Furuya Metal Co., Ltd. (7826.T) - Deep Dive Research Report

Prepared 23 June 2026. Fiscal year ends 30 June. Most recent reporting period: Q3 FY2026 (nine months ended 31 March 2026), disclosed 13 May 2026.

1. What the Company Does

Furuya Metal makes things out of two of the hardest-to-handle metals on the periodic table - iridium and ruthenium - and then takes those things back when they wear out, dissolves them, and makes them again. That closed loop is the whole business in one sentence.

Iridium and ruthenium are platinum-group metals (PGMs). They are vanishingly rare (iridium global mine supply is only a handful of tonnes a year, a by-product of South African platinum mining), and they are brutally difficult to work with. Iridium melts at about 2,446 degrees Celsius, the second-highest melting point of any usable metal, and it is brittle, so you cannot simply cast or forge it the way you would steel. Ruthenium is notorious for being nearly impossible to dissolve once it has been used, which is exactly the problem you must solve to recycle it. Only a small handful of companies on earth can melt, shape, sputter, and re-refine these metals to the purity that semiconductor and data-storage customers demand. Furuya is the largest of them, and on several specific products it has no real second source.

The company traces its roots to Furuya Shoten, a Tokyo precious-metals trading shop founded in 1951 that initially dealt in jewelry. The decisive pivot came in the 1970s, when the business moved from selling precious metal to fabricating it for industry, and in 1981 it achieved the milestone that defined everything after: manufacturing Japan's first domestically produced iridium crucible. That crucible is a container that can hold molten oxide at temperatures where almost nothing else survives, and it is used to grow the single crystals (sapphire, and oxide crystals for surface-acoustic-wave filters in phones) that sit inside electronics. From that one product, Furuya built outward into every high-temperature, high-purity niche where iridium and ruthenium are the only answer.

The value proposition is narrow and deep. Furuya does not try to be a broad precious-metals house like Tanaka or Umicore. It picked the two PGMs that are the hardest to process and the most supply-constrained, became the world's best at processing them, and then wrapped a recycling operation around its own products so that customers send their spent material back rather than to a refiner. The recycling loop is not a side business - it is the moat. It lowers the customer's effective metal cost, guarantees supply security in a market where the metal itself can quadruple in price, and locks the customer into Furuya's ecosystem.

"Only a small number of companies worldwide have the technological capability to process and refine these metals to the highest levels of purity and precision." - Furuya Metal corporate materials

A concrete example: a hard-disk-drive maker needs a ruthenium sputtering target, a disc of ultra-pure ruthenium that gets vaporised atom-by-atom inside a vacuum chamber to lay down the magnetic recording layer of every nearline data-centre drive. Furuya casts and machines that target to spec, ships it, and when the target is spent (a sputtering target is only ~30-50% consumed before it must be replaced), the customer returns the used puck. Furuya dissolves the otherwise-insoluble ruthenium, re-refines it to purity, and makes a new target. The customer never has to source virgin ruthenium on the open market, and Furuya keeps the metal circulating inside its own system.

2. Business Segments

Furuya reports through five segments. They are not really five different businesses so much as five different shapes that the same two metals take, plus a metal-trading arm. The technology core (melt, shape, purify, recycle iridium and ruthenium) is shared; the segments differ by what the final product is and which industry buys it.

Electronics (iridium compounds and crucibles for displays and crystal growth)

This segment supplies the iridium that ends up inside displays and single-crystal growth. Two things sit here. First, iridium organometallic compounds (phosphorescent iridium complexes) that are the emitter materials in OLED displays - Furuya supplies the high-purity iridium precursor that the OLED material formulators build on, and it holds roughly a 90% global share of this iridium-compound niche. Second, iridium crucibles used to grow oxide single crystals - the sapphire and SAW-filter substrates that go into LEDs, optical components, and the RF front-end of smartphones. The core capability is the ability to make iridium hold its shape and purity at crystal-growth temperatures for thousands of hours. The 90% share tells you replication is essentially not happening. Management talks about this as a high-margin, technology-defended core, and in the FY2026 upgrade it was named as one of the two segments (with Thin Film) driving the profit surge.

