Sakai Chemical Industry Co., Ltd. (4078.T)

Deep Dive Research Report

Basic Materials / Specialty Inorganic Chemicals | Tokyo Stock Exchange Prime | Report date: 3 September 2026

Reporting cadence and recency check. Sakai Chemical's fiscal year ends 31 March and it reports quarterly. Working forward from that: the quarter ended 30 June 2026 (Q1 of the year ending March 2027) was due in early-to-mid August 2026. It was in fact released on 7 August 2026, with an earnings briefing held the same day and a Japanese-language Q&A transcript posted on 12 August 2026. That is the most recent reporting period and it anchors this report. The six most recent management briefings with published transcripts are used throughout, the oldest dated 2 December 2024 and the newest 7 August 2026.

1. What the Company Does

Sakai Chemical makes powders. That sounds trivial, and it is the single most important thing to understand about the business.

Specifically, it makes inorganic powders in which the particle is the product: how big each grain is, how tightly the size distribution clusters around that average, what shape the grain takes, how well-formed its crystal is, and what molecular coating sits on its surface. The chemistry itself is often century-old and unremarkable. Barium titanate, zinc oxide, barium sulfate, titanium dioxide and calcium carbonate are all textbook compounds that any competent chemistry undergraduate could synthesise in a beaker. What Sakai sells is the ability to make them at industrial scale to a specification measured in tens of nanometres, batch after batch, for customers who will reject a shipment if the distribution widens.

Three examples make the point concrete.

A multilayer ceramic capacitor (MLCC) is a stack of alternating ceramic and metal-electrode layers, printed and fired into a chip the size of a grain of sand. The ceramic layers in the highest-performance parts are now thinner than half a micron and there can be more than 1,500 of them in one chip (passive-components.eu, Q2 2026). You cannot build a 0.4-micron layer out of 0.5-micron powder grains, and you certainly cannot do it if 5% of your grains are twice the average size, because those oversized grains puncture the layer and short the capacitor. Sakai's barium titanate is sold into exactly that constraint. Its answer is hydrothermal synthesis, a wet high-pressure route that grows spherical, uniformly-sized, well-crystallised perovskite particles with what the company calls an "extremely sharp particle size distribution," which it contrasts with the more common solid-phase and oxalate routes (Electronic Materials business strategy briefing, 1 December 2025).

A sunscreen has to block ultraviolet light without making the wearer look chalky. Zinc oxide is an excellent UVA blocker, but bulk zinc oxide is a white pigment, which is why old mineral sunscreens looked like paint. Grind the particle down to roughly 20 nanometres and it stops scattering visible light while still absorbing UVA. That is Sakai's FINEX line, and the company describes its grades as the most transparent available (Cosmetic Materials business strategy briefing, 2 December 2024). The particle also has to be surface-treated, because bare nanoscale zinc oxide is photocatalytically active and will degrade the other ingredients in the formulation. Sakai's catalogue lists grades coated with hydrated silica, hydrogen dimethicone, triethoxycaprylylsilane and stearic acid, in powder form and as pre-dispersed slurries in water, cyclopentasiloxane or ester oil (company product page).

A thermal interface pad inside an AI server has to move heat from a hot die into a heatsink. Sakai's LPZINC is a deliberately large-particle zinc oxide for that job, chosen not for surface area but for two properties alumina lacks: it is soft (low Mohs hardness, so it does not chew up the customer's mixing and moulding equipment) and, because it deforms, adjacent grains touch face-to-face rather than point-to-point, widening the heat path. The company sells it as a blend, large LPZINC-30S packed together with fine LPZINC-2 and -11, because the bimodal packing conducts better than either grade alone (LPZINC product page).

The founding story, and why it still matters

The company was founded in June 1918 as Sakai Refinery, on a corner of what is still the Sakai Works in Sakai City, Osaka, and its first product was zinc oxide (company chronology). The reason it existed was regulatory and public-health, not commercial. President Toshiyuki Yagura described it this way in a 2022 interview:

"Our company was established in 1918 when the white lead issue became more prevalent in society... We replaced it with zinc oxide, and ever since then, we have continued to evolve."

• Toshiyuki Yagura, President, The Worldfolio, 14 November 2022

White lead was the dominant white pigment of the era and it was poisoning painters. Zinc oxide was the safer substitute. That origin, a company created to sell the non-toxic replacement for an established material, turns out to describe most of what Sakai does today: spherical barium sulfate and calcium carbonate sold to replace microplastic beads that Europe is banning; inorganic UV blockers sold to replace organic UV absorbers; lead-free PVC stabilisers sold to replace lead-based ones the company itself is withdrawing from; chrome-free copper hydrogenation catalysts sold to replace chromium-containing ones; titanium polymerisation catalysts sold to replace antimony.

The rest of the chronology is a slow accretion of inorganic chemistries around the original zinc plant. Barium products from 1920. Incorporation as a joint-stock company in February 1932, with the current name adopted that November. In 1935 the company started the first domestic production of anatase titanium dioxide in Japan and, to feed it, operated its own barite mine at Otaru-Matsurekura in Hokkaido from 1935 to 1971 (Japanese Wikipedia). Listing on the Osaka exchange in 1950 and Tokyo in October 1961. Plants at Onahama, Yumoto and Izumi-Kita between 1963 and 1969. The traffic-safety and road-marking business from 1975. Then a run of acquisitions that pushed the company out of pure inorganics: plastic-additive maker Kyodo Pharmaceutical, organic-chemicals businesses through the 2000s, and in 2012-2013 an entry into pharmaceuticals through Katayama Seiyakusho, with the internal pharma division folded into what became Kaigen Pharma in April 2013. Resin-additive plants in Thailand (2018) and Vietnam followed.

The company's own framing is that it holds three core technologies: powder processing, organic synthesis, and GMP-compliant pharmaceutical manufacturing, sitting on 640 patents and roughly ¥2.7bn of annual R&D as of the year ended March 2025 (Integrated Report 2025). Powder processing is the one that recurs in every segment.

The value proposition, stated plainly

Sakai sells a qualified, spec-locked powder to a manufacturer whose product will not work without it and who cannot easily re-qualify a substitute. The value is not the compound. It is (a) the tightness of the specification, (b) the fact that the customer has already spent months proving the powder works in its process, and (c) reliability of supply from a supplier that has been shipping the same grade for decades. That combination is why an MLCC maker will pay a premium for one company's barium titanate over a chemically identical powder from another, and it is also why Sakai's pricing power is real but bounded: the customer is locked in on a given grade, but the customer also has a second and third qualified source, and in one important case has now built its own.

2. Business Segments

Sakai reports eleven segments. Management is aware this is too many. Asked directly about it in May 2025, President Yagura said the structure would stay for the current plan period and be reorganised in the next one (FY March 2025 full-year briefing, 27 May 2025).

Cutting across those eleven segments, the medium-term plan "Transformation: BEYOND2030" sorts the businesses into three strategic buckets, and this is the more useful lens (mid-term plan):

• Growth: Electronic Materials, Cosmetic Materials, Organic Chemicals
• Stable: Medical, Hygienic Materials, Contract Processing, Plastic Additives, Catalysts, Titanium Dioxide & Zinc
• Considered for efficiency improvement: pigment-grade titanium dioxide (exit), domestic lead-based PVC stabilisers (withdrawal), catalysts (site consolidation)

Revenue mix percentages below are derived from the segment table in the Q1 FYE March 2027 briefing materials (7 August 2026), the most recent quarter with a full eleven-segment disclosure.

2.1 Electronic Materials (~16% of revenue)

What it does. Two product families, sold almost entirely to MLCC manufacturers, plus a third that is still small but strategically interesting.

The first family is dielectrics: the finished ceramic powder that becomes the insulating layer of the capacitor. Sakai's catalogue covers barium titanate (BT series), barium titanate zirconate (BTZ), strontium titanate (ST), calcium titanate (CT), strontium zirconate (SZ) and calcium zirconate (CZ), and every one of them is made by hydrothermal synthesis (product page).

The second family is dielectric materials: the upstream raw inputs that an MLCC maker (or a rival powder maker) buys to synthesise its own dielectric. Principally high-purity barium carbonate (BW-E1, BW-KS, BW-K) and calcium carbonate (CWS).

The third is Sciqas spherical silica, an amorphous spherical filler offered at 0.05, 0.1, 0.4 and 0.7 micron with epoxy-silane, methacrylic-silane and phenylsilane surface treatments, sold into semiconductor sealing compounds, film materials, toner, dental materials and cosmetics. Its selling points are a very narrow size distribution and low moisture absorption, which matter in encapsulants where water uptake causes package failure.

The core capability. Hydrothermal synthesis at scale. The three competing routes to barium titanate produce meaningfully different particles: the solid-phase route is the industry workhorse and gives good uniformity across a wide size range; the oxalate route gives fine-to-medium particles; hydrothermal gives the finest and most uniform. Management's claim, made at the December 2025 strategy briefing, is that within Japan the number of competitors running hydrothermal at commercial scale is limited and that Sakai has an overwhelming production-capacity advantage in that specific route.

There is a nuance worth flagging because management raised it themselves and it cuts against them. Industry perception has been that oxalate-route powders perform better in the high-temperature, high-reliability applications (AI servers, automotive), and at the December 2025 briefing Sakai's answer was that this perception is wrong and that its hydrothermal powders are in fact being adopted in those applications. That is an assertion about a live technical argument, not a settled fact, and it is the single most important open question in the segment.

Why it exists separately. It sells to a different customer set from everything else in the company (electronic component makers rather than paint, plastics, cosmetics or pharma customers), it runs a distinct process technology, and its qualification cycles are far longer. Sakai unified the dielectrics and dielectric-materials sales organisations to present one face to MLCC customers (strategy briefing, 1 December 2025).

Competitive position. Third-party market research puts Sakai at roughly 25% of the global MLCC dielectric powder market, with the top five, Sakai, Nippon Chemical Industrial, Toho Titanium, Shandong Sinocera and Fuji Titanium Industry, holding an estimated 52-58% of revenue in 2025. Japanese producers dominate the high-end; Korean, Chinese and Taiwanese producers concentrate at the low end. Management's stated view is that the technical lead of Japanese MLCC makers in the high-end is itself Sakai's defence, because it keeps the value pool in a geography where Sakai is qualified.

How it fits into the group. This is the profit engine and the reason to own the story. It carries the highest operating margin of any segment, and it is where the AI-server demand shock is landing. In the Q1 FYE March 2027 quarter it was the largest single contributor to segment operating profit.

2.2 Cosmetic Materials (~2-3% of revenue)

What it does. Two product lines aimed at very different jobs inside a cosmetic formulation.

