhishu in China. Equipment and powder are tightly bound: using EOS equipment requires EOS-recommended powder parameter packs; using BLT equipment requires BLT-recommended powders. This "equipment-powder-process" lock-in is the moat of 3D-printing metal powder.

In H1 2026, two new trends emerged in China's 3D-printing metal powder. One is acceleration of superalloy powder localization on the COMAC C929 program; AECC's next-gen Changjiang engine is expected to grow powder usage 2 to 3× over LEAP. Two is AI server liquid-cooling cold plates beginning to use 3D-printed aluminum; some cold plates in NVIDIA GB200 NVL72 racks are now evaluating 3D-printed aluminum alloys. Both new scenes will sustain superalloy and aluminum powder demand from 2026 to 2030.

3D-printing metal powder technical specifications are stricter than traditional powder metallurgy. SLM powder requirements: tight 15–53 µm distribution; sphericity >0.95; flow <25 s/50g; apparent density >4.5 g/cm³; oxygen <100 ppm (titanium <50 ppm); clean surfaces, no satellites. EBM is stricter: 45–105 µm tight distribution; sphericity >0.97; flow <22 s/50g; oxygen <80 ppm; good electrical conductivity (EBM electron beam needs conductive powder). Only about 20 firms globally meet these specs; in China only six: Gaona, Yitong, Xi'an Sailong, AVIC Mete, BLT (in-house), Farsoon (in-house).

3D-printing metal powder cost structure is also distinctive. An SLM machine consumes about 1–5 tons of powder per year; an EBM machine 5–15 tons. At RMB 800,000 per ton superalloy powder, an SLM machine's annual powder cost is RMB 800,000 to RMB 4 million; at RMB 900,000 per ton titanium powder, an EBM machine's annual powder cost is RMB 4.5 million to RMB 13.5 million. Powder accounts for 30 to 50 percent of equipment operating cost and 10 to 20 percent of final part price. This "equipment + powder" double-high cost means 3D-printed parts mainly serve high-value aerospace and medical, unable to displace casting, forging, and machining in low-end automotive and consumer electronics.

With scale, 3D-printing metal powder has plenty of cost-down room. In 2018, superalloy powder was about RMB 2 million per ton; in 2021 about RMB 1.3 million; 2023 about RMB 1 million; 2025 about RMB 800,000; H1 2026 stable at RMB 800,000. By 2030 Gaona and Yitong capacity expansion and process optimization should bring superalloy powder below RMB 500,000 per ton. Titanium powder is dropping faster: about RMB 2 million per ton in 2018, RMB 1.2 million in 2023, RMB 800,000 H1 2026, expected below RMB 400,000 by 2030. Aluminum powder, with many new entrants, fell from RMB 800,000 per ton in 2021 to RMB 400,000 in H1 2026, expected below RMB 200,000 by 2030. Falling powder costs in turn open application space in new scenarios.

Another trend worth noting in 3D-printing metal powder is "specialty powder." Early on, powders were "general-purpose" — one powder served many applications. As applications fragmented, demands diverged. Aerospace superalloy powders (Inconel 718, 625, Hastelloy X, CM247LC) and medical orthopedic titanium powders (TC4 ELI, TA15) differ entirely in composition, particle distribution, and surface treatment. Aerospace lightweight aluminum powders (AlSi10Mg, AlMgScZr) and mold-insert tool steels (H13, CPM10V) likewise. Since 2025, leading powder makers have launched application-specific "specialty powders," with margins 10 to 20 percent above general-purpose. This signals the industry shifting from "scaling" to "specialization."

3D-printed metal parts' post-processing is another underestimated value-chain node. SLM-printed parts have surface roughness Ra 5–15 µm, far above Ra 0.4–1.6 µm in machining; internal microcracks and pores leave density at 98–99 percent; residual stress requires relief. Hence SLM parts typically undergo HIP densification, stress-relief heat treatment, precision machining, surface polishing, and inspection. An SLM machine needs a full post-processing suite — HIP furnace, heat-treatment furnace, five-axis machining center, CMM, X-ray CT. Post-processing is 30 to 40 percent of finished-part cost — as important as powder. Largest Chinese 3D-printing post-processing service providers: AVIC Mete, BLT, Farsoon, and specialized third-party service firms. This value-chain node is also growing fast.

Powder recycling for 3D printing is another concern. Both SLM and EBM melt powder layer by layer in a powder bed, but only 10 to 20 percent of powder is melted into parts; 80 to 90 percent is "unmelted powder." Theoretically recyclable, but during recycling powder gets partly oxidized, oxygen content rises, sphericity drops, distribution changes. Pure recycled powder noticeably degrades part quality. Industry practice is to mix recycled and new powder at a set ratio (commonly 20 percent recycled + 80 percent new), balancing cost and quality. Optimizing recycling (screening, deoxidation, surface activation) is a key cost-down tech. Chinese leaders are investing heavily in this area.

3D-printed metal parts have specific aerospace applications. First, aero-engine parts. LEAP fuel nozzles, PW1000G hot-end parts, Rolls-Royce Trent XWB static parts. China's Changjiang series engines (CJ-1000A, CJ-2000) plan over 15 percent 3D-printed content in pre-research, covering fuel nozzles, combustor liners, turbine cooling channels, casing stiffeners. A Changjiang engine's 3D-printed parts total RMB 10 to 20 million. Second, aircraft structural parts. Boeing 787 and Airbus A350 use 3D-printed titanium brackets and connectors. COMAC C919 is evaluating 3D printing for some critical titanium-alloy parts. Third, satellite structures. SpaceX Starlink, China's Hongyan constellation, Galaxy Space LEO sats — all heavily use 3D-printed lightweight titanium and aluminum structures. A LEO communication satellite's 3D-printed parts are RMB 500,000 to 1 million. Fourth, commercial space rocket engines. SpaceX Raptor heavily uses 3D-printed complex turbopumps, chambers, nozzles. China's Space Pioneer, Galactic Energy, iSpace, LandSpace, Deep Blue Aerospace likewise. A mid-size liquid rocket engine's 3D-printed parts: RMB 5 to 20 million.

3D-printed metal parts also see broadening medical use. First, orthopedic implants. 3D-printed titanium hip joints, intervertebral cages, vertebrae replacements, maxillofacial restorations — porous structures like bone trabeculae, promoting in-growth, shortening recovery. The killer medical app. China's Chunli, AK Medical, Double Medical have adopted 3D printing. A 3D-printed titanium hip joint sells for RMB 10,000 to 30,000, with RMB 1,000 to 3,000 in titanium powder cost. Second, personalized dental implants and surgical guides. Third, custom surgical tools. Fourth, tissue-engineering scaffolds. The most frontier medical application.

3D-printed metal parts also have important mold applications. Traditional mold cooling channels are drilled — straight lines only, with limited cooling efficiency. 3D-printed molds enable "conformal cooling channels" exactly following cavity shape, with 20 to 50 percent better cooling, 30 to 100 percent longer life, 15 to 30 percent shorter injection cycles. That is why 3D-printed molds are penetrating auto stamping, appliance injection, medical molds rapidly. Chinese 3D-printed mold makers: BLT, Farsoon, Yinxi, Aishikai. A 3D-printed mold insert: RMB 30,000 to 300,000.

3D-printing metal powder equipment-powder synergy also deserves elaboration. SLM print quality is affected by many factors: laser power, scan speed, hatch spacing, layer thickness, protective gas, platform preheat, scan strategy, support design. Each factor interacts with powder properties (particle distribution, flowability, oxygen). Optimal parameter combinations need extensive trial-error per powder. This is why 3D-printing equipment makers and powder makers necessarily form "process-pack lock-ins": each powder corresponds to an optimized process-parameter pack, and customers buying equipment typically also buy the matching powder and process pack. This "equipment-powder-process" integrated sales model raises 3D-printing metal powder's entry barrier further; new entrants can't just provide new powder — they must also offer process optimization.