Thin Film (sputtering targets for HDDs and semiconductors)

This is the ruthenium franchise and arguably the company's centre of gravity for the AI-data-centre thesis. Furuya makes ruthenium sputtering targets used to deposit the magnetic underlayer in perpendicular-magnetic-recording hard drives, holding roughly a 70% global share. It also makes ruthenium and platinum-alloy targets for the read/write magnetic heads. The same segment is now positioned for two newer waves: ruthenium targets for next-generation semiconductor wiring (the industry is moving advanced-node interconnect from copper toward ruthenium because ruthenium gives lower resistance at extreme miniaturisation), and ruthenium targets for EUV photomask blanks. The core capability is purity and grain structure - a sputtering target must be metallurgically uniform or it throws particles and ruins wafer yield, which is why qualifying a new target supplier takes a fab a very long time. This is the growth bet of the group: HDD demand for nearline data-centre drives is forecast to stay firm to around 2030 as HAMR (heat-assisted magnetic recording) drives ramp, and the semiconductor-wiring use is a genuinely new market on top.

Thermal (temperature sensors / thermocouples)

Furuya makes platinum- and iridium-based thermocouples - precision temperature sensors used inside semiconductor manufacturing equipment, where wafers are processed at tightly controlled high temperatures. It holds roughly a 70% domestic (Japan) share in semiconductor-equipment thermocouples. The capability here is that a thermocouple for a deposition or diffusion furnace must be exact and stable over long runs; an out-of-spec sensor scraps wafers. It exists as its own segment because the product, the customer (semiconductor equipment OEMs) and the qualification cycle differ from the metal-shaping businesses, even though the underlying metallurgy is shared. It is a steady, capacity-and-utilisation business that rides the semiconductor capex cycle.

Fine Chemicals / Recycling (refining, catalysts, green hydrogen)

This is the loop that makes the rest of the model work, plus the optionality. The recycling-and-refining operation recovers iridium and ruthenium (including from non-Furuya scrap) at high purity - critically, Furuya developed proprietary technology to dissolve ruthenium, the step everyone else struggles with. On top of refining sits the catalyst business: iridium-oxide catalysts for PEM (proton-exchange-membrane) water electrolysis, the technology that splits water into green hydrogen. Furuya is an established, academically-benchmarked supplier of iridium electrolysis catalyst and has signalled ambitions to scale PEM-catalyst capacity materially this decade. The core capability is the chemistry of getting these metals into and back out of solution at purity. Strategically this segment is both the cash-saving backbone (cheaper metal for every other segment) and the long-dated call option (green hydrogen).

Supply Chain Support (metal procurement / trading)

This is the metal-trading and procurement-security arm: Furuya sources and supplies raw iridium and ruthenium to help customers secure material, and the segment's revenue moves with metal prices and trading volume. It is the most volatile line in the group - in H1 FY2026 it grew roughly 460% year on year as iridium and ruthenium prices spiked - and it inflates reported revenue without carrying the same margin profile as the fabrication segments. It exists because in a market where the metal can quadruple, controlling procurement is itself a service customers will pay for, and it gives Furuya market intelligence on metal flows. Read this segment as a price-and-volume amplifier, not a quality-of-earnings driver.

3. Products and Business Detail

The full catalogue follows the two metals. From iridium: crucibles for oxide single-crystal growth (the 1981 founding product, ~70% global share for SAW-filter and sapphire growth), iridium organometallic compounds for OLED emitters (~90% share), iridium thermocouples, iridium-alloy friction-stir-welding tools (commercialised 2014, the world's first iridium-alloy FSW tool, used to weld high-melting-point materials), iridium targets for HAMR magnetic heads, and iridium-oxide catalysts for PEM water electrolysis. From ruthenium: sputtering targets for HDD recording media (~70% share), targets for semiconductor wiring and EUV photomask blanks, and ruthenium chemicals/catalysts. Across both: a refining and recycling service that recovers spent material to high purity.

What makes these hard to make is physical and tacit. Iridium's ~2,446 degree melting point and brittleness mean you cannot machine it conventionally; getting a defect-free crucible or a metallurgically uniform sputtering target requires process knowledge accumulated over decades, much of it not written down. Ruthenium's resistance to dissolution makes recycling a genuine chemistry problem that Furuya solved proprietarily. Then there is qualification: a sputtering target or a thermocouple that goes into a semiconductor fab or an HDD line must pass long customer qualification, because a particle defect scraps wafers or a sensor drift scraps a furnace run. Process knowledge plus qualification plus raw-material access is the three-layer barrier.