Sunscreen actives. Microfine titanium dioxide (STR series, 15nm primary particle) for UVB, and ultrafine zinc oxide (FINEX series, 20/35/90nm) for UVA, both sold as surface-treated powders and as pre-dispersed slurries (DIS and DIF series) at 40-70wt% solids so that formulators do not have to disperse nanopowder themselves. There is also a hexagonal flake zinc oxide (XZ series, 100nm-1000nm) that boosts UV shielding and substitutes for white pigment.

Makeup sensory modifiers. Flake-shaped barium sulfate (H series, 7-20 micron) in matte-to-lustre variants; spherical barium sulfate branded Barimaru (5 micron); spherical calcium carbonate branded Calmaru (5 micron). These give the soft-focus optical blur and the press-mouldability that formulators previously got from microplastic beads. And a fluorescent line, Lumate G/R/B, which absorbs UV and re-emits it as visible green, red or blue light to correct skin tone.

The core capability. The same powder processing, but pointed at optics and skin feel rather than dielectric constant, plus surface treatment and formulation work. The company also runs FDA-compliant GMP manufacturing for the US market, which matters because sunscreen is regulated as an over-the-counter drug in the United States.

Why it exists separately. Different regulatory regime, different customers (cosmetic brands and OEM formulators rather than industrial buyers), and a sales cycle driven by brand launch calendars. Sakai created a dedicated Cosmetic Innovation division in 2023 to run it.

Competitive position and market. At the December 2024 strategy briefing the company sized the sunscreen materials market at roughly ¥110bn growing about 5% a year, and the makeup sensory-modifier market at roughly ¥20bn also growing about 5%. It put its own share of 20nm ultrafine zinc oxide at approximately 20% and described itself as the second-largest player in Japan. The 2025 Integrated Report separately states approximately 15% global share in sunscreen zinc oxide and top share in cosmetic barium sulfate. Note those two share figures are measured on different bases and should not be reconciled.

The regulatory tailwind is real and dated: ECHA's microplastic restriction began in 2023, with phase-outs running from 2027 to 2035 depending on product category.

How it fits into the group. On paper this is the second growth engine. In practice it has been the group's biggest disappointment, and Section 9 covers that in detail. It is currently loss-making and its fixed assets have been written down to almost nothing.

2.3 Organic Chemicals (~9% of revenue)

What it does. Sulfur chemistry and pharmaceutical fine chemicals. The named product families are chain transfer agents (which cap polymer chain growth so the polymer comes out at a uniform molecular weight), thiol-based epoxy resin curing agents that cure at low temperature and short cycle times, organic-sulfur photocurable resin sensitisers that let a resin cure under a low-power LED rather than a mercury lamp, organic phosphate esters used as extreme-pressure lubricant additives and antioxidants, and pharmaceutical intermediates and reagents (product page).

Two products carry the segment. The first is beta-mercaptopropionic acid (BMPA), of which Sakai is the only manufacturer in Japan per the 2025 Integrated Report. The second is the thiol monomer chemistry that goes into high-refractive-index eyeglass lenses, a genuinely niche market where the company describes itself as holding a leading position in high-quality thiol products. There is also Multhiol, a water-resistant adhesive chemistry aimed at electronic materials, which the company classifies as one of its "Smart Material" launches.

The core capability. Handling thiols. Mercaptans are foul-smelling, corrosive and awkward to purify, which is a real barrier and a reason the niches stay niches. Alongside that sits a century of organic synthesis at subsidiary Katayama Seiyakusho, which began synthesising organic and inorganic reagents in Osaka in 1922, moved into photographic-paper chemicals in 1953, and into pharmaceutical fine chemicals in 1980.

Why it exists separately. It is organic synthesis, not powder processing, and it is the one part of the chemicals business with a services model attached: Katayama does contract manufacturing of active pharmaceutical ingredients and intermediates, and the strategy is an explicit shift from CMO (make what the client specifies) to CDMO (develop the process as well).

Competitive position. Sole-source or near-sole-source in narrow chemistries, which is a strong position with a low ceiling. Management has told investors that the two halves, eyeglass lens materials and pharmaceutical intermediates, have historically been roughly 50-50 in profit contribution, and that the pharma intermediate half was shrinking on volume (H1 FYE March 2026 briefing, 1 December 2025).

How it fits into the group. The third growth leg and the most consistently profitable of the three on a margin basis, but exposed to lumpy pharmaceutical order timing. In Q1 FYE March 2027 its profit fell sharply purely because the prior year had front-loaded pharmaceutical intermediate shipments.

2.4 Plastic Additives (~15% of revenue)

What it does. Additives that make PVC and polyolefins processable. Metal soaps: zinc stearate in wet-process (SZ-2000), dry-process (SZ-P) and fine-powder (SZ-TF) grades, calcium stearate (SC-100), magnesium stearate (SM-P), and zinc stearyl acid phosphate (LBT-1830) as a mould release and glidant. Hydrotalcites: HT-P (magnesium-aluminium layered hydroxide, acid scavenger for polyolefins) and HT-7 (magnesium-zinc-aluminium, for high transparency and low discolouration). Flame retardants: MC-2010N melamine cyanurate and MGZ-6R modified magnesium hydroxide (product page).

Why it exists separately. Acquisition history and geography. This segment lives partly inside subsidiaries: Kyodo Pharmaceutical in Hadano, Kanagawa (despite the name, it makes plastic additives), Sakai Chemical (Vietnam) in Ho Chi Minh City, and Siam Stabilizers and Chemicals in Rayong, Thailand.

The strategic story. Domestic Japanese PVC demand is shrinking and Sakai is deliberately exiting lead-based stabilisers at home, while ASEAN PVC demand approaches a million tonnes a year. The plan is to move the non-lead stabiliser know-how into the Vietnamese and Thai plants and grow there. A Sakai Shoji representative office opened in India in October 2025.

Competitive position. This is the most commoditised part of the company. Named competitors in PVC stabilisers include ADEKA, Nitto Kasei, Baerlocher, BASF, SONGWON and Akdeniz Chemson. Sakai's angle is that it makes the hydrotalcite that goes into calcium-zinc stabiliser packages, so it can sell components to formulators as well as finished packages.

How it fits into the group. Cash generator, mid-single-digit-to-low-teens margin, no premium narrative. Classified as "stable" with the domestic lead business under efficiency review.

2.5 Medical Business (~10% of revenue)

What it does. Two subsidiaries. Kaigen Pharma (Osaka, Doshomachi) sells pharmaceuticals, quasi-drugs, health foods, medical devices and beauty products. Its anchor product is barium sulfate X-ray contrast medium for gastrointestinal imaging, where the company puts its domestic share at roughly 50% and describes itself as number one in Japan. It also sells Alloid G, an alginate-based ulcer treatment developed in-house, Ethoxysclerol for oesophageal varices, and the consumer "Kaigen" cold-remedy brand. Katayama Seiyakusho (Hirakata, Osaka) does contract development and manufacture of APIs and intermediates.

Why it exists separately. Completely different regulatory environment, different customers (hospitals, wholesalers, pharmacies), and a pricing mechanism set by Japanese national drug-price policy rather than by negotiation.

The vertical logic. It is not an unrelated diversification. Sakai makes pharmacopoeia-grade barium sulfate (BAX series) in its inorganic materials segment and Kaigen formulates it into contrast medium. The group is selling the same compound at two points in the chain.

Competitive position and the problem. A ~50% share of a stable, low-growth, price-regulated market with almost no new entrants is a decent cash position. But on 22 December 2023 Osaka Prefecture and Hokkaido issued administrative sanctions against Kaigen Pharma under Japan's pharmaceutical law: a 39-day suspension of its marketing-authorisation business and a 36-day manufacturing suspension at the Okuzawa plant, plus business improvement orders. Eight products had been manufactured differently from their approval dossiers, with irregularities traced back as far as 1977 (PHARM TECH JAPAN, Nikkei Medical). The company's own business strategy materials list "restoring quality culture post-regulatory action" as the segment's first strategic task.

How it fits into the group. Classified "stable," but it is currently around breakeven and it is the segment with the highest governance risk.

2.6 Titanium Dioxide & Zinc Products (~9% of revenue, shrinking by design)

What it does now. Until the end of 2025 this was the pigment business the company was founded around expanding into. Sakai ended pigment-grade titanium dioxide production in December 2025 and pigment sales in March 2026, a decision guided a year in advance and delivered on time.

What remains is the specialty half: functional (non-pigment) titanium dioxide, titanium dioxide catalyst carriers (CS-300S and CS-750 spherical grades), and the zinc oxide franchise. Zinc oxide runs from commodity grades No.1/No.2/No.3 at 0.75 micron and 99.0-99.5% minimum purity, used in tyres, rubber, paint, ink, plastics, batteries and UV-absorbing fibres, up through Superfine (sub-0.3 micron) into two branded specialties: FighZinc with antibacterial and antiviral function, and LPZINC the thermal filler.

Why the exit. Pigment titanium dioxide is a global commodity where Chinese producers set the price. Management's stated posture at the December 2024 briefing was to pursue profitability over volume rather than compete on price against Chinese product. It also had a physical dividend: exiting the pigment process sharply cut industrial waste volumes, and the company agreed to sell its industrial waste disposal site to Daiei Kankyo for approximately ¥1bn (TipRanks).

The cost. Management quantified it: in H1 FYE March 2026 the Titanium Dioxide and Zinc segment earned approximately ¥1.27bn of operating profit, and more than half of that came from the pigment grade being exited (H1 FYE March 2026 briefing, 1 December 2025). That is the profit hole the growth businesses were supposed to fill.

How it fits into the group. This was the efficiency-review centrepiece and the restructuring is done. What is left is a specialty zinc business with genuine optionality in thermal management.

2.7 Contract Processing (~8% of revenue)

What it does. Toll manufacturing. Sakai rents out its powder processing capability to third parties across pigments, colorants, functional inks, catalysts and resin additives. It runs through subsidiaries Resinocolor Industrial (pigments, colorants, functional inks, Yodogawa Ward, Osaka) and Nippon Color Industrial (contract manufacture of chemical products, Sakai, Osaka), whose competitive asset per the company's business strategy materials is an unusually varied fleet of spray-drying equipment.

Why it exists separately. It is an asset-utilisation business, not a product business. The customer owns the formulation; Sakai owns the kit.

How it fits into the group. Structurally sensible: it monetises spare capacity on equipment that has to exist anyway, and it puts Sakai inside the process of companies that might otherwise be pure competitors. Margins in the low teens. Cyclical with raw material costs.

2.8 Hygienic Materials (~7% of revenue)

What it does. Materials for disposable diapers, principally the inorganic fillers that go into breathable polyolefin films (the backsheet that lets water vapour out and keeps liquid in), plus related materials supply. The company's business strategy materials cite roughly thirty years of experience supplying this value chain and a co-development capability with converters. Delivery in Southeast Asia runs partly through PT. S&S Hygiene Solution, an Indonesian joint venture formed in 2012 between Sakai Shoji and PT Multi Spunindo Jaya (Spunindo).