3D-printed metal parts' design freedom is the route's greatest appeal and its biggest challenge. Releasing that freedom requires entirely new design thinking (Design for Additive Manufacturing, DfAM); engineers can't take traditional designs and 3D-print them directly but must rethink topology, internal channels, thin walls, and lattice structures. Chinese engineers are still catching up on DfAM, but collaborative training between 2023 and 2025 with BLT and Farsoon has produced a cohort of DfAM-capable engineers, mainly in aerospace institutes, medical-device R&D centers, and consumer-electronics design houses. Over the next five years, DfAM will move from "scarce resource" to "industry-standard capability."

3D-printed metal parts' standard system is also rapidly maturing. Internationally, aerospace 3D-printed-part standards are set by ASTM F42, AMS-AM series, and ISO TC 261. China's SAC and the China AM Industry Alliance are accelerating standard-system buildup; in 2025 multiple foundational standards were issued covering metal AM process general rules, materials, and inspection. Standard maturation will normalize production and acceptance, helping lower customers' barriers to accepting 3D-printed parts.

3D-printing metal powder's full-chain competitive landscape is evolving. Beyond the traditional "powder maker vs equipment maker vs print-service provider" trio, several new integration models emerged in 2025. First, "powder-equipment-service integration," typified by BLT and Farsoon — both equipment and powder, plus print services — full chain. Strong synergies, high switching cost; downsides: high fixed cost and demand-swing exposure. Second, "powder maker vertical integration," typified by Gaona — powder, printed parts, end-application validation, but no equipment. Strength: deep tech; downside: equipment progress depends on outsiders. Third, "equipment-service led, powder outsourced," typified by industry-focused print service providers. Strength: light, flexible; downside: powder cost in others' hands. All three coexist in different niches.

3D-printing metal powder's global competition shifted notably in 2025-2026. US GE Additive (now spun off and run independently), Germany's EOS, UK's Renishaw, Sweden's Sandvik Additive Manufacturing, Canada's Equispheres ramped investment in Asia. China's BLT, Farsoon, Gaona, Yitong steadily gained international share, with some products entering European and North American customers' approved-supplier lists. Over the next five years, Chinese makers' global share in 3D-printing metal powder is expected to rise from 15 percent in 2025 to about 30 percent in 2030. A relatively rapid globalization process.

Chapter 9: Capacity Expansion — Dongmu Haimen, Yitong Taixing, Yuean Maanshan

Powder metallurgy is a capital-intensive business with long line-build cycles (typically two to three years from approval to ramp), so existing leaders' capacity pace largely determines supply curves over the next three years. Several core expansions:

Dongmu's metal soft-magnetic powder core phase II in Haimen, Jiangsu commissioned in H2 2025, adding 20,000 tons mainly serving BYD and CATL on-board chargers, Huawei Digital Power PV-storage, and Sungrow PCS. Phase II totals about RMB 1.2 billion; with it Dongmu's metal soft-magnetic core capacity rises from 30,000 tons in 2024 to 50,000 tons in 2026. Meanwhile Dongmu is restructuring its traditional press-and-sinter capacity in Datong, Shanxi, gradually switching low-end gear capacity to EV motor end parts and motor controller housings.

Poco's expansion has two pieces. One is the Heyuan phase II alloy soft-magnetic core, ramping in 2025 to bring total capacity from 40,000 tons in 2024 to 60,000 tons in 2025. Two is the Huizhou AI-server-power-specific powder-core line, commissioned H1 2026, adding 20,000 tons specifically for NVIDIA GB200 series, hyperscale data center power, and 800V storage PCS. Heyuan phase II is about RMB 1.5 billion; Huizhou about RMB 800 million.

Yitong's expansion centers on Taixing. In 2025 Yitong added a PREP spherical-powder line in Taixing with 500-ton capacity, mainly for COMAC C919/C929 titanium structures, AECC main engine superalloy powder, and CASIC solid rocket motors. In Hefei, Yitong has a new gas-atomized alloy-powder project commissioning H2 2026, adding 3,000 tons mainly for consumer electronics foldable MIM and AI server soft-magnetic cores.

Yuean's expansion centers on Maanshan, Anhui. In 2025 its phase II added 5,000 tons of carbonyl iron and 2,000 tons of gas-atomized alloy capacity, total investment about RMB 700 million. New capacity mainly serves consumer-electronics foldable MIM, AI server soft-magnetic cores, and EMI shielding materials. Yuean also planned a phase III in Zibo, Shandong with 2027 commissioning, adding 5,000 tons of carbonyl iron.

Gaona's expansion centers on Beijing, Hebei, and Sichuan. Beijing's Gaona Powder phase II commissioned in 2025, adding 1,000 tons of spherical superalloy capacity mainly for the COMAC C929 program and AECC. Hebei Zhuozhou's vacuum-induction-melted master alloy expansion commissions H1 2026. Sichuan Deyang's local AECC-supporting powder project commissions H2 2026.

GRIPM's expansion is in Yantai and Hefei. Yantai's water-atomized iron phase III commissioned in 2025, adding 30,000 tons mainly for Dongmu and FRT press-and-sinter orders. Hefei's battery-anode precursor expansion commissions H2 2026.

These projects collectively signal Chinese local players actively positioning for three incremental tracks: EV soft-magnetic cores, AI server soft-magnetic cores, and COMAC + AECC superalloy powder. AI server power emerged in late 2025 but is now in the two-year expansion plans of Poco, Dongmu, and Yuean simultaneously — showing the track's appeal.

Another implicit signal: Chinese makers are pulling ahead of overseas giants in capacity. Höganäs' 2025 global expansion plan adds only 30,000 tons of water-atomized iron; Sumitomo Electric's metal soft-magnetic expansion only 10,000 tons; GKN even closed an old UK plant in 2025 (capacity adjustment). Overseas giants are conservative, Chinese makers aggressive. Over the next five years, Chinese makers' voice in global supply will further rise.

The second tier is also moving fast. Tonglian invested in a second MIM base in Zhangshu, Jiangxi, commissioned end-2025, adding 50 million units mainly for foldable hinges and smartwatches. Jingyan's Changzhou phase II commissions Q1 2026, adding 30 million units mainly for consumer electronics, EVs, and medical. FRT's Nantong phase III commissions Q2 2026, adding 20 million units mainly for medical MIM. The three combined replicate roughly another 2024 Tonglian.

New entrants are also rushing in. Late 2025 saw a wave of traditional ferrite makers (Dongmagnetic, TDK-CCAR, Dongfang Electronics) enter metal soft-magnetic cores with about 20,000 tons planned. A wave of cemented-carbide makers (Zhuzhou Diamond, Zhangyuan Tungsten) entered 3D-printing superalloy and titanium powders with about 1,000 tons planned. They won't dethrone leaders short-term but will intensify competition.

Capacity-expansion financing is also worth observing. This round draws from three sources: operating cash flow (head firms net at RMB 500 million to RMB 2 billion, self-funding strong); debt financing (Dongmu, Poco, Yuean issued convertible bonds or corporate bonds in 2024-2025 totaling about RMB 3 billion); equity financing (Gaona, GRIPM, Dongmu completed or started private placements in 2025 totaling about RMB 2 billion). Combined RMB 7 billion of capital roughly covers head firms' expansion.

Local government support also backstops capacity. Haimen and Taixing (Jiangsu), Heyuan and Huizhou (Guangdong), Maanshan (Anhui), Yantai (Shandong), Beijing, Deyang (Sichuan) all classify powder metallurgy as "priority industry," offering land, tax breaks, R&D and talent subsidies. Combined effective benefit is about RMB 1 billion in cost savings for head firms. Local support also reflects powder metallurgy's strategic role in regional manufacturing upgrade.