Manufacturing is concentrated in Japan. The Tsukuba Plant (1990) is the large-scale manufacturing hub; the Tsukuba R&D Center (1998) drives thin-film technology; the Tsuchiura Plant (2007) doubled ruthenium refining capacity to roughly 15-20 tonnes a year; the Chitose Plant (2010) brought quartz-tube production in-house. The 2024 merger with Nano Cube Japan added nano-alloy production capability. Overseas, Furuya opened a South Korean subsidiary in 2011 (close to display and memory customers), a North American base in New Hampshire in 2013, and in 2025 it took full ownership of its Shanghai joint venture, establishing a wholly-owned China presence.

Milestones that changed the business: 1981 (first domestic iridium crucible, the founding act), 2006 (JASDAQ listing, the first industrial precious-metal maker to list there, funding global expansion), 2011 (capital alliance with Tanaka Kikinzoku Kogyo and first overseas subsidiary), 2014 and 2020 (selected as a METI "Global Niche Top" company, a government recognition of dominant niche share), 2022-2023 (moved up from TSE Standard to TSE Prime Market), and the 2024 Nano Cube merger. Raw-material access is secured partly through a relationship with Sibanye-Stillwater, a South African PGM miner and a roughly 4.7% shareholder, which matters because all the process skill in the world is worthless without metal to feed it.

4. Customers

The customers are the world's data-storage, display, semiconductor, and (increasingly) hydrogen industries, and they buy because in most cases there is no equally good alternative.

For the Thin Film segment, customers are the hard-disk-drive makers (the industry is effectively Seagate, Western Digital and Toshiba) and their component suppliers, plus advanced-logic semiconductor fabs evaluating ruthenium wiring. The buying decision sits with process/materials engineering, and the criterion is qualification: a target that produces uniform films with no particle defects. The sales cycle is long because qualifying a new target supplier on a production line is expensive and risky, which is exactly why Furuya's installed base and ~70% share are sticky. Switching away means re-qualifying, and few customers will do that to save a few percent on a non-bottleneck input.

For Electronics, customers are OLED material formulators and display makers (who buy the iridium emitter precursor) and crystal growers serving LED and smartphone-RF (SAW-filter) supply chains. The ~90% share in OLED iridium compounds reflects how few players can deliver that purity, and the buyer values supply security as much as price because iridium is supply-constrained.

For Thermal, customers are semiconductor-equipment OEMs designing furnaces and deposition tools; the sensor is qualified into the tool design, so once Furuya is in, it stays for the tool's life.

For Fine Chemicals / catalysts, the emerging customer set is electrolyzer manufacturers building PEM hydrogen capacity, where iridium loading is a known supply bottleneck for the whole industry.

The reason they choose Furuya is consistent across segments: capability (can anyone else even make this to spec), supply security (Furuya's recycling loop and Sibanye relationship insulate the customer from metal-price and metal-availability shocks), and the closed loop itself (returning spent material to Furuya lowers the customer's net metal cost). Switching costs are high - qualification testing, installed-base lock-in, and the loss of the recycling economics. Concentration exists (HDD is a three-customer industry) but reads more as a reflection of Furuya's quality position than as a fragility, since the customers cannot easily replace it. Contract structure is a mix of ongoing supply relationships in fabrication and price-and-volume-driven spot business in Supply Chain Support; the metal-trading line makes total revenue less predictable than the fabrication margin would suggest.

5. Competitive Landscape

The structure of this industry is a small number of global PGM houses, of which only a few can do iridium and ruthenium fabrication at the top tier, and Furuya is the specialist that went deepest on exactly those two metals.

The most direct competitor on iridium crucibles and PEM catalysts is Heraeus Precious Metals (Germany, private, family-owned). For broad PGM refining and catalysis, Johnson Matthey (UK) and Umicore (Belgium) are the large diversified players, though their centre of gravity is autocatalysts and chemistry rather than Furuya's electronics-grade iridium/ruthenium fabrication. Tanaka (the TANAKA Precious Metal group, Japan, private) is simultaneously a competitor in PGM products, a major raw-material supplier to Furuya, and a major shareholder (a roughly 17% relationship through TANAKA Precious Metal Technologies) - a tangled relationship where the same counterparty is rival, supplier, and owner. In sputtering targets specifically, JX Advanced Metals (Japan, TSE-listed) and Tosoh (Japan, via Tosoh SMD) are the relevant rivals, and Materion (US) competes in advanced materials and targets.