Why it exists separately. Different customer (hygiene converters and brand owners), different geography weighting, and a demand driver, birth rates and ageing, that has nothing to do with electronics or chemicals cycles.

Market. The company's own figure is roughly 5.7% annual growth overseas, with domestic growth concentrated in adult incontinence rather than baby diapers.

How it fits into the group. A quietly good business. It grew sales and nearly doubled operating profit year on year in Q1 FYE March 2027, and it is one of the few segments with a demographic rather than cyclical driver.

2.9 Catalysts (~3% of revenue)

What it does. Four families. Environmental: deNOx and de-dioxin catalysts for power stations and waste incinerators. Hydrogenation: stabilised nickel (SN-250, SN-750, SN-780) for oil and fat hardening, benzene hydrogenation and carbon-carbon bond reduction; and chrome-free copper (ACA-1, ACA-3, ACA-4) for higher alcohol production, methanol synthesis, ester and aldehyde hydrogenation, and reverse water-gas shift. Polymerisation: SPC-series titanium catalysts sold as an antimony replacement for polyester fibre, film and bottle production. Carriers: the spherical titanium dioxide grades above (product page).

Two development products sit here: Ir/ENETIA for water electrolysis and a methanation catalyst, both aimed at the energy transition.

Why it exists separately. It is the one place where Sakai's own titanium dioxide feeds an internal process rather than being sold as a product, and it has industrial-plant customers with multi-year qualification and long relationships.

The recent history. Catalysts was on the efficiency-review list. Production sites were consolidated and prices revised, and management said in February 2026 that the restructuring was complete and that the business had graduated to "stable." It has, however, been flattered by one-off items: at the Q3 FYE March 2026 briefing management confirmed the Q3 catalyst result contained one-time gains worth roughly ¥500m of operating profit and that Q4 would fall back.

How it fits into the group. Small, restructured, with genuine long-dated option value in electrolysis and methanation that is not in the numbers.

2.10 Inorganic Materials / Barium and Strontium (~7% of revenue)

What it does. The oldest specialty franchise after zinc. BARIFINE, ultra-fine barium sulfate from 0.01 to 0.1 micron, used in automotive topcoats, computer housings and catalysts, with BF-1 marketed as the strongest white in the series. BARIACE, dispersible grades from 0.3 to 0.7 micron for coatings, inks and plastics. Precipitated barium sulfate grades 100 through 300 for paint, ink, plastics, rubber and brake pads. Japanese-Pharmacopoeia-compliant special barium sulfate B-1(A)/B-1(B) for cosmetics, medical devices and dental materials, and BAX series contrast-medium grade. Lamellar barium sulfate. Strontium carbonate (SW-P-N, SW-K) for glass, ceramics and capacitors, and strontium chloride (STC-D) for fluorescent materials (product page).

The core capability. Making the same compound across a hundred-fold particle size range, from 0.01 micron to 20 micron, with pharmacopoeia-grade purity available at the top end.

Competitive position. Sakai is described in third-party surveys as one of the top barium companies globally. The largest volume competitor is Hebei Xinji Chemical Group in China, which describes itself as China's largest barium salts producer and exporter with annual barium and strontium salt output above 180,000 tonnes. Nippon Chemical Industrial competes in electronic-grade barium compounds and Solvay supplies barium and strontium intermediates in Europe.

How it fits into the group. The highest-margin segment after Electronic Materials in Q1 FYE March 2027, and the structural link between the industrial business and both the cosmetic and medical businesses.

2.11 Other (~15% of revenue)

The residual carries Sakai Shoji, the group's trading arm (import, export and sale of chemical products, synthetic resins and electronic materials, based in Nakanoshima, Osaka); Osaki Industrial (chemical products, road-marking materials and electronic materials, Sakai, Osaka), which traces to the traffic-safety business started in 1975; and Joban Kasei in Iwaki, Fukushima. Road-marking material is one of the nine categories Sakai lists on its public product menu, and it is a genuinely separate business: white and yellow thermoplastic road paint sold to public-works contractors. This bucket grew strongly in Q1 FYE March 2027 and carries a mid-teens margin.

Segment summary

3. Products and Business Detail

3.1 The manufacturing logic

Almost everything Sakai sells goes through a version of the same four-step chain:

1. Synthesis. Precipitate or grow the target compound. For dielectrics this is hydrothermal synthesis in a pressure vessel. For barium sulfate it is controlled precipitation. For zinc oxide it is oxidation. The controlled variable is nucleation and growth kinetics, which set the particle size and crystallinity.
2. Particle engineering. Classify, mill, or grow to the target size distribution. Sakai's differentiation lives here: it can hold a distribution narrow at 0.01 micron in barium sulfate and at 20 nanometres in zinc oxide, and it can deliberately grow large particles when the application wants them, as with LPZINC.
3. Surface treatment. Coat the particle so it disperses in the customer's medium and stops doing things you do not want. This is why one zinc oxide appears in a dozen catalogue grades: FINEX-50 bare, FINEX-50LP standard-treated, FINEX-30S-LPT low-viscosity, FINEX-52W-LP2 passivated, FINEX-50-OTS easy-dispersion. Same compound, same size, different coating, different price.
4. Delivery form. Powder, or pre-dispersed slurry at 40-70wt% in the customer's chosen carrier (water, cyclopentasiloxane, ester oil). Shipping a dispersion moves work from the customer's factory into Sakai's, which raises the price and deepens the lock-in.

The company also runs an internal AI tool called DeepCle to quantify microscopic and non-visible particle features, described by President Yagura in the 2022 Worldfolio interview.

3.2 Manufacturing footprint

The 2025 Integrated Report identifies five principal company sites: the Sakai works (Osaka, the 1918 original), Onahama and Yumoto (both in Iwaki, Fukushima), Semboku, and Otsurugi (the "Daiken" facility, completed in the 1991-2003 period, also in Iwaki). Group companies add: Kyodo Pharmaceutical (Hadano, Kanagawa, plastic additives), Katayama Seiyakusho (Hirakata, Osaka, APIs), Kaigen Pharma (Osaka, with the Okuzawa plant), Osaki Industrial (Sakai), Resinocolor Industrial (Yodogawa Ward, Osaka), Nippon Color Industrial (Sakai), Joban Kasei (Iwaki), Sakai Chemical (Vietnam) in Ho Chi Minh City, and Siam Stabilizers and Chemicals in Rayong, Thailand. Consolidated headcount was 2,355 including temporary staff as of the 2025 report. Capital is ¥21.838bn.

The Fukushima cluster is worth noting. Onahama, Yumoto and Otsurugi are all in Iwaki, which means a meaningful share of the group's inorganic production sits in one prefecture, and the company runs local advertising there (a Fukushima TV commercial launched April 2025, station video advertising at Iwaki from June 2026) which tells you something about how important local recruitment is to that cluster.

3.3 Geographies and export

Overseas selling runs mostly through Sakai Shoji rather than through direct subsidiaries, with manufacturing abroad limited to resin additives (Vietnam, Thailand) and a hygiene joint venture (Indonesia). Recent footprint moves are all Asian and all recent: a Sakai Chemical Vietnam sales office in Hanoi opened May 2025, a Sakai Shoji representative office in India opened October 2025.

Sales into Europe and the United States are largely export-led. In cosmetics that is changing: the company confirmed at the May 2025 briefing that shipments to a major European cosmetics manufacturer had begun, with the first shipment in the January-March 2025 window.

3.4 Milestones that changed the business

• 1918: zinc oxide, as the safe replacement for white lead.
• 1935: first domestic anatase titanium dioxide production in Japan; own barite mine in Hokkaido from 1935 to 1971.
• 1961: Tokyo Stock Exchange listing.
• 1975: entry into road-marking and traffic safety.
• 2012-2013: entry into pharmaceuticals; internal pharma division merged into Kaigen Pharma April 2013.
• 2018: Thai resin additive subsidiary acquired.
• September 2023: MF Material Co., Ltd. established by Fuji Titanium, Murata and Ishihara Sangyo. Not a Sakai milestone, but the most consequential single event in Sakai's most important market in a decade.
• June 2023: ¥3.0bn fourth convertible bond issued to an IA Partners fund, beginning a three-year restructuring partnership.
• December 2025 / March 2026: pigment-grade titanium dioxide production and sales ended, closing the business the company had been in since 1935.
• February 2026: new cosmetics multi-product plant for makeup materials completed, and immediately impaired.
• July 2026: the convertible bond fully converted; a ¥7.0bn buyback launched to offset it.

4. Customers

4.1 Who buys

MLCC manufacturers for Electronic Materials. Sakai does not name them, describing customers only as MLCC makers, both Japanese and international (strategy briefing, 1 December 2025). The identity of the pool is public knowledge even if the individual relationships are not: Murata (~40% of the global MLCC market), Samsung Electro-Mechanics (~18%), TDK (~12%), Taiyo Yuden (~10%), Yageo (~10%), plus Chinese makers Fenghua, Chaozhou Three-Circle and Eyang. Japan accounts for roughly 45% of MLCC production volume, Korea 20%, China and Taiwan 30%. Given Sakai's high-end positioning and Japanese base, the Japanese and Korean high-end makers must dominate its mix.

Cosmetic brand owners and OEM formulators for Cosmetic Materials. The December 2024 strategy briefing identified the customer types with deliberate vagueness: a major European cosmetics manufacturer (first adoption confirmed), Japanese majors, American cosmetics companies, and emerging Chinese local brands.

Eyeglass lens makers and pharmaceutical companies for Organic Chemicals. The pharma side is contract work, so the customer is the drug owner and the relationship is per-molecule.

PVC and polyolefin compounders and converters, increasingly in ASEAN, for Plastic Additives.

Paint, ink, plastics, rubber and brake-pad makers for Inorganic Materials.

Hygiene converters and brand owners for Hygienic Materials.

Power utilities, waste-incineration operators, oleochemical and polyester producers for Catalysts.

Hospitals, wholesalers and pharmacies for Medical, mediated by Japan's reimbursement system.

4.2 How the buying decision gets made

The buying centre differs sharply by segment, and this explains why Sakai's economics differ so much across the eleven.

In MLCC dielectrics, the decision sits with the customer's materials development organisation, not procurement. An MLCC maker's process is co-designed around a specific powder: the binder system, the tape-casting parameters, the firing profile and the electrode chemistry are all tuned to a given particle size distribution and crystallinity. Management made this point in an unusually revealing exchange at the December 2025 strategy briefing, saying that whether hydrothermal or oxalate powder is preferred depends on the individual MLCC maker's design philosophy. That is the correct description of the market: it is not one specification, it is n specifications, one per customer, each locked in by that customer's own process history. Qualification runs months and requires the customer to build, fire, and reliability-test parts. Once qualified, switching means redoing all of it and re-qualifying the finished capacitor with its customers, which in automotive means re-running AEC-Q200 reliability work.