The flip side of expansion is exit of laggard capacity. Low-end press-and-sinter has overcapacity; a dozen small plants began exiting in 2025 under environmental, energy, and tech-upgrade pressure. This "selection of the fit" makes room for head firms' share gains. Environmental policy mainly affects reduction-annealing furnaces, sintering furnaces, and HIP furnaces' energy use and emissions; head firms with modern, low-energy equipment and complete emission controls are largely unaffected. Mid-small plants upgrade or exit — the normal process of structural upgrading.

The global capacity map is also evolving. Chinese leaders expand both domestically and abroad. Dongmu set up an MIM plant in Mexico in 2025, mainly serving Tesla Mexico. Poco announced a Vietnam metal soft-magnetic plant in Q1 2026, mainly serving Southeast Asia solar inverters and storage PCS. Yuean announced a European (Germany or Poland) carbonyl iron plant in Q2 2026, mainly serving European MIM and EMI shielding. The essence is "follow the customer" — as Chinese leaders' customers (consumer electronics, EVs, AI servers) globalize, powder metallurgy as key upstream must follow geographically. This marks China's powder metallurgy moving from "national champion" to "global player."

Expansion pace deserves deeper analysis. Powder metallurgy expansion cycles are typically two to three years (from approval to ramp), broken into 6-12 months equipment procurement, 6-12 months civil works, 3-6 months equipment commissioning, and 3-6 months capacity ramp. Cycles look short but often slip in practice: EIA/safety approvals, land acquisition, key equipment supply (large sintering furnaces and HIP furnaces), process tuning, customer qualification. Slipping past the optimal window can miss share-grab timing. Head firms must carefully assess this in expansion planning.

Expansion also requires "capacity utilization" consideration. Healthy powder metallurgy utilization is typically about 80 percent; under 70 percent hurts profitability; over 90 percent struggles to absorb order swings. 2025 utilization was about 82 percent — healthy. By sub-track, utilization varies widely: metal soft-magnetic cores above 90 percent (undersupplied), 3D-printing metal powder about 70 percent (capacity outrunning demand), press-and-sinter about 75 percent. Utilization differences mirror supply-demand and inform head firms' planning.

Beyond the head firms, the "waist" (revenue RMB 100 million to RMB 1 billion) expansion is also worth noting. Though individually modest, combined capacity is significant. 2025 to 2027, about 30 waist plants announced new expansion plans, combined capacity about 40 percent of head firms' six-company combined addition. Waist firms expand in niches (a specific soft-magnetic core spec, a specific MIM application, a specific spherical-powder alloy). Their expansion pressures head firms but also enriches industry supply.

Another expansion observation is "greening." New capacity adopts higher-efficiency sintering furnaces (15-20 percent better than legacy), cleaner atmospheres (hydrogen replacing some natural gas), and improved emission controls (VOC to EU standards). Green expansion meets dual-carbon targets and high-end customer (Tesla, Apple, European OEMs) supply-chain carbon footprint requirements. By 2028, China's powder metallurgy industry's energy intensity should drop 15 percent vs 2025 and carbon intensity 20 percent — concrete green-transition embodiment.

Final expansion observation is "smartification." New capacity broadly adopts industrial internet, MES, AGV, machine-vision inspection, data-analytics platforms. These tools enhance traceability, controllability, and optimization, sharply improving process stability. Head firms allocate 10-15 percent of new-capacity investment to smartification. Smartification level has become a key differentiator between head and second-tier firms.

Capacity-and-demand fit is the most critical observation indicator for the next three to five years. Excess capacity vs lagging demand brings transitional overcapacity, price wars, and margin pressure; lagging capacity vs surprise demand brings order-delay and lost share to rivals. Head firms have seen both: 2018-2019 brief overcapacity (traditional auto miss), 2023-2024 brief undercapacity (EV demand surprise). The lesson: head firms manage capacity via "base + flex" — stable base, flex via outsourcing for rapid adjustment, allowing response to both upside surprises and demand misses. This capacity-management capability is a key head-firm soft skill.

Spreading the expansion picture, China's powder metallurgy supply is secured for the next three to five years. Head firms, second tier, and new entrants expand simultaneously; combined new capacity easily covers projected demand. But supply security doesn't equal stable competition: faster expansion and more entrants intensify competition, raising the bar for head-firm leadership maintenance. That is expansion's other side.

Chapter 10: Price Cycle — 2024-2026 Soft Magnetic Cores, Press-and-Sinter, MIM Unit Prices

Powder metallurgy price curves differ from those of bulk metals like copper, aluminum, and iron ore. The main drivers are three: raw-powder cost, downstream OEM demand intensity, and capacity utilization. Each sub-track has its own cycle.

Metal soft-magnetic core prices traced "drop, stabilize, rise" through 2024 to 2026. In 2024, slowing EV growth, end of PV pull-forward, and weaker storage installations softened demand, while new capacity (Dongmu Haimen, Poco Heyuan phase II) released — short price war ensued. One ton of Sendust core fell from RMB 80,000 average in 2023 to RMB 65,000 by H2 2024; non-crystalline nanocrystalline from RMB 150,000 to RMB 110,000. In H2 2025 it reversed. EV 800V rolled out broadly on BYD, Tesla, AITO, Li Auto, doubling powder-core demand for 800V OBC vs 400V. AI server power began large-scale procurement; NVIDIA GB200 NVL72 racks' BBU started erupting in H2 2025. The two new incremental scenes pulled Sendust prices from the 2025 low of RMB 65,000 back to RMB 85,000 by H1 2026. Poco's H1 2026 guidance specifically calls out AI-server-power orders' average-price contribution — 20 percent above traditional auto.

Press-and-sinter structural parts price cycles are relatively stable. Mature business, low capacity additions, low demand growth — pricing floats with water-atomized iron costs. 2024 water-atomized iron averaged RMB 12,000 per ton; 2025 rose to RMB 13,500 on iron ore and electricity costs, lifting part prices about 7 percent. H1 2026 water-atomized iron stable at RMB 13,000-14,000, with part prices steady. Dongmu and FRT achieved stable volume and price in H1 2026 press-and-sinter.

MIM part prices in 2024-2026 traced "titanium up, stainless flat, aluminum down." Titanium MIM rose from RMB 20 per hinge piece average in 2023 to RMB 28 in 2025 as foldable hinges shifted to titanium. H1 2026, with Jingyan, Tonglian, FRT capacity coming on and Huawei, Xiaomi, OPPO, Honor supply-chain diversifying, titanium MIM eased back to about RMB 25. Stainless MIM stable RMB 12-15. Aluminum MIM bumped up briefly in 2025 on AR/VR frame lightweighting, then fell to about RMB 8 in 2026 as new entrants arrived.

3D-printing metal powder 2024-2026 traced "superalloy stable, titanium down, aluminum down." Superalloy stable at RMB 600,000-1,000,000 per ton on aero-engine pre-research and COMAC C929 demand. Titanium dropped from RMB 1.2 million per ton in 2023 to RMB 900,000 in 2025 to RMB 800,000 H1 2026 as Yitong, Sailong, GRIPM expanded. Aluminum dropped from RMB 600,000 per ton in 2023 to RMB 400,000 H1 2026 as new entrants arrived in droves. 3D-printing powder margins held 40-50 percent because demand grows faster than capacity releases.

The price-cycle core rule: fast-growing tracks (soft-magnetic cores, 3D-printing powder) have high price elasticity, fast capacity expansion, prone to intermittent oversupply; steady tracks (press-and-sinter, stainless MIM) have stable prices floating with raw material costs; surge new scenes (AI server power soft-magnetic cores) often see premium prices because new entrants haven't yet sized up. Understanding this rhythm matters for investment and operations.

Customer-structure cycles deserve attention beyond price. Powder metallurgy's customer mix changed fundamentally from 2020 to 2026. Pre-2020 head firms' customers were mainly traditional ICE auto OEMs (60+ percent) and appliances (15 percent), with EV and consumer electronics under 20 percent. By 2025 EV customers exceeded 30 percent, consumer electronics rose to 15 percent, AI server power emerged from zero to 8 percent, PV-storage from 3 to 10 percent. Traditional auto fell from 60 percent to 30 percent. Not a simple "old out, new in" but head firms' deliberate strategic shift.