Furuya wins on focus and on the recycling loop. Against the diversified giants it wins because iridium/ruthenium electronics fabrication is its entire reason for existing, not one product line among hundreds, and because its closed-loop recycling lowers customers' metal cost in a way a pure fabricator cannot. It is most exposed where a competitor has equal process capability and deeper pockets - Heraeus on iridium crucibles and PEM catalyst, and the large target makers (JX, Tosoh) in semiconductor sputtering, where Furuya is challenging into adjacent share rather than defending a 70% incumbency.

Barriers to entry are genuinely high: the metallurgy of a ~2,446 degree metal, the chemistry of dissolving "insoluble" ruthenium, decades of undocumented process knowledge, customer qualification cycles, and access to scarce raw metal. A new entrant cannot buy its way past all four at once. The structural shift to watch is ruthenium's possible adoption in advanced-node semiconductor wiring and EUV masks, which could expand the pie and bring Furuya into more direct contact with the big semiconductor-materials incumbents.

Market-cap figures are approximate peer-size references as of June 2026 and move with the market; private companies marked "-".

6. Industry

Demand for Furuya's products is driven by some of the strongest secular trends in technology hardware: data-centre storage growth, OLED display adoption, advanced semiconductor manufacturing, and (longer-dated) green hydrogen. The unifying thread is that all of these need tiny quantities of an irreplaceable metal at extreme purity.

The supply side is the defining feature. Iridium and ruthenium are by-products of platinum-group-metal mining, roughly 80% of which is in South Africa. Iridium primary supply is only a few tonnes per year globally; ruthenium is in the low tens of tonnes. Because supply is fixed by platinum-mining economics rather than by iridium/ruthenium demand, prices are extraordinarily volatile: iridium rose from about ¥23,300 per gram in late December 2025 to about ¥33,200 per gram by mid-February 2026, and ruthenium from about ¥4,660 to about ¥7,190 per gram over the same weeks. This volatility is the single most important industry fact for Furuya: it inflates and deflates reported revenue (especially in Supply Chain Support), and it makes the recycling loop and procurement security genuinely valuable services rather than nice-to-haves.

Industry size at the application level is large and growing. HDD demand is concentrating into high-capacity nearline data-centre drives - more than 90% of HDD units shipped are forecast to be nearline by 2030 - with HAMR drives ramping toward 36-44TB by 2026 and high-volume HAMR shipments expected from 2027, each generation needing ruthenium recording layers. OLED continues to take display share from LCD. Advanced-node logic is beginning to substitute ruthenium for copper in interconnect because ruthenium scales better at extreme miniaturisation. Green-hydrogen PEM electrolysis is a multi-billion-dollar emerging market in which iridium is a recognised bottleneck.

Furuya sits at a specific point in the global supply chain: between the PGM miners/refiners (Sibanye-Stillwater, the South African producers, and partly Tanaka) and the device makers, as the firm that converts raw metal into qualified, high-purity functional components and then recycles them. There is little "import substitution" framing because this is a global, Japan-centred specialty niche rather than a domestic-versus-import market. Regulation is light on the product side but the metal supply is geopolitically concentrated. Cyclicality is real and double-layered: the semiconductor/HDD capex cycle drives volumes, while the metal-price cycle drives both revenue optics and customer behaviour. The current tailwind is the AI-data-centre storage build and the metal-price spike; the structural headwind is that the entire business rests on a scarce, price-volatile, geographically concentrated raw material.

7. Growth Triggers

All items below are drawn from Furuya's FY2025-FY2026 earnings briefings (the company is a Japanese mid-cap that publishes Japanese-language results briefings and disclosure documents rather than English transcripts; statements are taken from those briefings and the associated disclosures).

• Iridium melting-furnace demand for high-melting-point glass used in smartphones and semiconductor substrates is expected to expand (H1 FY2026 briefing, 9 Feb 2026).

President Takao Furuya outlined growing demand for iridium-based melting furnaces used in high-melting-point glass for smartphones and semiconductor substrates (H1 FY2026 results briefing, 9 Feb 2026).

