1Summary

Taikisha is a Japanese engineering contractor with two businesses that have almost nothing to do with each other. The first, Environmental Systems, is the same building-services work its listed peers do: designing and installing air conditioning for offices, hotels, hospitals and data centres, and the cleanrooms and precision temperature control that semiconductor, electronic-component, battery, pharmaceutical and food plants require. The second, Paint Finishing Systems, designs and installs the paint shops inside car factories — the booths, conveyors and ovens that coat car bodies and bumpers — and extends the same work to construction vehicles, rolling stock, aircraft and general industrial equipment. The group runs 32 consolidated subsidiaries and 3 affiliates across a network it describes as 20 countries and 30 locations.

That second business is why Taikisha does not read like its peers. It makes the company an exporter of capital equipment to global carmakers rather than a domestic contractor, with more than half of orders now coming from outside Japan, and it is the reason Taikisha's returns and its valuation sit apart from the group.

The year ended 31 March 2026 was a strong profit year on a flat revenue line. Revenue rose 3.6% to ¥286,127m while operating profit rose 29.8% to ¥23,320m and net income attributable to owners rose 41.4% to ¥15,594m. Orders, which the company reports separately, rose 26.8% to ¥351,740m, with overseas orders up 32.4%. The gap between +26.8% orders and +3.6% revenue is the shape of the year: the work was won, and most of it has not yet been built.

The single most unusual figure is cash flow. Operating cash flow went from −¥21,219m to +¥64,700m — a swing of ¥85,919m against a revenue base of ¥286bn — and cash and equivalents roughly doubled from ¥42,013m to ¥86,359m. This is a working-capital reversal, not an earnings event, and the [Results](#results) section takes it apart.

The first quarter of the current year, disclosed on 7 August 2026, went the other way: revenue −4.2%, operating profit −11.0%, net income −28.4%.

Three listed peers do the building-services half of this work and filed the same fiscal year: Takasago Thermal Engineering (1969), the largest, Sanki Engineering (1961), and Shin Nippon Air Technologies (1952), the smallest. With this review the cluster covers every listed Japanese air-conditioning installation contractor. All four reported a margin-led improvement for the year to March 2026 and then diverged in the June quarter — Sanki's operating profit roughly doubled, Taikisha's fell 11.0%, Takasago's fell 16.2%, and Shin Nippon's halved. Reading the four together separates a sector condition from a company outcome, and Taikisha is the one that tests whether the sector story survives contact with a company that is only half in the sector.

2Business

Two reported segments, matching the two businesses described above, with no "other" bucket.

Environmental Systems splits into building HVAC (offices, hotels, shops, schools, laboratories, theatres, halls, hospitals, data centres) and industrial HVAC (cleanrooms for semiconductor, electronic-component, battery, precision-machinery, pharmaceutical and food plants, plus plant factories). It is the larger segment and the one directly comparable to the peers. In the year under review it took orders of ¥216,588m, up 20.9%, and recognised revenue of ¥183,176m, up 8.1%.

Paint Finishing Systems designs and installs paint plant, mainly for the automotive industry. Taikisha describes itself as the domestic leader in this work. It took orders of ¥135,151m, up 37.6%, of which ¥113,138m — 84% — was overseas, and recognised revenue of ¥103,088m, down 3.6%.

How the money is earned is the same percentage-of-completion contracting as at the peers: revenue is recognised over the life of a project, so revenue in any year is a function of the order book taken in prior years and how fast it is worked off. Taikisha's version has two features the peers do not have. Its projects are more international — overseas was 52.7% of orders (¥185,480m of ¥351,740m) — so translation of foreign subsidiaries moves the reported result; the company discloses that US dollar, Thai baht, Chinese renminbi, Indian rupee and Philippine peso movements matter most, and quantifies each. And its paint-shop customers are car manufacturers making capital-investment decisions about electrification, which is a different demand cycle from building construction.

The overseas subsidiary disclosure shows where the growth is. Taikisha Engineering India's revenue went from ¥13,042m to ¥20,418m and its ordinary profit from ¥913m to ¥2,503m; Taikisha USA moved from an ordinary loss of ¥96m to a profit of ¥976m. China went the other way, revenue ¥16,831m to ¥13,529m and profit ¥1,938m to ¥1,153m.

