Pilot's Largest Customer Is Pilot

The Isesaki plant in Gunma makes ballpoints and mechanical pencils that retail for ¥1,000 (US$6.11) or less, and it ships ink and components to the group's factories abroad. The Hiratsuka plant in Kanagawa, running since 1948, makes fountain pens at the top of the price range, the lacquer-art pens sold under the Namiki name, jewellery, and ceramic industrial materials. Elsewhere the company explains that its pencil-lead extrusion technique yields micro-porous ceramics for car components and semiconductor equipment; it does not say at which plant. One issuer owns both lists. The distance between them is the distance between a factory opened in 1948 and one opened in 1978.

The turn toward lacquer began, by the company's own account, as a materials complaint. Ebonite barrels held up against ink but discoloured and lost their gloss, so a development team went looking for something better and arrived at urushi, yielding the technique the company calls Lacquanite — dated 1924 on the Namiki brand's history page and 1925 on Pilot's centenary site, the same firm disagreeing with itself about when. Once a pen is being coated in lacquer, someone is going to observe that maki-e is applied to lacquer. By 1925 the maki-e fountain pen existed, made under the technical direction of Rokkaku Shisui and his pupil Matsuda Gonroku, later designated a living national treasure.

The pen in the American drawer is, by Pilot's own account, the G-2, which the company says it launched in 1997 and makes at Isesaki. The US arm calls the G-2 the best-selling gel pen in America and the Precise V5 the best-selling roller, without naming a research firm, a period, a channel, or whether the ranking is by units or by dollars. The page is undated as well. The parent's report on the year to December 2025, disclosed 13 February 2026, is drier: G-2 volumes held up, Mexican demand was weak, the yen moved, and the Americas segment's revenue fell while its profit rose on lower costs and advertising.

The consolidated result for the year ended December 2025: revenue ¥126,391 million (US$771.6 million), up 0.2 percent — though the year also brought the Indian subsidiary PPIN into consolidation from the start of the first quarter, so the comparison is not like for like. Operating profit fell 6.5 percent to ¥16,649 million (US$101.6 million), which the company attributes mainly to higher labour and depreciation costs in Japan. Below the operating line the gap widens again: foreign exchange swung from a gain to a loss, the prior year's step-acquisition gain did not recur, special losses roughly doubled on environmental costs, and the tax charge rose. Net profit attributable to owners was ¥12,064 million (US$73.65 million), down 20.5 percent.

Against its own plan that was a miss, and the company sizes it. The 2025 revenue target was ¥133,000 million (US$812.0 million); the result came in ¥6,609 million (US$40.35 million) under. The operating margin target of 13.5 percent produced 13.2. The ROE target of 10.0 percent produced 8.5. The one target beaten was the total payout ratio, where the floor was 50 percent and the year came in at 87.3. Everything measuring what the company earns landed short; the thing measuring what it hands back landed over.

Currency does not close the gap. The plan assumed ¥150 to the dollar, ¥160 to the euro and ¥21.4 to the yuan; the year printed 149.61, 169.19 and 20.81, which are planning rates rather than amounts and stand unconverted here. Against the euro the yen came in more than nine yen weaker than assumed — 169.19 against 160 — which flatters every European sale on the way home. None of this is netted out anywhere. Across twenty-two pages of plan review there is no bridge chart, no volume-price-mix table, and no yen figure attached to any driver: the three causes the company names are labour costs, soft overseas consumer demand, and the cost of running capital expenditure high, each of them a phrase without a number. The same twenty-two pages report that US channel stock of the G-2 was cut by more than 20 percent in units against the end of 2024, meeting its target; that sits in the section on supply chain and not in the section on why sales missed. From this document the shortfall cannot be separated into volume, price and mix. The securities report and the results-briefing material may carry the split; they are not examined here.