The new customer mix brought margin uplift. EV, AI server power, and high-end consumer electronics customers pay 30-50 percent more per unit and at 10-20 percentage points higher margin than traditional ICE and appliance customers. This is why head firms' margins rose from 16 percent in 2020 to 24 percent in 2025 while industry revenue grew only 12 percent CAGR — margin expansion outpaced revenue, reflecting customer-mix upgrade.

The new customer mix also brought new risks. First, rising customer concentration: Dongmu's soft-magnetic top-five customers (BYD, CATL, Huawei Digital Power, Sungrow, Tesla via ODM) account for 65 percent of revenue. Poco's top-five 70 percent. Tonglian's top-three (Huawei, Apple EMS, Samsung EMS) 75 percent. Materially higher than ICE-era (top-five typically under 40 percent). Concentration weakens pricing power against a single big customer and amplifies single-customer order swings.

Second, longer qualification cycles. EV and AI server power qualifications run 12-24 months, far longer than traditional auto's 6-12 months. Once in, switching cost is prohibitive (involves chain rework of circuits, safety certs, reliability testing), so head firms supply stably 5-7 years. Conversely, missing the window means waiting for the next platform — a "winner-takes-all" dynamic intensifying competition among head firms.

Third, faster technology iteration. Traditional auto iterates 7-10 years per platform; appliances 5-7 years. EVs accelerated to 3-4 years; AI server power even 1-2 years. Head firms must complete new powder, new core design, and new process R&D in shorter time. This pace tests engineering teams, R&D intensity, and co-design capability with customers. Those who keep up benefit continuously; those who can't get sidelined.

The compounding of price and customer-structure cycles is the core dynamism of powder metallurgy through 2025-2030. Five years of rapid scale expansion and five years of intense internal reshuffling at the same time. Whether head firms keep first-mover advantage in new scenes, whether entrants find differentiated entry, whether overseas giants hold European-American share — all to be answered in these five years.

Order-confirmation mechanisms are another key variable. In powder metallurgy, head firms and major customers typically sign annual framework agreements stipulating projected annual volumes, target prices, payment terms, and quality — without constituting legally binding hard orders. Actual orders are confirmed monthly or quarterly. This "framework + rolling" mechanism creates fluctuation room between revenue and budgets: if demand surprises up, head firms can fulfill quickly; if not, framework orders may shrink. Understanding this helps explain why head firms' quarterly results sometimes swing significantly.

Relationships between price cycles and new-scene penetration also deserve elaboration. Typically a new scene moves from "sample qualification" to "small-volume production" to "scale ramp" over 12-36 months. During qualification, powder-core unit prices may be 30-50 percent above mature scenes but volumes are tiny. During small volume, prices roughly match mature, but volumes ramp, materially contributing to revenue. During scale ramp, prices fall 10-20 percent on competition, but volume growth pulls revenue strongly up. AI server power is now transitioning small-volume to scale ramp; 2026-2027 is the watch period. Successful ramp delivers material revenue and margin upside.

Last price-cycle dimension is "cost transmission lag." In powder metallurgy cost structure, raw powder, energy, and labor make up the bulk. Raw powder price moves transmit to core prices in 1-3 months; energy 3-6 months; labor remains relatively stable. Sharp raw moves (e.g. 2022 nickel) hit margins short-term, recovering as prices flow downstream. This lag is another driver of quarterly margin volatility.

Chapter 11: Policy and Downstream Drivers — Smart Manufacturing, Dual Carbon, COMAC

Powder metallurgy as foundational process rarely appears in direct policy documents, but virtually every major industrial-policy landing reverse-pulls powder metallurgy demand. Spreading the 2022-2026 policy cycle reveals three main lines.

The first is smart manufacturing and precision manufacturing pulling MIM and press-and-sinter parts demand. The MIIT's "14th Five-Year Smart Manufacturing Plan" (2021) and "Robot + Application Action Plan" (2023) listed precision gears, precision actuators, collaborative-robot joints, and servo-motor end covers in the critical-parts breakthrough list. These parts heavily use press-and-sinter gears and MIM precision parts. Meanwhile, consumer electronics — phones, smartwatches, smart glasses, TWS earbuds, AR/VR headsets — sustainably pull MIM penetration. In 2026, MIIT also launched the "Smart Terminal Parts High-End Supply Plan," listing foldable hinge MIM, smartwatch-case MIM, and AI-glasses-frame MIM as priority categories.

The second is the dual-carbon target pulling metal soft-magnetic cores. The 2020 "3060" goals (peak carbon 2030, neutrality 2060) propagated via EVs, PV, storage, and wind to metal soft-magnetic cores. EVs hold 45 percent of metal soft-magnetic core downstream; PV 20 percent; storage 15 percent; server power 10 percent; other industrial power 10 percent. Each main line has clear policy drivers: EV dual-credit, NEV-to-countryside, 800V fast-charge subsidies; PV county-level promotion, "Thousand Villages Wind Action"; storage mandates, shared storage; AI data center "East Data West Compute" and liquid-cooling builds. The five forces drove metal soft-magnetic from RMB 3 billion in 2020 to RMB 12 billion in 2025. H1 2026 NDRC + NEA joint "Measures on High-Efficient Energy Use System Building for Data Centers" capped data-center PUE at 1.25, further pulling 800V DC, high-power-density power, and soft-magnetic core demand.

The third is COMAC C919/C929, Changjiang engines, and commercial spaceflight pulling 3D-printing metal powder and superalloy powder demand. COMAC C919 delivered 38 in 2024, 60 in 2025, planned 100+ in 2026. C919 titanium content is 9.5 percent of fuselage weight, with some critical parts under 3D-printing evaluation. C929 widebody expected to first fly 2030, certify 2032, and deliver 2033; Changjiang engines (CJ-1000A, CJ-2000) pre-research plans over 15 percent 3D-printed content. In commercial spaceflight, SpaceX Falcon Heavy and Starlink reduced launch costs globally; China's Galactic Energy, iSpace, LandSpace, Deep Blue Aerospace are also iterating; bulk commercial satellite launches rapidly raise 3D-printing metal powder (titanium, aluminum, superalloy) demand. CASIC's 2025 "Hongyan constellation" Phase II plans 4,000 satellites; each satellite's brackets, fuel tanks, attitude-control engine nozzles use 3D printing extensively — combined metal-powder demand about 1,000 tons.

The shared trait: policies don't directly subsidize powder metallurgy but pull EVs, PV-storage, COMAC, consumer electronics demand, which reverse-forces powder metallurgy capacity, cost, and tech. This "downstream-driven" policy path means powder metallurgy sentiment tracks the four downstream's tightly. Read the downstream to read the upstream.

Beyond the three main lines, several secondary but equally important policies matter. First is the "New Materials Chain Strengthening Strategy." MIIT's "First Batch Application Demonstration Catalog" and "New Materials Chain Stability Improvement Program" since 2022 listed multiple critical categories in powder metallurgy as "bottleneck materials," including aerospace spherical superalloy powder, medical spherical titanium powder, fourth-generation metal soft-magnetic cores. Once listed, after domestic substitution succeeds, first-batch application can be supported by a government insurance mechanism, lowering downstream OEM risk concerns. This policy directly accelerated domestic substitution of high-end powder metallurgy products.

Second is the "specialized and innovative enterprise cultivation" policy. A cohort of mid-sized niche leaders (RMB 300 million to 2 billion revenue) are recognized as "Specialized and Innovative Little Giants" — eligible for R&D subsidies, tax breaks, IPO green channels. Yitong and Yuean both benefited, reducing financing pressure. This policy also encourages head firms to professionalize their niches rather than compete in "big and broad."