On strategy, the company published a ten-year plan (10-Year Plan 2035) on 15 May 2025 and is one year into the three-year mid-term plan inside it, which it labels "restructuring for transformation." The stated 2035 destination is revenue above ¥500bn — against ¥286bn today — with ROE of 12% or more and headcount of 7,200 against 5,525. During the year it acquired a North American automation company with an established customer base among the Detroit Three, to build out automation within Paint Finishing; it does not name the target in the securities report. It also created a regional strategy department with dedicated units for ASEAN, East Asia, India, North America and Japan.

Cross-shareholdings. Taikisha holds 20 listed stocks at a balance-sheet value of ¥31,783m and 20 unlisted at ¥421m. It discloses the ratio to net assets as 21.1%, up from 18.0% a year earlier, against a target of 15% or less by March 2028 and cumulative sales of ¥5.0bn across the three-year plan.

Two things about that are worth stating plainly, because both are visible in the filing and neither is drawn out by the company.

First, the ratio rose in a year when the company was selling. It disposed of two names in full — Canon, carried at ¥1,741m a year earlier, and East Japan Railway at ¥177m — for proceeds of ¥1,988m, roughly 40% of the three-year ¥5.0bn plan in year one. Yet the disclosed listed holdings went from about ¥25.2bn to ¥31.8bn. Selling on plan and still going backwards on the ratio is what a rising equity market does to a target expressed as a ratio of a mark-to-market number, and the company says so: it had already missed this target once for the same reason, hitting 20% only in the final year of the previous plan.

Second, the 21.1% includes holdings that are not on its own balance sheet. Listed plus unlisted is ¥32,204m, which against consolidated net assets of ¥170,232m is 18.9%. Adding the two deemed holdings contributed to the retirement benefit trust — Murata Manufacturing at ¥3,181m and Takeda Pharmaceutical at ¥566m, over which Taikisha retains voting instruction rights — gives ¥35,951m, or 21.1%. That reconciles exactly, and it means the company is disclosing on a wider definition than the balance-sheet figure. This is the reverse of the usual direction of travel and, on the evidence of the reconciliation, a conservative choice. It also means the number is not comparable with a peer ratio computed from listed holdings alone. Computed the way this site computes it for the other three — listed holdings over consolidated net assets — Taikisha is at 18.7%, the lowest of the four.

The largest positions are Sumitomo Realty & Development (¥8,424m), Murata Manufacturing (¥6,861m), Hulic (¥3,790m), Suzuki (¥2,979m) and Tokyo Tatemono (¥1,748m). Every one is justified as a customer of one of the two segments — the property companies buy building HVAC, Suzuki buys paint plant. The board reviews each name every April against, in the company's words, whether the benefit and risk of holding it are commensurate with the cost of capital.

3Results

Full year to 31 March 2026 (consolidated)

One convention worth explaining. Ordinary profit (経常利益) is a Japanese GAAP line between operating and pre-tax profit: it adds recurring non-operating items — interest, dividend income, FX, equity-method results — but excludes one-off extraordinary gains and losses. Japanese investors watch it as the "normal earning power" line. Taikisha also reports its segment profit on this basis rather than on operating profit, which is why the segment table below does not foot to operating profit.

What management said

The figures in this section come from the company's own MD&A in the same securities report (docID S100YCPQ) rather than from the machine-verified headline extract, and are presented as the company's disclosure.

Orders received rose 26.8% to ¥351,740m, of which overseas was ¥185,480m, up 32.4%. Completed works revenue rose 3.6% to ¥286,265m on the segment basis, of which overseas was ¥144,168m, up 9.4%. (The ¥286,265m includes inter-segment transactions; the consolidated revenue line is ¥286,127m.) Gross profit on completed works rose ¥10,253m to ¥55,259m — a gross margin of 19.3% against 16.3% the year before, on the company's own numerator and denominator. That three-point move, not volume, is the profit story, exactly as at the three peers.

By segment, on the company's basis including inter-segment transactions and with segment profit stated as ordinary profit:

Environmental Systems carried the profit increase almost entirely: ¥5,520m of the ¥5,629m combined segment improvement. Paint Finishing grew orders 37.6% and added ¥109m of profit. Management's explanation for the Paint Finishing revenue decline is a base effect — a large domestic project contributed in the prior year — offset by growth in India and Europe. Within Environmental Systems, building HVAC grew orders 43.4% while industrial HVAC grew 11.2%, which is a change of mix for a company whose industrial side is twice the size of its building side.