The reporting segments are geographic rather than product-based: Japan, the Americas, Europe, Asia. Japan booked external sales of ¥37,456 million (US$228.7 million) and internal sales of ¥46,460 million (US$283.6 million). The domestic segment sells more to its siblings than to outsiders. It also carries ¥11,815 million (US$72.13 million) of the ¥16,535 million (US$100.9 million) in segment profit, as reported, before the intersegment elimination the company is careful to flag. The subsidiaries in the Americas, Europe and Asia are described as manufacturing as well as selling; what stays exclusively Japanese is toys, jewellery and industrial materials. In the accounts, Japan is a segment whose largest customer is the rest of the company.

Divide that ¥11,815 million (US$72.13 million) of Japanese profit by the ¥37,456 million (US$228.7 million) of external sales and the answer is 31.5 percent, which is the figure Pilot prints against Japan in its own April 2026 summary. What the summary does not print beside that percentage is the internal sales the same segment booked, so the second denominator cannot be built from that page. It can be built for earlier years, because the segment note in the securities report carries both series. On external sales alone the Japan margin computes to 64.5 percent for 2022 — ¥22,591 million (US$137.9 million) of profit over ¥35,051 million (US$214.0 million) — and to 34.5 percent for 2023, on ¥13,482 million (US$82.31 million) over ¥39,062 million (US$238.5 million). On the base that includes internal sales, ¥86,796 million (US$529.9 million) and ¥81,685 million (US$498.7 million), the same numerators give 26.0 and 16.5. Same numerator, two denominators, both of them the company's own, and only the larger result is the one that gets published with a percent sign attached. The internal sales are struck at third-party transaction prices, the company says, and the profit on them is taken back out one line down.

That line does not sit still. Segment profits are struck before intersegment elimination, whose composition the filing does not disclose — unrealised profit held in inventory is the standard example, and the note itself names nothing. The adjustment was minus ¥4,720 million (US$28.82 million) in 2022 and plus ¥1,842 million (US$11.25 million) in 2023 in the securities report's segment note; the April 2026 summary shows plus ¥114 million (US$0.6960 million) against 2025 in an adjustment column it does not define, which is a different document and not demonstrably the same construction. So between 2022 and 2023 the reported segment profits fell ¥8,804 million (US$53.75 million), from ¥25,965 million (US$158.5 million), while consolidated operating profit fell ¥2,241 million (US$13.68 million), from ¥21,244 million (US$129.7 million). That is 33.9 percent against 10.5 percent: one series says a third of the profit went, the other says a tenth. The filing calls the difference intersegment elimination and stops there.

The first quarter of 2026 put the mechanism in a single line. Revenue was ¥31,526 million (US$192.5 million), up 8.3 percent; the four regional segments earned ¥5,183 million (US$31.64 million) between them, up 40.0 percent; consolidated operating profit was ¥4,154 million (US$25.36 million), down 10.4 percent. The company gives the reason in one sentence: the adjustment for unrealised profit in inventory went from plus ¥931 million (US$5.684 million) to minus ¥1,028 million (US$6.276 million), a swing of ¥1,959 million (US$11.96 million). Take the second from the segment total and you land a million yen above the reported figure, which is rounding. Every region sold more and the group made less. External sales rose in all four: the Americas supplied ¥1,256 million (US$7.668 million) of the ¥2,408 million (US$14.70 million) increase and lifted profit 46.5 percent, Japan added ¥335 million (US$2.045 million) and lifted profit 53.8 percent, and the other two moved the other way at the profit line — Asia grew 6.8 percent and earned 31.8 percent less, Europe grew 6.9 percent in yen, declined in local currency by the company's own account, and turned a profit into a loss of ¥19 million (US$0.1160 million). Two of four regions earned more, and all four sold more. A quarter is a thin window here: roughly 60 percent of American and 40 percent of European sales ship to mass retailers ahead of the new school year.

How foreign the company is depends on which definition you accept. By customer location, overseas revenue was ¥97,037 million (US$592.4 million) against domestic ¥29,353 million (US$179.2 million), or 76.8 percent of the ¥126,391 million (US$771.6 million) total. By the entity booking the sale, the three non-Japan segments' external sales of ¥88,933 million (US$542.9 million) over the same total come to 70.4 percent. The company declines the arithmetic entirely and states in its risk disclosure that overseas is roughly three quarters of sales while most of the manufacturing is at home, which is the honest version of both numbers.