• Ruthenium adoption as advanced-semiconductor wiring material - moving interconnect from copper toward ruthenium at advanced nodes (H1 FY2026 briefing, 9 Feb 2026; a theme repeated across briefings as the AI-data narrative built through FY2026).

• Expansion of ruthenium EUV photomask target supply for EUV lithography (H1 FY2026 briefing, 9 Feb 2026).

• HDD ruthenium target demand firm to around 2030, driven by nearline data-centre drives and the HAMR transition, with iridium targets supplied for next-generation HAMR magnetic heads (H1 FY2026 briefing, 9 Feb 2026, reiterating the data-centre demand thesis).

• Electronics and Thin Film strength from data-centre and telecom capex named explicitly as the driver of the FY2026 upward revision (Q3 FY2026 disclosure, 13 May 2026).

• PEM water-electrolysis (green hydrogen) catalyst scale-up - iridium-oxide catalyst capacity expansion as a longer-dated growth line (FY2025-FY2026 briefings; flagged as strategic optionality).

• China full subsidiary (Shanghai, 2025) and Nano Cube nano-alloy capability (2024 merger) broadening manufacturing reach and product capability (FY2025 results, 7 Aug 2025).

8. Key Risks

Metal-price and raw-material concentration. The entire business rests on iridium and ruthenium, whose supply is a by-product of South African platinum mining and is therefore inelastic to demand. Prices can double in weeks (iridium ¥23,300 to ¥33,200/g, Dec 2025-Feb 2026). The mechanism cuts both ways: a price spike inflates Supply Chain Support revenue and can lift fabrication margins on inventory, but a sharp reversal can compress margins, and a genuine supply disruption in South Africa would starve the fabrication lines. This is a high-probability, recurring volatility risk and a low-probability catastrophic-supply risk in one. Management's own answer - the recycling loop and the Sibanye-Stillwater relationship - is a mitigant, not an elimination.

Quality of the revenue beat. The spectacular FY2026 numbers were partly driven by the Supply Chain Support trading segment (up ~460% in H1) and by metal-price tailwinds, not solely by fabrication volume. The mechanism of disappointment: if metal prices normalise, reported revenue and the trading-segment contribution fall sharply even if the underlying fabrication business is healthy, and the year-on-year optics turn ugly. A reader who anchors to the ¥88bn record-revenue forecast without separating trading pass-through from fabrication margin will misjudge the run-rate.

Customer concentration in HDD. Thin Film leans on an HDD industry that is effectively three customers. If the HAMR transition slips, if nearline demand softens, or if a customer dual-sources ruthenium targets, the segment's growth thesis weakens. Mitigated by qualification lock-in and ~70% share, but the concentration is real.

Technology-substitution risk inside the moat. The bull case for ruthenium-in-semiconductor-wiring and EUV is also a risk if it goes the other way: advanced-node roadmaps could favour a different material, or HDD could lose ground faster than expected to NAND in some tiers. Furuya's products are mission-critical but tied to specific technology choices it does not control.

The Tanaka relationship. TANAKA is supplier, ~17% shareholder, and competitor at once. The mechanism of harm is governance and strategic conflict: a major shareholder that also competes in PGM products and sits in the raw-material supply chain has interests that may not always align with minority holders. Furuya has felt the need to publicly clarify the relationship and governance, which itself signals the market's sensitivity to it.

Founder/insider ownership transition. The president-founder reduced a large block in early 2026 (see Section 11). The mechanism is signalling and overhang risk - large insider selling, even for benign reasons, can pressure the stock and raises questions about the founder family's long-term posture.

9. Walk the Talk

The six reporting periods used: H1 FY2025 (disclosed 7 Feb 2025), Q3 FY2025 (~May 2025), full-year FY2025 (disclosed 7 Aug 2025), Q1 FY2026 (~Nov 2025), H1 FY2026 (disclosed 6 Feb 2026, briefing 9 Feb 2026), and Q3 FY2026 (disclosed 13 May 2026). As a Japanese mid-cap, Furuya publishes Japanese-language results summaries and briefing materials rather than English transcripts; the assessment below is built from those disclosures.