In cosmetics, the decision sits with the brand's formulation lab, with legal and regulatory holding a veto. Criteria are transparency (does the sunscreen look white on skin), SPF/PA contribution per unit loading, formulation stability, sensory feel, and regulatory clearance in every market the brand sells into. The cycle is long and staged: sample, lab formulation, stability testing, clinical SPF testing, then a brand launch calendar. The most useful detail management gave about this was that adoption does not equal revenue, because the brand then rolls the ingredient across its portfolio over years. Asked in December 2024 how the European win would ramp, the answer was initial shipment January-March 2025, brand rollout from the following fiscal year, and third-party adoption expected by FYE March 2027.

In pharmaceutical CDMO, the buyer is the drug owner's CMC organisation and the constraint is regulatory. Once a supplier's site and process are named in a drug's approval dossier, changing them requires a regulatory filing. This is the highest switching cost in the group and it cuts both ways: it is also why Kaigen's 2023 sanction was so serious.

In plastic additives and pigments, the buyer is procurement and the criterion is delivered cost with adequate quality. Sales cycles are short and price resets annually or faster.

4.3 Why they choose Sakai

For dielectrics, three specific reasons and one general one. First, particle uniformity from hydrothermal synthesis, which is a requirement not a preference as layers thin below half a micron. Second, capacity: management's claim is that within the hydrothermal route it has an overwhelming production-capacity advantage, which matters enormously to a customer in a shortage. Third, the ability to supply both the finished dielectric and the upstream barium carbonate, which management explicitly framed as creating multiple touchpoints with the same customer (strategy briefing, 1 December 2025). Some customers make their own dielectric and buy Sakai's raw material; some buy the finished powder; Sakai sells to both.

The general reason is the one President Yagura gave: close collaboration to understand and develop against a specific customer need, which he characterised as a distinctly Japanese approach.

For cosmetics, transparency of the ultrafine zinc oxide, the breadth of surface treatments, the availability of pre-made dispersions, and FDA-compliant GMP manufacturing for the US market, where sunscreen is an OTC drug.

4.4 Switching costs, honestly assessed

They are high in electronic materials, very high in pharmaceutical CDMO, moderate in cosmetics, and low in plastic additives, pigments and commodity zinc oxide.

But there is an important qualifier that management raised themselves, and it is the kind of admission worth taking seriously. At the May 2025 briefing an analyst asked whether the hydrothermal route was failing to gain adoption while solid-phase dielectric materials grew, and the answer included the phrase that differentiation by manufacturing method has diminished (FY March 2025 full-year briefing, 27 May 2025). Read alongside the December 2025 briefing where management pushed back on the perception that oxalate powders win in high-temperature applications, the picture is of a company that has a real process advantage in a shrinking subset of the market and is arguing hard for its relevance in the growing subset.

4.5 Concentration

Sakai does not disclose customer concentration figures. Structural inference is possible and unflattering: AI-grade MLCCs are roughly 80-85% Murata and Samsung Electro-Mechanics combined. Any supplier whose growth story is AI-server dielectrics is therefore selling into a duopoly, and one half of that duopoly has taken a stake in a competing barium titanate producer.

4.6 Contract structures

Sakai has not disclosed a contract mix. What can be observed from management commentary is that pricing is negotiated and revisited rather than set by long-term formula. In Q1 FYE March 2027 the company implemented price revisions across all segments to pass through raw-material cost increases from Middle East disruption around the Strait of Hormuz, and was explicit that these were cost pass-through, not value capture (Q1 FYE March 2027 briefing, 7 August 2026). Separately, on dielectrics, management described an ongoing rather than episodic negotiation:

"Regarding price negotiations for our dielectrics and dielectric materials, for products we want purchased at an appropriate price, we continue to request price pass-through, price correction, price revision, on an ongoing basis."

• FY March 2026 full-year briefing, 25 May 2026

That is the language of a supplier with leverage but without contractual escalators, having the same argument every year.

5. Competitive Landscape

5.1 MLCC dielectric powder: a concentrated oligopoly with one new and dangerous entrant

Third-party market research puts the global MLCC dielectric powder and formulations market at roughly USD 1.5bn in 2023 growing to about USD 2.4bn by 2032, a 5.5% CAGR, with barium titanate the largest chemistry (Dataintelo). Sakai is put at approximately 25% of that market, and the top five (Sakai, Nippon Chemical Industrial, Toho Titanium, Shandong Sinocera, Fuji Titanium Industry) at an estimated 52-58% of 2025 revenue.

Where Sakai wins. Against the Chinese producers, on high-end specification: the technology gap in MLCCs themselves (sub-0.5 micron layers and 1,500+ layers in Japan and Korea versus roughly 1 micron and 1,000 layers in China) means the premium powder demand sits with customers Sakai is already qualified at. Against most Japanese peers, on scale in the hydrothermal route specifically.

Where Sakai is exposed. Two places.

First, Nippon Chemical Industrial is a focused competitor with roughly 49% of its revenue in functional products including barium titanate, battery materials and electronic materials, versus roughly 47% in basic chemicals (phosphorus products, chromium salts, silica). It has high domestic share in phosphorus and chromium salts funding it, and it has been aggressive on shareholder return, targeting a 40% total payout ratio and raising its dividend from ¥92 to ¥120 (kabutabi comparison note). Its overseas sales are only around 12%, so it is more domestically concentrated than Sakai.

Second, and far more important, MF Material Co., Ltd. In September 2022 Murata Manufacturing, Ishihara Sangyo Kaisha and Fuji Titanium Industry signed a basic agreement to build barium titanate capacity together; the joint venture was formally established on 1 September 2023, headquartered in Nobeoka City, Miyazaki Prefecture, with Fuji Titanium the majority holder and Murata and Ishihara as minorities (reported splits vary between 55/35/10 and 70/20/10). Its stated business covers manufacture and sale of barium titanate, raw material procurement, quality assurance and productivity improvement, and it plans to bring a new production facility online by 2027 (Murata release, passive-components.eu).

This is the most important competitive fact in the report. The largest MLCC maker in the world, holding roughly 45% of AI-server MLCC share, has taken an equity position in a barium titanate producer and is adding capacity into a market where Sakai is the largest merchant supplier, timed to arrive in 2027. Sakai has not publicly addressed it in the six briefings reviewed here. When asked at the December 2025 strategy briefing about the risk of MLCC makers internalising production, management's answer was narrower than the question: that in-house hydrothermal production among MLCC makers is limited, that Korean makers do some internal synthesis, and that broader internalisation had not been observed. That is a true answer about the hydrothermal route. It does not address a majority-Fuji-Titanium joint venture with Murata inside it.

5.2 Titanium dioxide: a market Sakai chose to leave

Ishihara Sangyo Kaisha has the largest titanium dioxide production volume of any company in Japan as of April 2025 and is the only Japanese producer using the chloride process (ISK). Tayca Corporation, founded 1950 and also headquartered in Osaka, competes in both pigment and functional/cosmetic grades. Globally the pigment market is set by Tronox, Chemours, Kronos and the Chinese producers led by Lomon Billions. Sakai's response was to leave the pigment segment rather than fight on cost, which on the evidence was the right call and was executed on schedule.

5.3 Cosmetic ultrafine oxides

Sakai puts itself second in Japan in 20nm ultrafine zinc oxide with roughly 20% share (December 2024 strategy briefing). The competitive set overlaps heavily with titanium dioxide: Tayca and Ishihara Sangyo both supply cosmetic-grade inorganic UV materials, and international suppliers compete in surface-treated and dispersed forms. But the real competitor here is not another powder maker. It is organic UV absorbers, and they are winning on price. Management's own diagnosis of the cosmetics impairment named "increased adoption of cheaper organic UV absorbers" against its premium inorganic positioning (Q3 FYE March 2026 briefing materials, 12 February 2026).

5.4 Barium and strontium

Hebei Xinji Chemical Group, founded 1947 with 5,000 employees and over 180,000 tonnes a year of barium and strontium salts, is China's largest producer and exporter and sets the volume price. Nippon Chemical Industrial competes at electronic grade; Solvay supplies Europe. Sakai's defence is the specialty tail: 0.01 micron BARIFINE, pharmacopoeia grade, flake and spherical morphologies, none of which is a volume commodity.

5.5 Plastic additives

The most crowded market Sakai is in. ADEKA, Nitto Kasei, Baerlocher, BASF, SONGWON and Akdeniz Chemson are the named competitors in PVC stabilisers, with calcium-zinc chemistry taking around 40% of 2025 volume as lead is phased out under EU REACH and Chinese standards. Sakai's position is component supply (hydrotalcite into calcium-zinc packages) plus ASEAN manufacturing presence, not technology leadership.

5.6 Barriers to entry

Genuine but uneven. In MLCC dielectrics the barrier is customer qualification plus process know-how plus capital, and it is high enough that the merchant market has stayed a five-player oligopoly for years. It is not high enough to stop a customer-backed joint venture, which is precisely what happened. In pharmaceutical CDMO the barrier is regulatory filing lock-in and it is very high. In cosmetics the barrier is formulation qualification and regulatory clearance, moderate. In plastic additives, pigments and commodity zinc oxide there is effectively no barrier beyond cost position.

5.7 Structural shifts underway

Three at once. AI-driven bifurcation in MLCCs: high-end AI and automotive grades are tight and expensive while commodity grades sit near balance, described as a "K-shaped recovery." Customer vertical integration upstream, via MF Material. Regulatory substitution running in Sakai's favour in three places at once (microplastic beads, lead stabilisers, organic UV absorbers on environmental grounds) and against it in one (organic UV absorbers winning on cost).

5.8 Competitor comparison

Market caps are peer-size references only, drawn from public quote pages on the dates shown, and they move.

6. Industry

6.1 What drives demand

Sakai sits behind five largely independent demand engines, which is the single best structural feature of the company.

Electronic component build rates, for Electronic Materials. This is currently the dominant swing factor and it is in an unusual state.

Sun-care and colour-cosmetics consumption plus chemical regulation, for Cosmetic Materials. The company's figures: sunscreen materials ~¥110bn growing ~5%, makeup sensory modifiers ~¥20bn growing ~5%.

Vision correction and pharmaceutical development, for Organic Chemicals. The company's assumptions: thiol products for eyeglass lenses growing ~3%, pharmaceutical API and intermediate demand ~6%, high-refractive-index lenses ~6%.