Third is "Made in China 2025" and "Manufacturing Champions" selections. MIIT and CFLP have selected "Manufacturing Single-Item Champions" since 2016. Powder metallurgy entries include Dongmu (auto powder metallurgy parts), Poco (metal soft-magnetic cores), Yuean (carbonyl iron), Gaona (superalloy). Champion designation grants national brand endorsement, valuable for international expansion and major-customer qualification.

Fourth is export-control policies. H2 2025 and H1 2026, MOFCOM and Customs issued multiple notices on rare earths, special alloys, high-end powder, mainly affecting spherical superalloy, spherical titanium, rare-earth permanent magnet exports. Controls protect domestic strategic materials and push exporters toward the domestic market. For 3D-printing metal powder this is neutral-to-positive: lost some overseas orders but won more domestic aerospace, medical, commercial space share. Net positive.

Fifth is dual-carbon and green manufacturing. Powder metallurgy itself has relatively low energy/carbon intensity, but as critical upstream of EVs, PV, storage, and wind — pillars of carbon neutrality — its strategic importance is elevated. Green finance (green bonds, green credit, ESG investment) tilts toward powder metallurgy, reducing head-firm financing costs. Dongmu and Poco both issued green bonds in 2025-2026 at 20-50 bps below comparable plain bonds.

Sixth is local government investment promotion. Jiangsu, Guangdong, Anhui, Shandong, Sichuan county and city governments broadly classify powder metallurgy as "priority introduction," offering land, taxes, talent, R&D subsidies. These local government supports effectively provide implicit subsidies for capacity expansion.

Taken together, powder metallurgy enjoys a relatively loose, friendly policy environment from 2025 to 2030. This supports both industry sentiment and capacity expansion. But friendly today doesn't mean friendly forever; head firms need to use the window to complete tech upgrade and market positioning, so they can keep advantage when policy dividends fade.

Downstream-driven transmission mechanisms deserve specifics. Take EV OBC: the policy → powder demand path is roughly — dual-credit and EV subsidies stimulate OEMs to raise EV production, OEM OBC modules order Tier 1, Tier 1 orders second-tier soft-magnetic core suppliers, second-tier suppliers order upstream powder. Full chain transmits in 6-12 months; from policy issuance to actual demand spike takes about half a year to a year. This lag means powder metallurgy sentiment lags policy by half a year to a year. Understanding this helps pre-position capacity at policy announcement rather than reacting after demand spikes.

Sub-tracks have different paths and lags too. Metal soft-magnetic cores have shortest lag (about 6-9 months) — fast downstream iteration, quick stock turn. Press-and-sinter has longer lag (about 9-15 months) — long auto supply-chain management, large inventories. MIM is shortest (3-6 months) — extreme consumer-electronics iteration speed and supply-chain responsiveness. 3D-printing powder is longest (1-3 years) — aerospace/medical product qualification is extremely long, slow ordering. Knowing each sub-track's lag is key to understanding policy-to-sentiment impact.

For 2026-2030 specifically, several policy directions warrant early attention. First, deepening "AI+" strategy. The State Council's 2024 "AI+ Action Plan" and 2026 H1's "AI Compute Infrastructure 3-Year Plan" treat AI compute infrastructure as national strategy, sustainably pulling AI servers, liquid-cooled data centers, high-power-density power supplies — and reverse-pulling metal soft-magnetic core demand. Second, "dual circulation" domestic-demand expansion. State Council's expansion of domestic demand, consumption stimulus, and new urbanization sustainably grow appliances, autos, consumer electronics. Third, deepening "manufacturing power" strategy. MIIT's "Key Industry Chain Supply-Chain Stability Improvement" and "Specialized-and-Innovative Enterprise Cultivation" sustainably benefit basic process industries like powder metallurgy. Fourth, full-scale roll-out of "green low-carbon transition." NDRC, MEE, PBoC joint green-finance, green-manufacturing, green-supply-chain policies will sustainably support powder metallurgy's greening with capital, tech, and market access. Fifth, full-scale "Digital China" deployment. Cyberspace Administration and SASAC digital infrastructure, smart manufacturing, industrial internet policies will sustainably push powder metallurgy's digital and smart upgrade.

These five directions each emphasize their own angles but all are concrete landings of the national "modern industrial system" strategy. As key basic-process material in the modern industrial system, powder metallurgy finds its strategic place in each direction. This is the firmest political ground over the next five to ten years.

Chapter 12: Research Institute Judgments — Six Calls on a 3- to 5-Year View

After spreading out the first eleven chapters' facts, we can offer several 3- to 5-year judgments from Tianxia Gongchang Research Institute's perspective.

First judgment: metal soft-magnetic core penetration will keep rising, but the competitive landscape will shift from "top three split the market" to "top three + new entrants." Poco, Dongmu's subsidiary, and Yitong will keep top positions, but the 2025-2027 wave of AI server, 800V OBC, and storage PCS will attract many new entrants (including traditional ferrite-magnetic, inductor, and transformer makers). By 2028 total industry capacity of metal soft-magnetic cores should double 2025 to over 200,000 tons; top-three share will fall from 75 percent in 2025 to about 60 percent in 2028. This is the inevitable pull between capacity ramp and demand growth.

Second judgment: MIM penetration in consumer electronics is near saturation; new growth comes mainly from medical, robotics, and EVs. Foldable hinges, smartwatch cases, AR/VR frames are near saturated by 2026; 2027-2030 growth slows from 30 percent to about 10 percent. Real growth: one, robotics — humanoid and collaborative robots need large precision MIM joints; leading firms (UBTECH, Unitree, Tesla Optimus, Boston Dynamics) begin small-volume production in 2026, MIM demand to surge. Two, medical — orthodontics, orthopedic implants, endoscope effectors keep rising. Three, EVs — 800V fast-charge modules, thermal management actuators, battery BMS connectors expand.

Third judgment: 3D-printing metal powder domestic substitution mostly completes between 2026 and 2030. COMAC C919/C929's domestic superalloy and titanium are on fast-track; by 2030 domestic share should rise from 40 percent in 2025 to over 70 percent. Equipment-powder-process lock-in will make Gaona, Yitong, BLT, Farsoon a "Chinese 3D-printing metal ecosystem." This ecosystem's COMAC + AECC share will steadily rise.

Fourth judgment: overseas giants' China share will keep being squeezed, but global share remains secure. GKN, Höganäs, Sumitomo Electric China share will fall from 20 percent in 2025 to under 10 percent in 2030, because Chinese locals dominate EV, consumer electronics, and AI server supply chains. But in European-American traditional auto OEM chains, overseas giants' moat is intact; Chinese "true globalization" depends on entering European-American Tier 1 chains. Dongmu's 2025 IMECO acquisition entered Japanese auto OEM supply — a sample worth tracking.

Fifth judgment: M&A consolidation will accelerate. Past the wild-growth phase, leading firms have stable cash flow, reasonable valuations, and good consolidation ability; 2026-2030 will see several flagship deals: top firm acquires niche leader (e.g., Dongmu acquires an AR-glasses MIM specialist; Poco acquires an amorphous nanocrystalline tape maker); powder feedstock firms acquire downstream core or MIM firms (e.g., Yuean acquires an MIM firm); SOE powder coordination consolidates (Gaona, AVIC Mete, GRIPM consolidate).

Sixth judgment: powder metallurgy's "data infrastructure" becomes the industry's new productive force. Based on: powder metallurgy's downstream is extremely fragmented (hundreds of mid-sized plants), the traditional "by relationship, by trade shows, by phonebook" matching is low-efficiency, costly, asymmetric. Such plant-data platforms indexed by process/material/capacity/qualification will become new infrastructure for industry supply-demand matching, compressing finding factories, sourcing, capacity-checking, qualification, and chain research costs to a tenth of current. The core: B2B data infrastructure's value isn't replacing human judgment but liberating it from "finding info" to "making decisions."

The premise of these six judgments: China's manufacturing fundamentals don't systemically reverse over 3-5 years, and EVs, AI servers, COMAC, consumer electronics remain healthy. If the premise holds, powder metallurgy is in a 3-5-year expansion worth tracking.