The story this sequence tells is of a deeply conservative management that under-promised and then massively over-delivered, with the help of a metal-price tailwind. Through FY2025, the picture was mixed-to-cautious: the full year (disclosed 7 August 2025) delivered record revenue of about ¥57.4bn (+20.7%), but operating profit (about ¥9.5bn, -2.8%) and net income (about ¥6.4bn, -12.7%) fell, with management pointing to foreign-exchange, raw-material costs, and overseas-market uncertainty as the squeeze. So the message entering FY2026 was sober: top line growing, profitability under pressure.

That caution carried straight into FY2026 guidance. The company's initial FY2026 forecast (set with the August 2025 results) was for revenue of roughly ¥55bn and net income of roughly ¥4.8bn - a forecast that sat below the FY2025 actual it had just reported. In other words, management guided for a down year on profit.

What actually happened was the opposite, and the correction came fast. At the H1 FY2026 results (6 February 2026), the full-year forecast was revised dramatically upward to roughly ¥88bn revenue and ¥11.0bn net income - revenue guidance lifted by about 60% and net income guidance more than doubled in a single revision. H1 itself showed revenue of about ¥43.9bn (+63.7%), with the Supply Chain Support segment up about 460% as iridium and ruthenium prices spiked.

Furuya revised its full-year FY2026 outlook from roughly ¥55bn revenue / ¥4.8bn net income to roughly ¥88bn / ¥11.0bn (H1 FY2026 disclosure, 6 Feb 2026).

The momentum did not stop. At Q3 FY2026 (13 May 2026), the company raised again: nine-month ordinary profit reached about ¥16.9bn (roughly 2.2x year on year), the full-year ordinary-profit forecast was lifted 37.5% from ¥16.0bn to ¥22.0bn, a fresh record, and the standalone January-March quarter showed operating margin jumping from 17.3% to 32.1% year on year. The dividend forecast was raised twice in the process - to ¥120 and then to ¥155 per share.

The honest read: this is conservative-to-a-fault management whose forecasts are not a reliable guide to the upside, but who do not over-promise. The pattern (lowball the initial number, then revise up repeatedly as results come in) is credible and customer-trustworthy, but the enormous FY2026 beat owes a great deal to a metal-price spike and the trading segment, which is luck as much as execution. They do what they say on the downside (the FY2025 profit dip they flagged did happen) and they bury their own guidance on the upside. The risk for the analyst is the mirror image: the same metal-price leverage that produced the upside can reverse, and management's conservatism on the way up offers little read on how the next normalisation will land.

10. Shareholder Friendliness Index

Dividends. Furuya pays a single annual dividend that is profit-linked and therefore lumpy. On a comparable basis the trajectory has been upward but bumpy: about ¥85 per share in FY2022 and FY2023, an elevated ¥286 in FY2024 (a figure marked as "other," reflecting a special/one-off element alongside the record FY2024 profit and complicated by a 3-for-1 stock split that took effect 27 June 2024), ¥96 in FY2025, and a FY2026 forecast that was raised twice during the year (to ¥120, then to ¥155) as profit surged (FY2026 disclosures; investing-data dividend calendar; stock-split records). Reported payout ratios have run in the mid-30s percent range, so the dividend is well covered and the volatility reflects management's policy of paying out in line with a swinging, metal-price-sensitive profit rather than smoothing it. The direction of travel into FY2026 is a clear increase.

Buybacks and dilution. The MoatMap database records zero buybacks in the trailing ~90 days (since 25 March 2026), and a search of the company's disclosures and financial news for the prior three years did not surface an executed share-repurchase programme either - capital return at Furuya runs through dividends, not buybacks. Share count has moved the other way: shares outstanding rose over the period (roughly 21m to about 25m), driven by the 3-for-1 split of June 2024 (which mechanically tripled the count) and the 2024 Nano Cube merger (share-funded), so the company is a net creator of shares rather than a retirer of them. There is no evidence of buyback authorisation or execution over the last three years to offset that.

Verdict: Returns Capital - via a rising but profit-linked dividend, not buybacks; shares outstanding have grown rather than shrunk, so the capital-return story is dividend-only.

11. Insider Activities

The data below is sourced from MoatMap's cross-market disclosure database (market: Japan), which is the canonical source for recent Japanese insider/large-shareholder filings here because the underlying EDINET/TDnet large-shareholding portal is API/auth-gated. Japan's 5%-rule large-shareholder reports disclose share counts and ownership percentages but generally no execution price, so values are not shown.