Construction and infrastructure PVC consumption, for Plastic Additives. Declining in Japan, growing in ASEAN toward a million tonnes a year.

Demographics, for Hygienic Materials and Medical. Roughly 5.7% annual overseas growth in hygiene materials; domestic growth in adult incontinence rather than baby diapers; a stable but drug-price-controlled medical market.

6.2 The MLCC cycle, and why 2026 is not a normal year

This deserves detail because it is the whole near-term story.

A general-purpose enterprise server uses about 2,500 MLCCs. An eight-GPU AI server uses 15,000 to 25,000. An NVIDIA GB200 NVL72 rack consumes roughly 440,000 MLCCs. Next-generation platforms are quoted at roughly 330,000 (NVIDIA VR300) and 600,000 (Google TPU v8) passive components per rack (passive-components.eu, Q2 2026).

The parts AI servers need are high-capacitance X5R/X7R devices in small 0402 and 0201 footprints at 10-100 microfarads and 2.5-16 volts, with ultra-low equivalent series inductance. Those are the hardest MLCCs to make. Yields run around 40% and production cycles are roughly twice as long as standard parts, which is why AI servers consume only 2-3% of global MLCC units but roughly 10% of global capacity.

The result through 2026 is a genuine shortage in the high-end. Book-to-bill ratios at Murata, Samsung Electro-Mechanics and Taiyo Yuden reached 1.30, 1.31 and 1.25 respectively by late June 2026, the highest since the pandemic. Lead times for AI and automotive grades stretched from 8-10 weeks to 20-40 weeks. All three leaders raised prices 15-35%, with Samsung Electro-Mechanics implementing a 30% increase from 1 August 2026 and Taiyo Yuden reportedly targeting 1 September (TrendForce, 29 July 2026). New supply is constrained: Murata's Izumo plant is not expected at full capacity until 2027. Murata's own forecast has MLCC shipments into AI servers compounding at roughly 30% to about 3.3 times 2025 levels by 2030.

Sakai's own read, given at the Q3 FYE March 2026 briefing on 12 February 2026, was that the capacitor market grew about 10% year on year. That is a company-level observation about its whole customer base, and it sits well below the AI-specific numbers above, which is exactly the point: the growth is concentrated in a narrow slice of parts.

6.3 Where Sakai sits in the chain

Two steps upstream from the AI server. Sakai sells barium carbonate and barium titanate to an MLCC maker, who makes a capacitor, which goes onto a power-delivery board next to a GPU. That distance has consequences. It insulates Sakai from the extreme spot-price behaviour in finished MLCCs (individual scarce part numbers have reportedly moved three to ten times), because Sakai negotiates powder prices with a handful of large industrial buyers rather than selling into a spot market. It also means the volume signal reaches Sakai with a lag and gets smoothed by customer inventory, which is exactly what management described when Q3 FYE March 2026 dipped on customer inventory adjustment before recovering in Q4.

6.4 Regulation

Regulation is unusually load-bearing for this company, and it runs in several directions.

Microplastics. ECHA's restriction on intentionally added microplastics began in 2023, with product-category phase-outs from 2027 to 2035. This is the entire commercial rationale for Barimaru and Calmaru, and management's read is that the driver is not only the legal mandate but brand risk aversion, producing what they called a drastic shift from microplastic beads toward silica, with secondary concern about silica itself opening a further opportunity for inorganic alternatives (Cosmetic Materials strategy briefing, 2 December 2024).

Lead. The EU banned lead stabilisers in most consumer PVC in 2023 and Japan, the US and China are tightening progressively, which is both why calcium-zinc took ~40% of 2025 stabiliser volume and why Sakai is withdrawing from domestic lead-based stabilisers.

Sunscreen as a drug. In the United States sunscreen actives are regulated as OTC drugs, which is why FDA-compliant GMP manufacturing appears in Sakai's cosmetics capability list rather than a quality certificate.

Pharmaceutical GMP. The Kaigen sanction in December 2023 is the clearest demonstration in this report of how sharp regulatory teeth are in Japan: a 39-day marketing-authorisation suspension and a 36-day plant suspension for approval-dossier deviations, some dating to 1977.

Drug pricing. Japan's biennial price revisions structurally compress the medical segment.

Environmental permitting. Sakai operated its own industrial waste final disposal site and publishes monthly maintenance monitoring data on it. Exiting pigment titanium dioxide cut waste volumes enough to make the site saleable.

6.5 Cyclicality

Electronic Materials is the most cyclical business in the group and the swings are large. Management's own account of the FYE March 2025 recovery was that the prior year had been severely depressed and the rebound reflected capacitor-market normalisation rather than market expansion (FY March 2025 briefing, 27 May 2025). Cosmetics is less cyclical but exposed to Chinese consumer weakness, which is what broke it. Plastic additives track construction. Catalysts track utility and industrial capex. Hygiene and medical are close to non-cyclical.

6.6 Tailwinds and headwinds at industry level

Tailwinds: AI infrastructure build-out driving high-end MLCC demand at roughly 30% compound growth on Murata's forecast; regulatory substitution away from microplastics and lead; ASEAN PVC growth; global ageing supporting hygiene and diagnostics; MLCC price increases of 15-35% flowing up the chain and improving customers' willingness to accept material price revisions.

Headwinds: Chinese capacity in commodity inorganics setting the price in barium, strontium, zinc and titanium dioxide; customer vertical integration in dielectrics; organic UV absorbers taking share on cost; Japanese drug price controls; energy and freight cost volatility, with Middle East disruption around the Strait of Hormuz already forcing raw material cost pass-through in Q1 FYE March 2027.

7. Growth Triggers

Sourced only from the six management briefings listed in Section 9, plus company disclosures made at those briefings.

• AI-server dielectric demand is inflecting, with specific product lines multiplying. Management identified individual part numbers whose annual sales roughly doubled in dielectric materials and roughly tripled in dielectrics, and attributed them to AI-server applications (FY March 2026 full-year briefing, 25 May 2026). Repeated at the Q1 FYE March 2027 briefing, where management said such lines had doubled year on year and momentum was continuing at a moderated rate (7 August 2026).

  "We are not given full disclosure of the application for each product, but for those we can tell are for AI servers, the growth is clearly abnormal. For dielectric material part numbers, annual sales have roughly doubled; for dielectrics, some have tripled."

  • FY March 2026 full-year briefing, 25 May 2026

• Electronic materials capacity expansion is now under concrete evaluation. Utilisation is approaching maximum and management said equipment investment and capacity expansion are being examined in specific terms (Q1 FYE March 2027 briefing, 7 August 2026). This is a change of stance: at the December 2025 strategy briefing the position was that no major facility replacement was anticipated and that investment would depend on utilisation approaching full.

  "As for capital investment and capacity expansion, we are already examining it in quite concrete terms."

  • Q1 FYE March 2027 briefing, 7 August 2026

• Price revisions implemented across all segments. Cost pass-through in response to raw-material increases stemming from Middle East tension around the Strait of Hormuz, plus continuing price-correction requests on dielectrics and dielectric materials (Q1 FYE March 2027 briefing, 7 August 2026; dielectric pricing repeated from the 25 May 2026 briefing).

• The cosmetics makeup-materials multi-product plant completed in February 2026, with revenue contribution deliberately pushed into the next medium-term plan period (Q3 FYE March 2026 briefing, 12 February 2026).

• European cosmetics major adoption, with more in negotiation. First shipment in the January-March 2025 window; at that point less than 10% of segment revenue with a three-year ramp planned (FY March 2025 briefing, 27 May 2025). Management had earlier said several additional major manufacturers were in final-stage evaluation with third-party adoption expected by FYE March 2027 (Cosmetic Materials strategy briefing, 2 December 2024). Note that by February 2026 management said adoption had happened but was growing below plan because brand rollout was delayed.

• Cosmetics three-part recovery plan: expanded overseas selling into Southeast Asia and China; a strengthened volume-zone product lineup to compete on price against organic UV absorbers; and acceler ated development and production optimisation for faster launch of differentiated products (Q3 FYE March 2026 briefing, 12 February 2026).

• Pigment-grade titanium dioxide exit completed to schedule, production ending December 2025 and sales ending March 2026, removing a structurally loss-prone commodity and freeing the industrial waste disposal site for sale (guided at the FY March 2025 briefing, 27 May 2025; repeated at the H1 FYE March 2026 briefing, 1 December 2025; confirmed complete at the FY March 2026 briefing, 25 May 2026).

• Catalyst site consolidation complete, with the business promoted from efficiency-review to stable. Management said FYE March 2027 should sustain roughly current levels or better, while flagging that the Q3 result contained roughly ¥500m of one-off operating profit that would not repeat (Q3 FYE March 2026 briefing, 12 February 2026).

• Pharmaceutical CDMO capacity build. Roughly half of capital expenditure directed to growth investment, with pharmaceutical CDMO facilities and cosmetic makeup-material multi-line production named as the two targets (H1 FYE March 2025 briefing, 2 December 2024). Subsidiary Katayama Seiyakusho's on-site continuous-manufacturing project was subsequently selected for a large NEDO economic-security technology programme on 25 July 2025 (Katayama release).

• Smart Material programme: five or more certified products by 2030 with a ¥2bn sales target and a 50% gross margin, of which three had been commercialised as of the FY March 2026 briefing, named as Lumate, Multhiol and the SZR series (FY March 2026 briefing, 25 May 2026; targets originally set in the mid-term plan).

• ASEAN and India expansion in plastic additives, transferring non-lead stabiliser technology into the Vietnam and Thailand subsidiaries as domestic PVC shrinks (H1 FYE March 2026 briefing, 1 December 2025; Hanoi sales office opened May 2025, Sakai Shoji India representative office October 2025).

• ¥7.0bn share buyback with cancellation on completion. Authorised 30 June 2026 for up to 1,520,000 shares (9.9% of shares issued excluding treasury), running 1 July to 31 December 2026, explicitly to offset the convertible bond conversion and improve per-share value and capital efficiency (Q1 FYE March 2027 briefing, 7 August 2026).

• Segment reporting will be consolidated in the next medium-term plan, reducing the current eleven segments (FY March 2025 briefing, 27 May 2025).

Trigger summary

8. Key Risks

8.1 Customer vertical integration in dielectrics (moderate probability, high severity)

Mechanism. MF Material Co., Ltd. is majority-owned by Fuji Titanium, a competing barium titanate producer, with Murata and Ishihara Sangyo as minority partners, and plans to bring a new production facility online by 2027. Murata holds roughly 45% of AI-server MLCC share and roughly 40% of the global MLCC market. If Murata sources incremental barium titanate from a venture it part-owns, Sakai does not lose its existing qualified positions immediately, because those are locked into specific process recipes. It loses the incremental AI-driven volume, which is precisely the volume the entire growth story rests on. Worse, the merchant market's price discipline weakens when the largest buyer has an internal alternative.