Stringing the six judgments together produces a more macro observation: China's powder metallurgy is at a critical "double-upgrade" node. First upgrade: product mix — from traditional press-and-sinter dominance to compound drivers of soft-magnetic cores, MIM, and 3D-printing powder. Second upgrade: market mix — from domestic-led to "domestic + global" dual engine. Both upgrades simultaneously are both opportunity and challenge. Opportunity: industry valuation ceiling rises; head firms' market caps could double or triple in five years. Challenge: simultaneous strategic, technical, and organizational upgrades for head firms; any laggard could fall behind.

Over the next five years, powder metallurgy may see several anticipatable "milestones." First, a Chinese powder metallurgy head firm enters global top three (by combined scale and technology). Currently GKN, Höganäs, Sumitomo Electric are global top three; Chinese head firms sit between fourth and eighth. If Dongmu becomes global top across press-and-sinter, soft-magnetic, and MIM combined, or Poco becomes single-category global top in soft-magnetic, the rankings could rewrite. Second, Chinese 3D-printing metal powder first enters mainstream European-American commercial-aircraft supply chains. Airbus and Boeing critical 3D-printed parts are mostly from European-American powder suppliers; Chinese makers breaking in by 2028-2030 mark China's true entry to the global first tier of high-end powder metallurgy. Third, China births its first RMB 100 billion market-cap powder metallurgy firm. The largest now is RMB 20-30 billion; breaking RMB 100 billion in five years marks capital market's complete revaluation of the industry.

Another important direction is "material-process-equipment" trinity coordinated innovation. Powder metallurgy progress historically followed "material leads, process follows, equipment last" or "equipment leads, process follows, material last." From 2025, head firms drive trinity coordinated innovation: powder makers and equipment makers jointly define next-gen SLM parameters; powder makers and downstream power OEMs jointly define next-gen metal soft-magnetic core alloys; powder makers and phone brands jointly define next-gen titanium MIM hinge geometry. Trinity coordinated innovation markedly raises Chinese head firms' chain voice and is key to their persistent leadership in new scenes.

Last worth mentioning is the rise of "industry research institute" type organs. Powder metallurgy is technically complex, sub-track-divided, info-asymmetric; specialized industry research institutes are needed to keep delivering independent analysis. Historically this work was done by broker research, associations, consultancies, but coverage was limited. Since 2023 more specialized industry research institutes, technical service organizations, and data analytics platforms have entered the field, providing deeper, more granular, more forward-looking industry analysis. The formation of an "industry research institute ecosystem" will sharply raise powder metallurgy's information transparency, lowering supply and demand sides' transaction and decision costs.

To capital-market view, six judgments can be extended to investment logic. First, niche leaders beat diversified platforms. In an environment of widening sub-track divergence, niche leaders' (Poco in soft-magnetic, Yuean in carbonyl iron) growth certainty beats diversified platforms (Dongmu). Platforms must invest persistently in multiple tracks, dispersing capital and R&D, hard to achieve extreme. Niche leaders can focus, hard to match in their niche. That's why Poco's 7-year compound market-cap growth doubled Dongmu's.

Second, "downstream-tied" firms beat "powder upstream" firms. Powder metallurgy has two camps: powder material providers (GRIPM, parts of Yuean, parts of Yitong) and powder-part downstream service providers (Dongmu, Poco, Tonglian, Jingyan, FRT, BLT, Farsoon). The latter directly engage downstream OEMs with stronger pricing power, higher margins, stickier customers. The former are upstream but pricing-power constrained by downstream part makers, with naturally lower margins. Unless they're "bottleneck" high-end powder (spherical titanium, fourth-gen metal soft-magnetic), where pricing matches downstream.

Third, "R&D-driven" firms beat "scale-expansion" firms. Powder metallurgy's moat is process R&D, not capacity. R&D-to-revenue ratio, engineer count, patents, customer co-design ability — these decide whether a firm keeps leading in new scenes. Poco's R&D-to-revenue rose from 3 percent in 2018 to 7 percent in 2025 — the key to its repeated next-gen passivation and AI-server-power leadership.

Fourth, "globalized" firms beat "single-domestic" firms. Powder metallurgy is highly globalized; overseas share matters greatly for long-term growth. Chinese head firms now sit at 15-30 percent overseas; if this rises to 40+ percent over the next five years, growth space opens. Dongmu acquired IMECO and set Mexico plant; Poco set Vietnam; Yuean Europe. This globalization is a key watch point.

Fifth, "deep tech reserves" beat "short-term performance elasticity." Powder metallurgy tech iterates incrementally, but every new process, material, scene breakthrough takes 3-5 years of reserve. Whether a firm started watching amorphous nanocrystalline 10 years ago, fourth-gen passivation 5 years ago, 800V OBC powder 3 years ago — decides whether it leads when 2025-2030's new scenes erupt. Firms with short-term elasticity but shallow reserves fall behind on the next switch.

Capital markets are reshaping their view of powder metallurgy. Pre-2020 the labels were "traditional manufacturing," "low margin," "mature industry," "low valuation ceiling." The 2022-2026 EV/AI/COMAC wave is breaking those labels. Head firms' margins rose from 20 to 35 percent; revenue growth from single digit to 30 percent; valuations from low teens to 20s and 30s PE. This valuation reset just started in the past three years and has further room over the next three to five. This is the industry's biggest "rerating opportunity."

Looking longer, several "decade-class" opportunities await. First, powder production of solid-state battery cathode and anode materials. Next-gen solid-state batteries may require entirely new powder metallurgy to produce, spawning a new sub-track. Second, fusion-reactor first-wall material development. Commercial fusion is far off, but related materials are in pre-research. Fusion first-wall needs extreme heat, neutron-radiation, plasma-sputtering resistance — main candidates are tungsten-based, beryllium-based, vanadium-based alloys, all powder metallurgy. Third, space mining and in-situ manufacturing. SpaceX, Blue Origin's crewed lunar and Mars plans if successful spawn in-situ-manufacturing demand; powder metallurgy is among the most suitable space-manufacturing routes. Fourth, quantum computer and superconductor parts demand. Quantum computers' cryogenic operating environments need special low-temp alloy precision parts, some via powder metallurgy. Fifth, brain-computer interfaces and neural repair devices. BCI core implants need biocompatibility, corrosion resistance, extreme shape complexity; powder metallurgy fits.

These "decade-class" opportunities won't materially affect short-term sentiment but represent the industry's long-term potential. Chinese head firms must invest in these frontiers alongside short-term market-share expansion. This "short + long" dual layout is the head firms' fundamental guarantee of sustained leadership.

Finally, this chapter's judgments should return to a plain industry observation: powder metallurgy has evolved along the "forming from powder" main line for decades. From the earliest press-and-sinter parts, to mid-stage metal soft-magnetic cores and MIM, to today's 3D-printing metal powder — each new process opens new application scenes, each new application attracts new entrants, each wave of new entrants pushes process evolution further. This "process → application → maker → process" loop is the industry's most vital trait. Chinese locals seized key windows in this loop over the past decade, catching up and surpassing overseas giants in many niches. The loop will continue over the next five to ten years; we have reason to remain long-term optimistic about this "forming from powder" business.

Chapter 13: Risks and Variables

Every judgment requires laying out corresponding risks. Powder metallurgy's 3 to 5 years has several worth being alert to.

First, consumer-electronics demand softens. MIM's consumer-electronics share exceeds 50 percent. If foldable phone growth slows from projected 35 percent in 2025 to under 10 percent in 2027, smartwatch saturates, AR/VR "next iPhone moment" doesn't arrive — overall MIM growth falls from 25 percent in 2025 to under 10 percent. Massive valuation pressure on Jingyan, Tonglian, FRT.

Second, overseas giants' price competition. Höganäs, GKN, Sumitomo Electric may cut prices to defend European-American share. Water-atomized iron down 15 percent, press-and-sinter down 10 percent, soft-magnetic cores down 15 percent — Chinese makers' export margins compressed materially. Dongmu and Poco overseas share expansion may be impeded.