Buys - reading the signal. The only buy in the window is MUFG's 1,345,900-share purchase on 30 March 2026 that pushed it above the 5% reporting line. The role label is "政策投資" - a policy/strategic (cross-shareholding) holding by a financial institution, the kind of position banks build and trim for portfolio and threshold-management reasons rather than fundamental conviction. Three weeks later MUFG sold 1,109,100 shares (20 April 2026), dropping back below 5%. Taken together, MUFG's buy-then-sell looks like mechanical movement around the 5% disclosure threshold by a financial holder, not a conviction signal in either direction. There is no open-market buying by operating insiders (directors or officers) in the window.

Sells - working out the why. The transaction that matters is the president and founder-family representative Furuya Takami selling 1,857,396 shares (7.55% of shares outstanding) on 5 March 2026. That is a large block from the person whose name is on the company. The 5%-rule filing does not disclose a reason or a price. Plausible benign explanations for a founder reducing a stake of this size include diversification/estate planning, and the broader reshaping of Furuya's ownership structure around the TANAKA relationship and the move to TSE Prime (which rewards free float) - but none of these is confirmed in the filing, so the honest statement is: reason not disclosed.

Net assessment. Insiders were net sellers in the last 12 months, but the activity is narrow (two parties) and largely mechanical: MUFG's pair of trades is threshold housekeeping by a financial holder, leaving the founder-president's 7.55% disposal as the one substantive signal. A large founder sale with no disclosed reason is a mild concern - it creates overhang and invites questions about long-term founder commitment - but it is not, on this evidence, a fundamental red flag about the business, especially against a backdrop of repeatedly raised guidance. Read it as mild concern, weighted by the absence of any offsetting operating-insider buying.

12. Scenarios

Bull case. The AI-data-centre build keeps pulling nearline HDD demand, HAMR ramps on schedule from 2027, and every new high-capacity drive needs Furuya's ruthenium recording layers and iridium head targets - and Furuya keeps its ~70% share because no one re-qualifies a target supplier to save pennies. On top of that installed base, the two new ruthenium markets arrive: advanced-node logic adopts ruthenium interconnect at scale, and EUV photomask blanks become a steady volume line, both of which the diversified giants are slower to serve at Furuya's purity. The iridium glass-melting-furnace and OLED-compound businesses ride smartphone and display demand, and the PEM hydrogen catalyst line finally scales into a real third leg as electrolyzer capacity is built out, with Furuya's recycling loop making it the low-net-cost supplier of the scarce iridium catalyst. Metal prices stay firm enough to keep the trading segment fat without choking the customers. In this world Furuya is a structurally short metal supply meeting structurally rising demand, and the conservative management keeps revising up.

Base case. Furuya remains the dominant specialist in iridium and ruthenium fabrication, defending its franchise shares in HDD targets, OLED compounds, crucibles, and thermocouples while the new ruthenium-semiconductor and EUV uses grow but more slowly and lumpily than the bull hopes. Revenue and profit stay genuinely volatile because so much of the reported top line is metal-price- and trading-driven; some years print records on a price spike, others give part of it back as prices normalise, and the fabrication margin grinds higher underneath. Management continues its lowball-then-raise pattern, the dividend stays profit-linked and broadly rising, and the Tanaka relationship persists as a manageable governance wrinkle rather than a rupture. The business compounds on the strength of an irreplaceable position in two scarce metals, without the step-change the bull case needs.

Bear case. The metal-price cycle turns hard. Iridium and ruthenium, having spiked, fall back, and the Supply Chain Support segment collapses year on year, dragging reported revenue down and exposing how much of the FY2026 "record" was price and trading rather than durable fabrication earnings - the stock de-rates as the optics reverse. Simultaneously the HDD thesis disappoints: HAMR slips, nearline demand softens, or a key drive maker dual-sources ruthenium targets, hitting the Thin Film growth engine. The much-hyped ruthenium-in-semiconductor-wiring transition stalls as roadmaps favour an alternative, removing a pillar of the forward story. The founder's large early-2026 sale proves to be the start of a sustained reduction, creating overhang, and the unresolved tension of a competitor-supplier-shareholder (Tanaka) holding ~17% surfaces as a real governance problem. None of this breaks the core franchise - the process moat is intact - but it turns a "record-breaking growth" narrative into a cyclical, governance-clouded specialty-metals story.