Calibration. The 2027 facility timing collides exactly with the year Sakai says it is now evaluating its own capacity expansion. This is the single risk most likely to make the current AI narrative fail to convert into sustained profit, and it has not been addressed publicly by Sakai in any of the six briefings.

8.2 The hydrothermal advantage may be narrowing (moderate probability, high severity)

Mechanism. Sakai's entire premium positioning in dielectrics is the claim that hydrothermal synthesis produces the finest, most uniform particle and is therefore required as MLCC layers thin. If MLCC makers can achieve the required specification with solid-phase or oxalate powders, Sakai's premium collapses toward commodity powder economics.

Management said the quiet part out loud when asked directly:

"Differentiation by manufacturing method has diminished."

• FY March 2025 full-year briefing, 27 May 2025

At the December 2025 strategy briefing they then argued the opposite direction, pushing back on the industry perception that oxalate powders perform better in high-temperature applications and asserting actual adoption of Sakai's material in AI-server and automotive uses. Both statements can be true (differentiation narrowed in mid-range while remaining in high-end), but an investor should treat the durability of this moat as unresolved rather than established.

8.3 Cosmetic materials structurally losing to organic UV absorbers (high probability, moderate severity)

Mechanism. Inorganic UV blockers cost more than organic absorbers. When consumer spending tightens, brands reformulate toward cheaper organic chemistry, and Sakai's premium transparency argument stops mattering. Management's own explanation of the impairment named exactly this:

"In our mainstay sunscreen materials, earnings deteriorated due to the effects of the slowdown in the Chinese economy."

• Q3 FYE March 2026 briefing materials, 12 February 2026

with additional weak US demand from inflation and increased adoption of cheaper organic UV absorbers.

Calibration. This has already happened. It produced a ¥2.4bn impairment in Q3 FYE March 2026 plus roughly ¥0.3bn more in Q4 on the newly completed multi-plant. Asked what tangible fixed assets remained in cosmetics after the write-down, management answered that it was fine to understand almost nothing remained. The forward risk is not another impairment (there is little left to impair); it is that the segment stays loss-making and never becomes the second growth engine the plan requires. FYE March 2027 guidance has it still in operating loss.

8.4 The growth-business shortfall is structural, not timing (high probability, moderate-to-high severity)

Mechanism. The medium-term plan required growth businesses to deliver roughly ¥6.5bn of operating profit and to reach 60% or more of group operating profit by FYE March 2027. Current planning has them at roughly ¥3.0bn, a ¥3.5bn gap, while the efficiency-review businesses over-delivered (roughly ¥4.1bn against a ¥1.9bn baseline). The company's stated reason:

"Profit expansion in the growth businesses is taking more time than we assumed."

• FY March 2026 full-year briefing, 25 May 2026

Calibration. The concerning reading is that Sakai executed the restructuring half of its transformation well and the growth half poorly, which means the earnings improvement delivered so far came from subtraction (exiting pigment titanium dioxide, consolidating catalyst sites, withdrawing from lead stabilisers) and that source is now exhausted. Future improvement has to come from the growth businesses, which is the half with the weaker track record.

8.5 Plant safety at the Fukushima cluster (low probability, high severity)

Mechanism. Sakai handles zinc dust, flammable gases and fine powders, which is a combustible-dust and gas-explosion environment. It has a history. The Yumoto plant zinc powder facility suffered a dust-explosion fire, the investigation attributed the cause to zinc dust explosion, and the company subsequently withdrew from the zinc powder business because rebuilding and recovering the investment was judged unviable (Nikkei; company investigation committee report). More recently, on 8 June 2025 at around 08:30, a fire broke out at flammable-gas production equipment in the catalyst plant at the Daiken (Otsurugi) works in Iwaki City, Fukushima; it was extinguished by nitrogen purging of the equipment interior, one employee suffered a burn to the left wrist and was treated, and the company reported no property damage (first report, 9 June 2025, second report 4 July 2025).

Calibration. Individual incidents have been contained. The precedent that matters is that one of them ended a business line permanently. Three of the group's principal plants sit in Iwaki, so a serious event there has concentrated consequences.

8.6 Pharmaceutical compliance at Kaigen Pharma (low-to-moderate probability of recurrence, high severity)

Mechanism. On 22 December 2023 Osaka Prefecture and Hokkaido imposed a 39-day suspension of Kaigen Pharma's marketing-authorisation business and a 36-day manufacturing suspension at the Okuzawa plant, plus improvement orders, after eight products were found to have been manufactured differently from their approval dossiers, with the oldest deviations traced to 1977. Trade coverage attributed it to low GMP-compliance awareness and to problems specific to a company formed by merger. A repeat would suspend the group's ~50% share of Japanese gastrointestinal contrast media.

Calibration. The company's own strategy documents list restoring quality culture as the segment's first task, which is the right acknowledgement. The medical segment is currently around breakeven, so the earnings at risk are modest; the reputational and licence risk is not.

8.7 Dilution from remaining IA Partners warrants (near-certain if in the money, low severity)

Mechanism. The ¥3.0bn fourth convertible bond issued to an IA Partners fund in June 2023 at a ¥1,975 exercise price was fully converted in July 2026, issuing 1,519,000 shares (9.5% of prior shares outstanding) and taking shares issued to 17,518,987. The fund also holds warrants equivalent to 1,012,000 shares, exercisable until December 2027 (Nikkei, July 2026). With the shares well above the CB strike, exercise is the rational default.

Calibration. Low severity because the company has responded with a ¥7.0bn buyback-and-cancel, but investors should model roughly a million more shares arriving before the end of 2027 and ask whether they will be offset again.

8.8 Concentration of the AI narrative in a duopoly customer base (moderate probability, moderate severity)

Roughly 80-85% of AI-grade MLCCs come from Murata and Samsung Electro-Mechanics. Sakai's AI upside therefore depends on remaining qualified at two accounts, one of which is a partial owner of a competing powder producer. Any qualification loss or share shift between those two changes Sakai's outcome materially.

8.9 Commodity price-taking in half the portfolio (high probability, low-to-moderate severity)

Barium, strontium, commodity zinc oxide, plastic additives and contract processing are all price-taking businesses where Chinese capacity sets the level. Sakai's Q1 FYE March 2027 price revisions were explicitly cost pass-through rather than margin expansion, which is the honest description of pricing power in these lines.

8.10 Disclosure opacity (certain, low severity, but it raises the cost of underwriting)

Eleven reporting segments, no named customers, no disclosed customer concentration, and no publicly quantified capacity. An analyst put this to management directly in May 2025, saying the segment count had made things harder rather than easier to understand; the reply was that it would be reorganised in the next plan. Until then, every judgement about mix and margin has to be assembled from an eleven-row table.

9. Walk the Talk

The six briefings used, oldest to newest:

1. H1 FYE March 2025 results briefing, 2 December 2024 (52 attendees; President Toshiyuki Yagura and Executive Officer Shinji Ogama), accompanied the same day by the Cosmetic Materials business strategy briefing (Director and Executive Officer Koichiro Magara)
2. FYE March 2025 full-year results briefing, 27 May 2025 (Yagura and Ogama)
3. H1 FYE March 2026 results briefing, 1 December 2025 (67 attendees, hybrid, 51 minutes; Yagura and Ogama), accompanied the same day by the Electronic Materials business strategy briefing (Executive Officer Wataru Ibaraki and Functional Materials Division sales manager Hiroo Yoshikawa)
4. Q3 FYE March 2026 results briefing, 12 February 2026 (50 attendees, 22 minutes; Ogama)
5. FYE March 2026 full-year results briefing, 25 May 2026 (Ogama), preceded by a small meeting on 13 May 2026 (Ogama)
6. Q1 FYE March 2027 results briefing, 7 August 2026 (46 attendees, webcast; Ogama)

The most recent is 27 days before this report date, comfortably inside 90 days. Note the cadence itself changed: through FYE March 2025 the company held briefings only at the half and full year; from FYE March 2026 it added Q3, and from FYE March 2027 it added Q1. That is a genuine improvement in investor access and it happened during the IA Partners engagement.

A titles note, per the company's board page: Toshiyuki Yagura is President and Representative Director; Yasuhiro Okamoto is Managing Director and Executive Officer; Koichiro Magara and Shinji Ogama are Directors and Executive Officers; Wataru Ibaraki is an Executive Officer. The company publishes names and current titles only. No prior employers, education, tenure or career history is disclosed for any of them in the board listing or the 2025 Integrated Report, so none is asserted here.

The narrative

Start at December 2024. The company was one quarter past the halfway mark of a three-year plan whose targets were ¥9.0bn of operating profit and 8% ROE by FYE March 2027, and it was in the middle of the restructuring half. Electronic materials had just surged on semiconductor recovery, with dielectric material sales volume recovering substantially. Cosmetics guidance was being cut on Chinese weakness. Titanium dioxide's pigment exit was confirmed for the year ending March 2026, with management framing the choice as pursuing profitability over volume rather than competing on price against Chinese product. Notably, despite a first-half beat, full-year guidance was left unchanged, with management citing an uncertain environment. That is the first data point on temperament: this is a conservative-guidance management team.

The same day, the cosmetics team laid out the most specific, most falsifiable set of promises in the entire six-briefing sequence. Segment sales were roughly ¥2.5-3.0bn; the target was two to three times that by 2030; and, crucially, operating profit of roughly ¥1.5bn by the end of the medium-term plan. The path was a first European major already shipping from January-March 2025, several more in final-stage negotiation, and third-party adoption expected by FYE March 2027, plus a makeup-materials line to capture microplastic-bead substitution.

Move to May 2025. The pigment exit was reconfirmed with dates attached:

"Production of pigment-grade titanium dioxide will end in December 2025, with sales ending March 2026."

• FY March 2025 full-year briefing, 27 May 2025

The FYE March 2026 dividend was set at ¥130 (¥65 interim, ¥65 final) on a 3% DOE benchmark, a ¥2.5bn buyback was authorised with cancellation of the repurchased shares, and ROE was described as on track for the 8% target having moved from -9.2% to 6.6%. Capital expenditure was set at roughly 40% growth and 60% maintenance.

Two things were already wobbling. Asked how large the European cosmetics customer had become, the answer was less than 10% of a segment forecast of roughly ¥2.7-2.8bn, with meaningful expansion pushed to the third year. And asked about the hydrothermal route's adoption, management conceded that differentiation by manufacturing method had diminished, blaming the heavy automotive weighting of its dielectrics mix.