Product detail exists only inside Japan. There is no stationery segment and no writing-instruments segment; what gets disclosed is a breakdown of Japanese sales, in which stationery came to ¥29,896 million (US$182.5 million) and writing instruments ¥26,173 million (US$159.8 million), the latter at 91.9 percent of the prior year. Part of that fall is not a fall: sales to Malaysia and India left the Japan segment when those subsidiaries were consolidated into Asia, which the company also gives as the main reason Asia's sales rose, from ¥20,817 million (US$127.1 million) to ¥23,422 million (US$143.0 million). How much moved is not disclosed, and the company says the change in consolidation scope makes the two segments non-comparable year on year.

Two categories outside stationery are small against revenue and appear only as lines in a summary: toys at ¥4,009 million (US$24.48 million), and industrial materials and other at ¥3,561 million (US$21.74 million) in the April 2026 summary, which describes the industrial line as ceramics used in flow-control parts, sensor holders and semiconductor equipment. They are worth the space because the 2030 vision is written in their ratio. With stationery other than writing instruments at ¥5,676 million (US$34.65 million), the three sub-lines add to ¥13,246 million (US$80.87 million) against a stated non-writing total of ¥13,247 million (US$80.87 million), a one-million-yen rounding gap that at four significant figures does not survive translation into dollars at all. That summary is also where the whole non-writing side finally carries a number: ¥13,247 million, or 10.5 percent of revenue, against writing instruments at ¥113,143 million (US$690.7 million). Two years earlier the same split ran ¥12,317 million (US$75.20 million) and ¥106,273 million (US$648.8 million). Non-writing grew 7.6 percent over those two years and writing instruments 6.5 percent, so the mix moved a tenth of a point, from 10.4 to 10.5. The 2030 vision puts non-writing at 25 percent. The 14.5-point gap is drawn nowhere: the target and the ratio sit on different pages, and the subtraction is mine.

FriXion passed five billion units cumulatively, series-wide, as of 31 December 2025, and it is the only brand-level number the company publishes. What FriXion earns is not stated, because no brand's revenue is stated and no product's profit; the segment note knows only geography. The share figures come with the same texture: second place in Japan at about 22 percent, second in the United States at about 13 percent, second across eight European countries at about 13 percent, each of them the company's own survey, with the basis, the period and the eight countries unstated. Who holds first place is not named either.

The 2025-2027 plan reset its own numbers on 13 February 2026. The 2026 revenue target went from ¥139,000 million (US$848.6 million) to ¥133,000 million (US$812.0 million), the 2027 target from ¥145,000 million (US$885.2 million) to ¥139,000 million. The 2026 operating margin floor went from 15 percent to 13.5, the 2027 floor from 16 to 15. The 2027 ROE floor went from 11 percent to 10; the 2026 floor stayed at 10. Which leaves every year holding the target that used to belong to the year before it: 2027's new revenue line is 2026's old one, 2027's new margin floor is 2026's old floor, and 2026's revised pair is what 2025 was supposed to deliver. The company describes this as revising the time axis for achievement and says the growth it planned needs more time. It does not name the year it expects instead. Guidance for 2026 — revenue ¥133,000 million and operating profit ¥18,000 million (US$109.9 million), a margin of 13.5 percent on that division — no longer sits under the plan, because the plan came down to meet it.

The regional ambitions kept their numbers and changed their units. The same day's progress deck asks for India up 30 percent and ASEAN's ten markets up 15 percent by 2027, both measured against 2024, while the financial targets a few pages earlier are absolute yen for 2027 measured against nothing in particular. The 2025 actuals arrive in a third and fourth form: India up 11.4 percent, on an axis denominated in millions of rupees, and ASEAN up 1.3 percent in yen, where ASEAN means the external sales of three subsidiaries rather than ten markets. The bars carry no value labels, so the amounts cannot be read off them. A rupee series, a yen series, a percentage target against a base year the financial plan does not use, and a definition of ASEAN that changes between the results page and the target page: nothing here converts into anything else without a figure the deck does not print.