Third, ferrite cost-performance rebound. Soft-magnetic powder cores' biggest substitute is ferrite. In low-power-density low-cost scenes ferrite still has price advantage. If iron oxide, manganese, zinc oxide raw material prices drop sharply, or ferrite breaks through in high-power-density (ultra-high Bs or low-loss ferrite), metal soft-magnetic core penetration pace breaks.

Fourth, EV growth slowdown. China EV penetration exceeded 50 percent in 2025; 2026-2030 growth naturally slows. If 2030 China EV sales grow from 15 million in 2025 to only 20 million (6 percent CAGR) instead of forecasted 25 million (10 percent CAGR), OBC soft-magnetic demand growth falls materially.

Fifth, AI server power standard shifts. NVIDIA GB200 uses 800V DC with metal soft-magnetic; whether next-gen (GB300, Rubin) keeps metal soft-magnetic depends on NVIDIA and hyperscalers' (Google, Amazon, Microsoft, Meta, ByteDance, Alibaba, Tencent) tech choices. If next-gen moves to GaN high-frequency switching (kHz to MHz), powder-core demand drops materially.

Sixth, COMAC C929 and Changjiang delays. 3D-printing metal powder's biggest superalloy and titanium incremental scene is COMAC C929 and Changjiang. If C929's first flight slips from 2030 to 2032, Changjiang installation from 2030 to 2033, 3D-printing metal powder demand lags two to three years.

Seventh, industry over-expansion. As noted, head firms expand aggressively; if 2026-2028 downstream lags expectations, intermittent overcapacity, prices fall further, head-firm margins compress.

Spreading these seven, the core isn't powder metallurgy's process risk but downstream demand variables. As "basic process," powder metallurgy's own tech progress is stable; what really determines sentiment is the four downstream players. Read the downstream to predict risk.

Beyond the seven main risks, several secondary risks deserve attention. First, geopolitical risk. Key equipment (SLM, EBM, HIP, vacuum induction) and key raw materials (some rare metals: rhenium, tantalum, hafnium) still depend on overseas import. If geopolitical tensions cut or restrict imports, Chinese head firms' expansion paces disrupt. US-China tech friction since 2023 already affected some high-end 3D-printing metal equipment (US GE Additive, EOS imposed restrictions on some Chinese clients); 2026-2030 pressure may intensify. Chinese head firms respond by accelerating equipment domestic substitution: BLT, Farsoon SLM equipment achieved domestic substitution of mainstream German EOS in 2025, but EBM domestic substitution lags.

Second, raw price violent volatility. Powder metallurgy raw powders (iron, aluminum, copper) are relatively stable, but key alloy elements (nickel, molybdenum, chromium, tungsten, titanium, rhenium, cobalt) move significantly. Nickel and cobalt particularly: 2022 Indonesia nickel-ore export controls drove 30 percent monthly spikes; cobalt fluctuates with Congo politics. These movements transmit to downstream soft-magnetic, MIM, 3D-printing powder costs, hitting head-firm margins.

Third, IP litigation risk. Chinese head firms' rapid domestic substitution may touch overseas giants' core patents. Höganäs, GKN, Hitachi Metals own large foundational patents in powder formulation, sintering, passivation. If Chinese head firms' overseas share expands enough, infringement suits' probability rises. Litigation typically leads to legal costs, settlements, license fees.

Fourth, new-process route disruption risk. Powder metallurgy process routes are relatively mature, but disruptive tech may emerge. Cold spray, nano-imprint, laser-melting deposition (LMD) — maturing routes may substitute press-and-sinter, MIM, HIP, SLM in some niches. If new routes break through in specific segments, some existing capacity faces obsolescence. Chinese head firms respond by laying out new routes in parallel, but layout intensity requires foresight.

Fifth, talent loss risk. Powder metallurgy's core competence is process R&D, whose core asset is experienced engineers. Chinese head firms' core process engineer teams typically number 50-100, mostly with 10+ years of experience. Cluster loss (poached, founding own firms, switching industries) seriously hits R&D ability. Since 2025 with more new entrants, head firms face greater poaching. Retaining core process talent is another key management challenge.

Sixth, ESG and carbon constraints risk. Powder metallurgy's own carbon intensity is relatively low, but as critical upstream for carbon-neutrality pillars (EVs, storage, wind), its carbon data is included in downstream OEMs' chain carbon statistics. Tesla, European OEMs, Apple have all proposed "net-zero" timelines for suppliers (chain net-zero by 2030 or 2035). Chinese head firms must complete energy structure decarbonization (green power, clean energy replacing coal) over the next 5-10 years, or face losing some high-end orders. This is how ESG transmits to Chinese manufacturing.

Seventh, M&A integration risk. Powder metallurgy M&A will accelerate 2026-2030. But integration carries culture conflict, management mis-integration, weaker-than-expected synergy risks. Chinese head firms' past major deals (Dongmu IMECO, Poco overseas soft-magnetic) had mixed integration outcomes; future big deals' success directly affects head-firm market-cap performance.

Eighth, value-chain bargaining imbalance. Powder metallurgy downstream is mainly large customers (auto OEMs, consumer-electronics brands, power Tier 1, aerospace primes); upstream is powder material providers. In chain bargaining, mid-stream powder-part makers' position is most awkward: weak vs large downstream customers, and not necessarily strong vs upstream. If downstream slows and upstream raw prices rise, powder-part makers may be squeezed both ways, with margins falling materially. Typical cyclical-industry risk. Head firms respond by "extending up and downstream": up by self-producing powder (Poco done, Dongmu progressing), down by making modules (single magnetic core → inductor module; single MIM part → assembly). This extension defines head-firm position in chain bargaining.

Ninth, tech leak and supply-chain security risk. Powder metallurgy core tech (powder formulation, sintering process, passivation) mostly exists as patents, trade secrets, process skills, vulnerable to leak. Especially when core engineers jump to rivals or start their own firms with key tech. Head firms protect via non-competes, NDAs, equity incentives, but full prevention is impossible. Supply-chain security risk: if a key raw (a rare metal, a special chemical) supplier has issues, may affect production-line operation. Head firms mitigate by diversifying suppliers and keeping appropriate raw inventories.

Tenth, industry-wide tech misjudgment risk. Powder metallurgy's expansion relies on future demand judgments. If industry-wide misjudgment occurs (overestimating 800V OBC penetration speed, underestimating AI server power cap, overestimating 3D-printing penetration of traditional industries), industry-wide over- or undercapacity results. Once it occurs even head firms can't survive alone. Mitigation is via dynamic demand tracking and capacity adjustment, releasing in stages by actual downstream progress rather than upfront.

After laying out the ten risks, powder metallurgy's risk picture is more complex than surface. But it should also be noted: most risks are "manageable" not "irreversible." Head firms with proper management can mostly mitigate. The truly irreversible: geopolitical shocks cutting key equipment/raw materials, and core tech blockaded by overseas giants via legal means. Low probability but huge impact — strategic bottom-line thinking and contingency plans needed.

Beyond risk identification, response strategies matter. Common powder metallurgy head-firm responses fall into: First, "diversify downstream applications" — avoid over-reliance on a single downstream. Dongmu typifies: laying out press-and-sinter, metal soft-magnetic, MIM simultaneously, plus auto, electronics, appliances, EV, minimizing single-downstream impact. Second, "tiered capacity planning" — not all-at-once release but staged based on market. Poco Heyuan phase II released in 2-3 batches based on orders. Third, "diversified customer mix" — actively cap single-customer share, typically requiring top customer under 30 percent and top five under 70 percent. Active control may lose short-term orders but avoid long-term volatility. Fourth, "cash reserves" — head firms typically keep 1-2x operating cash flow in reserve. Fifth, "diversified tech reserves" — beyond main R&D, modest forward-looking R&D in adjacent tech (amorphous nanocrystalline, SMC, new powder formulations) to hedge main-business route change. Sixth, "globalization to disperse risk" — overseas plants, sales, raw sourcing diversify geopolitical and single-market risks.