December 2025. The best briefing of the six. Full-year operating profit guidance was raised at the half, driven by dielectric material volume, with management quantifying growth business volumes when asked: electronic materials up 5-10% in H1, cosmetics volume down in line with sales, optical lens materials down 5-10%. The titanium dioxide exit was on schedule and management quantified the cost, saying the Titanium Dioxide and Zinc segment earned roughly ¥1.27bn of operating profit in the half with more than half of it from the pigment grade being exited. Cash conversion cycle had improved 18 days against baseline. Asked whether the efficiency-review work was essentially complete, President Yagura confirmed most of it was and that resources would now redirect to growth businesses.

The Electronic Materials strategy briefing the same day is where a specific forward statement was made that would age badly within eight months. Asked about equipment investment, management said no major manufacturing facility replacement was anticipated and that future capital investment was contingent on capacity utilisation approaching full.

February 2026. The reckoning on cosmetics. A ¥2.4bn impairment on cosmetics manufacturing equipment, with a further ¥0.3bn expected in Q4 upon completion of the multi-plant in February. Asked what tangible fixed assets would remain in the segment afterwards, the answer was that it was fine to understand almost nothing would. Fourteen months after guiding ¥1.5bn of segment operating profit by the end of the plan, the assets that were supposed to produce it had been written down to close to zero, and revenue contribution from the new plant was pushed into the next medium-term plan.

To management's credit, the same briefing raised the dividend from ¥130 to ¥145 and confirmed the ¥2.5bn buyback complete at 914,000 shares with 1,000,000 shares cancelled. They also volunteered, unprompted in substance, that the catalyst segment's strong quarter contained roughly ¥500m of one-off operating profit that would reverse in Q4. Flagging your own non-recurring gains is a good sign.

May 2026. The plan was formally missed and admitted. The FYE March 2027 operating profit target of ¥9.0bn was replaced with guidance of ¥6.0bn. The stated reason:

"Profit expansion in the growth businesses is taking more time than we assumed."

• FY March 2026 full-year briefing, 25 May 2026

The arithmetic of the miss is instructive: growth businesses planned at roughly ¥6.5bn came in around ¥3.0bn, a ¥3.5bn shortfall, while efficiency-review businesses roughly doubled from a ¥1.9bn baseline to roughly ¥4.1bn. Cash conversion cycle went from 216 days at plan start to 178 days, beating the 180-day target. Shareholder returns reached roughly ¥9.4bn against the ¥8bn three-year commitment, 117% of plan. ROE fell to 3.5% on the impairment while management maintained the 8% ambition for the final year.

There was also a telling exchange on electronic materials. An analyst pointed out that FYE March 2027 revenue was guided roughly flat despite the generative-AI story and asked how profit could rise so much on so little revenue growth. Management's answer was that orders were robust and that the plan should be understood as conservative, then disclosed the doubling and tripling of AI-attributed part numbers. So the company was simultaneously reporting a demand inflection and guiding as if it might not persist. On the small-meeting call twelve days earlier, the FYE March 2027 buyback was left as "undecided," with management explaining that this preserved discretion based on profit performance and cash position.

August 2026. Two things reversed. The buyback that had been "undecided" in May became a ¥7.0bn programme for up to 1,520,000 shares authorised on 30 June, with 1,330,800 shares already repurchased for ¥5,336,436,500 by 31 August, and cancellation planned on completion (TDnet, 1 September 2026). And the capacity position articulated in December 2025 flipped: utilisation is approaching maximum and expansion is now being examined in concrete terms. Full-year guidance was left at the 13 May level pending observation of Q2 AI-server demand.

Promise versus outcome

Assessment

This management team does what it says on things it controls directly, and consistently over-promises on things that depend on customers adopting new products.

The restructuring record is genuinely good. The pigment titanium dioxide exit was a hard decision, announced with dates a year ahead, and delivered on the announced dates, including the awkward downstream consequence of disposing of a waste site. Working capital beat its target early. Capital returns beat their three-year commitment by 17%, the dividend has more than doubled in three years, buybacks have been executed and the shares cancelled rather than parked in treasury, and when the convertible bond diluted holders the company came back with a larger buyback within days rather than shrugging. Management also volunteers unflattering detail: the ¥500m catalyst one-off, the admission that method-based differentiation has diminished, the plain statement that almost no tangible assets remain in cosmetics, and the acknowledgement that the segment structure is confusing. That is not a promotional team.

The growth record is poor, and cosmetics is the clearest case in the report of a specific, dated, quantified promise being falsified. In December 2024 a director stood up at a dedicated strategy briefing and put ¥1.5bn of operating profit on the board by the end of the plan, backed by a named European win and several more in final negotiation. Fourteen months later the division was impaired to near zero and the revenue was moved to the following plan period. The failure was not entirely their fault, Chinese consumer weakness and price-driven substitution to organic UV absorbers are real external forces, but the promise was made with knowledge of a market that was already softening, and the recovery plan announced in February 2026 (sell more in Southeast Asia, add cheaper volume-zone grades, develop faster) reads as a restatement of intent rather than a changed strategy.

The most instructive single data point is the December 2025 statement that no significant capacity investment was anticipated, followed eight months later by capacity near maximum and expansion under concrete evaluation. That is not dishonesty. It is a company that was so conditioned by a decade of cyclical disappointment in electronic materials that it under-estimated the one upcycle that finally arrived, and is now potentially arriving late to capacity in a market where a customer-backed competitor has a facility scheduled for 2027. The conservatism that makes their guidance trustworthy is the same conservatism that may cost them the AI window.

Verdict: credible operators and reliable capital allocators, unreliable forecasters of their own growth businesses. Discount the growth-segment targets; take the restructuring and capital-return commitments at face value.

10. Shareholder Friendliness Index

Dividends. Dividend per share was ¥70 for the year ended March 2024 (¥35 interim, ¥35 final), ¥135 for the year ended March 2025 (¥62.5 interim, ¥72.5 final), and ¥145 for the year ended March 2026 (¥65 interim, ¥80 final), with ¥160 forecast for the year ending March 2027 (irbank dividend history; Integrated Report 2025; Q3 and full-year FYE March 2026 briefings). That is a near-doubling in two years and a further raise on top, with no cuts. Two features are worth naming. First, the policy changed: the company abandoned a payout-ratio approach in favour of a 3% dividend-on-equity benchmark, which is why the dividend kept rising through a year in which earnings fell on impairment. Second, the FYE March 2026 dividend was raised mid-year, from ¥130 to ¥145 at the Q3 briefing on 12 February 2026, in the same announcement that disclosed a ¥2.4bn impairment. Payout ratio consequently reached roughly 82% in the year ended March 2026 against roughly 44% the year before, which tells you the dividend is now anchored to the balance sheet rather than to earnings and will not flex down with a weak year.

Buybacks and dilution. Two programmes in the last three years, and I could not verify any repurchase in the year ended March 2024. In the year ended March 2026 the company authorised ¥2.5bn for up to 1,000,000 shares on 14 May 2025, completed it by 2 December 2025 having acquired 914,000 shares for ¥2.5bn, and cancelled 1,000,000 shares on 23 December 2025 (company news releases; Q3 FYE March 2026 briefing). In the current year, following full conversion of the ¥3.0bn fourth convertible bond held by an IA Partners fund (exercise price ¥1,975, exercised July 2026, issuing 1,519,000 shares and taking shares issued to 17,518,987), the company authorised on 30 June 2026 a much larger programme of up to ¥7.0bn for up to 1,520,000 shares, or 9.9% of shares issued excluding treasury, running 1 July to 31 December 2026, with cancellation planned on completion. Progress in the trailing ~90 days per MoatMap's disclosure feed and the company's own monthly filings: 596,100 shares for ¥2,397,132,000 at an average of JPY 4,021 in July 2026, and 734,700 shares for ¥2,939,304,500 at an average of JPY 4,001 in August 2026, giving cumulative 1,330,800 shares for ¥5,336,436,500 as of 31 August 2026 (TDnet, 1 September 2026). On share count: issued shares were roughly 17.0m before the December 2025 cancellation, roughly 16.0m after it, 17,518,987 after the July 2026 conversion, and will return toward roughly 16.0m once the current programme is cancelled. So over three years the count is roughly flat, with a genuine 1m-share retirement offset by 1.5m shares of convertible dilution and then re-offset by a larger buyback. One remaining overhang: the IA Partners fund holds warrants equivalent to 1,012,000 shares exercisable to December 2027. Total shareholder returns across the three-year plan reached roughly ¥9.4bn against a stated target of more than ¥8bn.

Verdict: Returns Capital. The dividend has more than doubled in three years under an explicit DOE floor, and when the convertible bond diluted holders by 9.5% the company responded within days with a ¥7.0bn buyback-for-cancellation rather than letting the dilution stand.

11. Insider Activities

Japan's disclosure portals (EDINET for large shareholder reports, TDnet for timely disclosure) are API-gated and return blocked or empty stubs to general web search. The transaction data below is therefore taken from MoatMap's cross-market disclosure database, which scrapes the EDINET 5%-rule filings directly, and is the sole source for recent insider dealing in this venue. Data current as of 3 September 2026.

Recent transactions, last 12 months

Both entries are EDINET large-shareholder reports (大量保有報告書 / 変更報告書, dated as shown). Japan's 5%-rule filings disclose the change in holding and the purpose of holding but no execution price, which is why the value column is blank. The share count and direction are the signal.

Buys: read the signal

There were none. Zero purchases by any director, officer or substantial shareholder appear in the trailing twelve months. There is therefore no conviction signal to read, positive or negative, from insider buying. It is worth stating plainly what is absent: through a period in which the shares roughly doubled, the AI-server demand inflection became visible in the company's own product-level data, and management launched the largest buyback in its recent history, not one director or executive officer bought stock on the open market. That is not bearish in itself, since Japanese executives at a company of this size hold relatively small personal positions and typically accumulate through restricted stock rather than market purchases (the company disposed of treasury shares for restricted stock units in July 2025 and again in July 2026), but it means the strongest available bullish signal is simply not present.

Sells: working out the why

Both sells are explained by their filers' own stated purpose of holding, and neither is a view on the business.

Nomura Securities, 91,130 shares (0.56%). The filing states the holding exists for securities-business trading inventory and for operating cumulative investment (accumulation-plan) services. This is a broker-dealer's market-making and client-facilitation book. Positions in such a book move mechanically with client flow and hedging, not with a house view. Reason effectively disclosed: inventory management.

Sumitomo Mitsui Trust Asset Management, 164,800 shares (1.02%). The filing states the holding is for management under investment trust and discretionary investment mandates. This is a passive and quasi-passive asset manager. A holding change of this size around a rebalancing window is routine index and mandate flow. Reason effectively disclosed: fund management activity, not a discretionary sell decision about Sakai's prospects.