The money moved with the targets. Capital expenditure across the plan period was ¥45,000 million (US$274.7 million) as first published and is now ¥26,356 million (US$160.9 million): 2025's actual ¥7,856 million (US$47.96 million), then ¥10,000 million (US$61.05 million) for 2026 where ¥17,000 million (US$103.8 million) was planned, then ¥8,500 million (US$51.89 million) for 2027 where ¥20,500 million (US$125.2 million) was planned. A cut of 41 percent. The named reason is the rebuild of the Isesaki plant: it had been scheduled to start in 2026, and it is now smaller, because the company revised its demand assumptions, and runs three years from 2027, which is to say out of this plan and into whatever follows it. Systems spending of ¥3,000 million (US$18.32 million) is unchanged, and none of it was spent in 2025, against a plan of none. Depreciation for the same three years falls 2.4 percent, from ¥20,500 million (US$125.2 million) to ¥20,007 million (US$122.1 million), against the 41 percent taken out of capital expenditure; the deck sets the two figures side by side and connects them with nothing.

What did not get cut is the money going out. On the same day the shareholder-return target moved the other way, from a total payout ratio of 50 percent or more to 70 percent or more, inside a ¥40,000 million (US$244.2 million) return frame for the three years. The 2025 figure of 87.3 percent included ¥5,999 million (US$36.62 million) of buybacks. On 24 February 2026 the company took 2,119,000 shares, 5.68 percent of the shares outstanding excluding treasury, for ¥10,266 million (US$62.67 million), all of it off-auction on the Tokyo exchange's ToSTNeT-3 facility in an order the board resolution says was good for that trading window alone, against a frame of ¥10,600 million (US$64.71 million): 100.00 percent of the authorised share count and 96.85 percent of the authorised yen, both the company's own progress figures. Subtract one from the other and ¥334 million (US$2.04 million) of the yen frame expired unused, because the binding cap was the share count. On 31 March it cancelled 3,500,000 shares, 8.56 percent of shares outstanding including treasury, taking ¥16,186 million (US$98.82 million) out of retained earnings.

Which leaves the ROE line doing something quiet. Revenue targets came down, margin floors came down, the 2027 ROE floor came down, and the 2026 ROE floor did not: 10 percent or more, unchanged, with the lines beneath it lower and the payout target above it higher. Net assets fell ¥8,403 million (US$51.30 million) in the first quarter, to ¥137,675 million (US$840.5 million), and the equity ratio went from 80.8 percent to 79.2. The company puts no causal sentence between the two decisions. It also publishes no dividend payout ratio for 2025: the five-year table carries 16.6, 22.5, 28.9 and 30.1 percent for the four prior years and a dash for the fifth, with the total payout ratio standing in its place. Dividing the ¥120 annual dividend by reported earnings per share of ¥317.02 fills the blank at 37.9 percent, which is my arithmetic on two of the company's own figures and not a disclosed number. The dividend is ¥120 for 2025 against a planned ¥126 for 2026 before the one-for-three split, a raise of ¥6.

On conversion: every yen figure in this piece is translated at ¥163.8 = US$1, the rate quoted for 24 July 2026 by TradingEconomics rather than anything the company discloses, and applied uniformly to every period regardless of when the yen amount was reported. Because one rate is used throughout, each dollar figure is a fixed multiple of its yen figure and period-to-period comparisons come out identical in either currency; what is lost is the rate that actually applied, since by construction none of these dollar amounts uses the average or closing rate of the year it describes. Comparability is limited for a second reason the company supplies itself: PPIN Private Limited was newly consolidated from the start of the first quarter of 2025, which the filing flags as a material change in the scope of consolidation. Per-share amounts are not converted here; those the company reports are pre-split, the one-for-three split taking effect 1 July 2026. Planning and realised exchange rates, ratios, percentages and multiples are left unconverted.