Specifically, response strategies vary by company. Dongmu's emphasis: "business diversification + overseas expansion"; Poco's: "sustained tech investment + active customer mix management"; Yuean's: "niche leader + cash flow stability"; Yitong's: "deep specialization + IPO funding R&D"; Gaona's: "SOE platform + customer deep binding"; GRIPM's: "scale advantage + cost control." Each company combines based on its resources, business mix, customer base.

Downstream risks also deserve reverse thinking. Powder metallurgy downstream is the key risk source but also the key opportunity source. If downstream sees new, surprise demand (humanoid robot eruption, AR-glasses surprising adoption, controllable fusion commercialization, space mining), powder metallurgy demand will also surprise upward. So head firms in risk management need to consider both downstream miss downside and reserve capacity flex for upside surprise. This "two-way risk management" is head-firm's high-water mark.

Chapter 14: Data Sources and Further Reading

This report's factual basis comes from the following public channels:

Company level: Dongmu Co. 2025 Annual Report and Q1 2026 Report; Poco Holding 2025 Annual Report and Q1 2026 Report; Yuean Advanced Materials 2025 Annual Report and H1 2026 Guidance; Yitong New Materials 2025 NEEQ disclosure and annual report; Gaona Aero Material 2025 Annual Report and Q1 2026 Report; GRIPM Advanced Materials 2025 Annual Report; Jingyan Technology 2025 Annual Report; Tonglian Precision 2025 Annual Report; FRT 2025 Annual Report; BLT 2025 Annual Report; Farsoon 2025 Annual Report. Overseas company reports: GKN Powder Metallurgy (under Melrose Industries) 2025 Annual Report; Höganäs AB 2025 Annual Report; Sandvik AB 2025 Annual Report Sandvik Materials segment; Sumitomo Electric Industries 2025 Annual Report Powder Metallurgy Division; Proterial Ltd. (former Hitachi Metals) 2025 Annual Report Magnetic Materials Segment.

Industry level: CGMCIA Powder Metallurgy Branch 2025 Industry Statistical Bulletin; China Power Supply Society 2026 Metal Soft Magnetic Core White Paper; China Additive Manufacturing Industry Alliance 2025 Annual Report; SmarTech Analysis 2025 Global Metal Additive Manufacturing Report; European Powder Metallurgy Association (EPMA) Annual Industry Review 2025; Metal Powder Industries Federation (MPIF) State of the PM Industry 2025; Japan Powder Metallurgy Association Annual Statistics 2025.

Downstream application level: Counterpoint Research global foldable phone shipment monthly reports 2025-2026; IDC global smartwatch quarterly reports 2025-2026; Bloomberg New Energy Finance (BNEF) Electric Vehicle Outlook 2025; International Energy Agency (IEA) World Energy Investment 2026; Nikkei Asia coverage of Asian powder metallurgy supply chains; Reuters coverage of NVIDIA AI server power standards; COMAC C919 and C929 program progress; COMAC 2025 Annual Sustainability Report; AECC 2025 public technical reports.

Plant level: Tianxia Gongchang (4.8 million in-production factory B2B platform) plant capability data reverse-indexed by process, material, capacity, customer qualification, and region. The platform supports keyword searches such as powder metallurgy, metal injection molding, press-and-sinter, metal soft-magnetic, soft-magnetic core, Sendust core, amorphous nanocrystalline, 3D printing metal powder, spherical titanium powder, superalloy powder, water-atomized iron powder, gas-atomized alloy powder, carbonyl iron powder, selective laser melting, electron beam melting, hot isostatic pressing, phone hinges, titanium MIM, stainless MIM, common-mode inductors, PFC inductors, on-board charger, PV inverter, storage converter, AI server power, aero engine, orthopedic implants, titanium structural parts, precision gears, foldable hinges.

Report scope disclaimer: market sizes, unit prices, capacities, revenues cited are aggregated from public channels and research-institute estimates; they may differ from companies' own disclosures. 3- to 5-year judgments are independent research-institute calls, not investment advice.

Further reading suggestions: interested readers should explore: (1) head companies' annual reports, quarterly reports, earnings calls, IR records — most authoritative company-level source; (2) CGMCIA Powder Metallurgy Branch, China Additive Manufacturing Industry Alliance, China Power Supply Society statistical bulletins, industry research reports, technical white papers; (3) EPMA, MPIF, Japan Powder Metallurgy Association statistics and tech reports; (4) SmarTech Analysis, Wohlers Associates annual reports — most authoritative 3D-printing metal powder data; (5) Counterpoint Research, IDC, Canalys consumer-electronics market data — key for MIM downstream demand; (6) BNEF, IEA, Bloomberg, Reuters new-energy data — key for metal soft-magnetic downstream; (7) aerospace journals: Aerospace, Aviation Week, Journal of Aerospace Power, Propulsion Technology — important for 3D-printing aerospace applications; (8) medical-device journals: Medical Device and Diagnostic Industry, China Medical Devices — important for MIM and 3D-printing medical applications.

The research institute will continue tracking powder metallurgy developments and update analysis regularly. Next-issue topics: "Deep penetration of metal soft-magnetic cores in AI compute power chains," "3D-printing metal powder domestic substitution progress in COMAC C929," "Granular sizing of humanoid robots' MIM demand." We welcome value-chain firms, investors, and researchers to continue following.

Research institute methodology rests on four principles: First, "facts first" — all judgments built on publicly verifiable facts, avoiding subjective speculation; second, "data cross-validation" — key data cross-checked across multiple independent sources to reduce single-source bias; third, "complete logic chain" — every judgment with full reasoning, supporting reader scrutiny and challenge; fourth, "clear time dimension" — all data and judgments tagged with explicit time windows to avoid "time-misalignment" misleading. Based on these principles, the institute aims to deliver "credible, readable, usable" industry research, serving investment decisions, industry operations, policy research, and academic reference.

Powder metallurgy, as a key basic material process in China's advanced manufacturing and the "manufacturing power" strategy, holds a unique strategic position. This "forming from powder" business, though not as glamorous as EVs or AI compute, is the key base supporting those bright industries. To understand powder metallurgy is to understand China's manufacturing "root." This "root" grew quietly for decades, only recently rediscovered by capital markets as EVs, AI servers, COMAC, and consumer electronics exploded over the past five years. Over the next 5 to 10 years it will keep growing, deepening, and nourishing the overall upgrade of Chinese manufacturing. This is our most plain and firm view of China's powder metallurgy.

The institute's long-term view of powder metallurgy is positive yet sober. Positive: ample tech reserves, diverse downstream, friendly policy, complete talent pipeline, ongoing capital-market rerating. Sober: intense global competition, capacity expansion overflow risk, faster tech iteration, geopolitical uncertainty. On balance, we lean toward seeing this industry maintain high sentiment over the next 5 to 10 years, breeding several global-class head firms and becoming a key pillar of Chinese manufacturing's leap to the high end. To readers who reached this far: if you are an industry person, may this report inform your operations; if an investor, an independent angle on judgment; if a researcher, facts and data underpinning your work; if a policymaker, first-hand industry observation; if a general reader, a different understanding of Chinese manufacturing's "foundation." Powder metallurgy, seemingly obscure yet truly critical, deserves more attention and study. We look forward to ongoing exchange with all kinds of readers, progressing together on this industry-research path. The institute believes that truly deep industry research is not chasing hot topics but seeing those overlooked basic industries clearly, explaining them plainly, researching them thoroughly. Powder metallurgy, as a representative of such "foundational yet critical" industries, is one of the institute's long-term tracking priorities. This "forming from powder" process line walked a century, accelerates now, and will keep going. We will, with industry peers, witness and record this business's next chapter, capturing every process breakthrough, every new-scene eruption, every company's rise with rigorous research methods.