There is a plausible technical driver worth flagging, though it is inference rather than disclosure. Both filings fall in early-to-mid June 2026, immediately ahead of the 30 June 2026 authorisation of the ¥7.0bn buyback and the July 2026 conversion of the convertible bond that added 1,519,000 shares, or 9.5% of shares outstanding, to the register. A 9.5% increase in shares outstanding mechanically dilutes every holder's percentage and, for a 5%-rule filer sitting near the threshold, forces a change report. Whether these two filers were trimming into an anticipated share-count increase or simply crossing a reporting threshold because the denominator changed is not disclosed and should not be guessed at. What can be said is that the June 2026 window was one in which the register was being reshaped by corporate action, and both filers are institutions whose reporting obligations are triggered by exactly that.

Net assessment

Insiders are net sellers on a headline reading, but the label is misleading. Neither seller is an insider in the sense that matters for signalling: there are no director or officer transactions at all in the window, and the two reported sells come from a broker-dealer's trading inventory and an index-and-mandate asset manager, each disclosing a purpose of holding that is inconsistent with an investment view. Combined, the two sells total roughly 256,000 shares, about 1.6% of shares outstanding, against a company that itself bought back 1,330,800 shares (over five times as many) in the two months that followed.

The far more informative capital-markets event in the period is not on this table at all. It is the exit of IA Partners from an advisory role. The fund invested ¥3.0bn through a convertible bond in June 2023, supported the portfolio transformation and financial strategy for three years, saw the advisory partnership formally concluded in May 2026, converted the entire bond in July 2026, and thereby became the largest shareholder excluding trust accounts. It still holds warrants for a further 1,012,000 shares to December 2027. An activist-adjacent investor converting rather than selling, and retaining warrants for eighteen more months, is a constructive signal about intent, even as it is a dilution overhang.

Read: neutral. Nothing in the insider record supports a bullish thesis, and nothing undermines one. The two disclosed sells are mechanical institutional flow, not information. The absence of any director or officer open-market buying during a period of visibly improving fundamentals is the one mildly disappointing feature, and the company's own aggressive buyback-for-cancellation is the offsetting fact.

12. Scenarios

Bull case

The AI-server MLCC shortage turns out to be structural rather than a spike, and Sakai finds itself holding a genuinely scarce asset: the largest merchant hydrothermal barium titanate capacity in the world at a moment when the finest, most uniform powder is the binding constraint on layer thinning. Murata's own forecast of roughly 30% compound growth in AI-server MLCC shipments to 2030 plays out. The part numbers management identified as doubling and tripling keep multiplying, and because dielectric powder is a tiny fraction of a capacitor's cost and an absolute prerequisite for it, Sakai converts the price revisions it has been requesting every year into real margin rather than cost pass-through. The capacity expansion now under concrete evaluation is announced and commissioned quickly enough to catch the wave, and Sakai's argument that hydrothermal powder performs in high-temperature AI and automotive applications is validated in production rather than merely asserted, so the mid-range share loss to solid-phase and oxalate routes stops mattering.

Around that core, the second and third growth legs finally contribute. The written-down cosmetics division becomes an accidental asset: with its fixed assets impaired to near zero, incremental sales drop almost straight through, and the ECHA microplastic phase-out that begins biting from 2027 pulls Barimaru and Calmaru into formulations that previously used plastic beads. The European major that has been slow to roll out finishes its brand-wide adoption, the several other majors that were in final negotiation in late 2024 convert, and the makeup multi-plant completed in February 2026 runs at a level the previous asset base never could. Organic chemicals adds a CDMO layer on top of its sole-source thiol positions, with the Katayama continuous-manufacturing programme selected for NEDO funding turning into a differentiated on-site manufacturing capability that Japanese and US pharma clients pay a premium for.

Meanwhile the boring segments quietly get better. LPZINC finds a home in the thermal interface materials that every AI rack needs, giving Sakai a second, entirely separate exposure to data-centre build-out from a zinc franchise it has run since 1918. Hygienic materials compound at their demographic rate through the Indonesian joint venture. Plastic additives complete their migration from a shrinking Japanese market into ASEAN. Catalysts contribute something real from methanation and Ir/ENETIA electrolysis. The next medium-term plan collapses eleven segments into four or five, the story becomes legible to generalist investors for the first time, and ROE finally reaches the 8% the company has been promising, then heads toward the 12% it has pencilled in for 2030. The capital-returns machine keeps running: DOE-anchored dividends rising, the IA Partners warrants absorbed by another buyback-for-cancellation, share count trending down rather than flat.

Base case

The most likely path is that Sakai remains a competent, unglamorous specialty inorganics company that has fixed its worst problem and is still looking for its next engine.

Electronic materials carries the group. AI-server demand stays strong but Sakai captures a smaller share of it than the headline industry numbers imply, because it sells two steps upstream into negotiated annual contracts rather than a spot market, and because the incremental volume at its largest customer partly routes to MF Material once that facility comes online. Sakai does expand capacity, probably announcing it within the next year, and the expansion is sized cautiously for the same reason the company has always been cautious: a decade of cyclical whipsaw in this segment. Prices get revised annually, mostly recovering input costs rather than expanding margin. The segment grows and its margin stays the best in the group, but it does not re-rate the company on its own.

Cosmetics does not become the second engine. It stops losing money at some point in the next plan period, helped by having almost no depreciation left, and it grows respectably off a small base as European adoption widens and microplastic rules tighten, but the ¥1.5bn of operating profit promised for the end of the current plan arrives late or not at all. Organic chemicals continues as it has: steady eyeglass lens chemistry, lumpy pharmaceutical intermediates, a CDMO transition that takes longer than the slide deck suggests. The stable businesses do what stable businesses do, with catalysts giving back the one-off gains management already flagged and hygiene and contract processing grinding forward.

Management delivers roughly what it now guides rather than what it originally planned, which is the pattern of the last three years. The ¥9.0bn operating profit target for the year ending March 2027 is not met; the guided figure roughly is. ROE recovers from the impairment year but reaching 8% depends on the buyback shrinking the equity base as much as on earnings. The next medium-term plan is published with fewer segments, more modest growth-business targets and a continued DOE-anchored return policy, and management's credibility improves simply by promising less. The IA Partners warrants are exercised and largely offset. Investors get a company whose earnings quality is much better than three years ago, whose capital returns are dependable, and whose growth story is still a work in progress.

Bear case

The bear case is not a demand collapse. It is that Sakai wins the argument and loses the business.

MF Material's new facility comes online in 2027 as planned. Murata, holding roughly 45% of AI-server MLCC share and an equity stake in the venture, directs its incremental barium titanate requirement there. Sakai keeps its existing qualified positions, because process lock-in is real, but it does not get the growth, and the growth was the entire thesis. Simultaneously the technical argument goes against it: MLCC makers find they can hit the specifications they need with solid-phase or oxalate powders they can buy more cheaply, exactly as management themselves conceded when they said differentiation by manufacturing method had diminished. Shandong Sinocera, already China's most advanced producer and investing heavily in higher-purity grades, closes enough of the quality gap to take the mid-range while holding its cost advantage in volume. Sakai's 25% share of the merchant market erodes from both ends, and the hydrothermal capacity advantage it spent decades building becomes an advantage in a route that fewer customers specify. If it has by then committed capital to the capacity expansion currently under evaluation, it owns new assets in a market that has stopped growing for it.

At the same time the restructuring dividend runs out. The profit improvement of the last three years came from subtraction: exiting pigment titanium dioxide, consolidating catalyst sites, withdrawing from lead stabilisers, squeezing 38 days out of the cash conversion cycle. Every one of those is now done, and management has said so. Pigment titanium dioxide took more than half of a ¥1.27bn half-year profit stream out of the group permanently. Catalysts gives back its one-offs. There is nothing left to cut, and the growth businesses that were supposed to take over are running roughly ¥3.5bn behind the plan they were set.

Cosmetics never recovers. Organic UV absorbers keep winning on price, Chinese consumer demand stays weak, the European brand rollout that was already delayed stalls, and a division whose assets have already been written down to near nothing continues to consume management attention and working capital while contributing losses. The Smart Material programme, targeting ¥2bn of sales by 2030, remains a rounding error.

Then something breaks. The plant safety record is the specific worry: a zinc dust explosion at Yumoto already ended one business line permanently because rebuilding could not be justified, a fire hit the Daiken catalyst plant in June 2025, and three of the group's principal plants sit in one Fukushima prefecture cluster. Or the governance risk recurs: Kaigen Pharma was suspended for 39 days in December 2023 for manufacturing eight products differently from their approval dossiers, with deviations traced to 1977, and a repeat would suspend a roughly 50% share of Japanese gastrointestinal contrast media. Either event lands on a company whose dividend is now anchored to a 3% DOE floor rather than to earnings, meaning the payout does not automatically flex down when profit does, and whose share count is about to grow again by roughly a million shares if the IA Partners warrants are exercised. In that world an investor owns eleven mediocre businesses, a well-intentioned management team that has already missed the targets it set, and a dividend commitment that outlives the earnings that funded it.

Sources

Company filings and briefings: FY2027 Q1 results and briefing transcript, 7 August 2026; FY2026 full-year briefing, 25 May 2026; small meeting, 13 May 2026; Q3 FY2026 briefing, 12 February 2026; H1 FY2026 briefing and Electronic Materials strategy briefing, 1 December 2025; FY2025 full-year briefing, 27 May 2025; H1 FY2025 briefing and Cosmetic Materials strategy briefing, 2 December 2024; Integrated Report 2025; Transformation: BEYOND2030 mid-term plan; company chronology; group companies; board members; product pages for electronic materials, cosmetic materials, barium and strontium, plastic additives, catalysts, organic chemicals, zinc oxide and LPZINC; news releases; buyback status, TDnet 1 September 2026; Daiken plant fire first report, 9 June 2025; Yumoto explosion investigation report.

Insider and buyback data: MoatMap cross-market disclosure database (EDINET 5%-rule filings, market JP), current 3 September 2026.

Industry and third-party: passive-components.eu, MLCCs in the Age of AI, Q2 2026; TrendForce, 29 July 2026; Dataintelo MLCC dielectric powders market; Murata MF Material establishment release; passive-components.eu on the Murata JV; Ishihara Sangyo titanium dioxide business; irbank dividend history; Nikkei on the IA Partners CB conversion and buyback; Nikkei on the Yumoto zinc dust explosion cause; PHARM TECH JAPAN on the Kaigen Pharma sanction; Nikkei Medical on the 39-day suspension; TipRanks on the waste site disposal; Katayama Seiyakusho NEDO selection, 25 July 2025; PT. S&S Hygiene Solution; Worldfolio interview with President Yagura, 14 November 2022; Japanese Wikipedia company history; kabutabi 4078 vs 4092 comparison; competitor market caps from stockanalysis.com and public quote pages as dated in the Section 5 table.