Two disclosures on the parent-company (non-consolidated) basis are worth carrying across. The negotiated order ratio — work won by direct negotiation rather than competitive tender — rose from 67.5% to 75.3%. And the parent's order backlog carried into the current year rose from ¥135,391m to ¥154,516m, with ¥84,023m of that in building HVAC against ¥52,895m a year earlier. Neither figure is given on a consolidated basis, so both understate the group, but the direction is the forward revenue line.

Named projects completed in the year included battery plant work for Obayashi at Shimoyama, a Murata production building in Izumo, a new Suzuki paint plant at Kosai, and a paint booth installation at Nissan's Oppama plant. Work in hand includes a Softbank data centre at Tomakomai due February 2027 and an Obayashi redevelopment in Sapporo due July 2027.

Cash flow. This is the year's outlier. Operating cash flow was +¥64,700m against −¥21,219m. Management's stated cause is a decrease in trade receivables and pre-tax profit, partly offset by a decrease in trade payables. The balance-sheet movement confirms it: receivables and amounts due on completed construction fell ¥39,020m while cash on deposit rose ¥43,108m, and advances received on uncompleted contracts rose ¥7,781m. Investing cash flow was +¥748m. Financing cash flow was −¥23,475m, on a net repayment of short-term borrowings (down ¥11,952m) and dividends.

The reading is that the prior year's negative operating cash flow was a receivables build on large projects, and this year is the collection of it. A contractor recognising revenue over the life of a project can run cash flow far away from profit in either direction for a year or two; ¥64.7bn of operating cash flow on ¥15.6bn of net income is as unrepresentative of steady state as −¥21.2bn was. The useful figure is the two years together: about ¥43.5bn of operating cash flow against ¥26.6bn of net income over the two years, which is an ordinary-looking relationship.

R&D was ¥1,600m, of which ¥836m in Environmental Systems, run out of two centres in Kanagawa.

Against the mid-term plan. The plan's own table puts several FY2026 outcomes ahead of the FY2028 targets. Orders of ¥351.7bn already exceed the ¥296.0bn planned for March 2028; ordinary profit of ¥24,790m exceeds the ¥22.7bn target; net income of ¥15,594m is level with the ¥15.8bn target; ROE of 10.1% is close to the 10.3% target. Revenue is the exception: ¥286.1bn against a ¥336.5bn target. A plan whose profit targets are met in year one, with revenue two years behind, is a plan built on converting a backlog that has since grown faster than expected. The company's own FY2027 forecast row shows orders falling back to ¥330.5bn.

First quarter to 30 June 2026 (disclosed 7 August 2026)

Japanese quarterly reporting is cumulative from the start of the fiscal year, so these are three-month figures that become six-month figures in November. They are not annualised.

Company forecast for the full year to March 2027: revenue ¥307,000m, operating profit ¥23,800m, ordinary profit ¥25,000m, net income ¥18,000m, dividend ¥119 per share. Japanese listed companies publish their own full-year forecast under Tokyo Stock Exchange practice; it is management's number, not an analyst consensus, and revising it is itself a disclosable event. No revision has been disclosed since the 7 August Q1 release.

Read against that forecast, Q1 revenue is 20.0% of the full-year number, operating profit 14.9%, ordinary profit 15.4% and net income 11.3%. Progress below a straight-line quarter is normal for this industry — completions cluster at the fiscal year end — but the net income line is running furthest behind, and the forecast requires net income to grow 15.4% for the year from a quarter that fell 28.4%.

One balance-sheet movement deserves a flag. Owners' equity rose from about ¥160.9bn at 31 March (implied by the 56.1% equity ratio on ¥286,820m of assets) to ¥174,521m at 30 June, and the equity ratio from 56.1% to 60.5%, while quarterly net income was ¥2,034m. Something other than retained earnings added roughly ¥11.6bn of equity in three months. The two candidates visible in the annual report are the valuation reserve on the securities portfolio and the foreign-currency translation adjustment, both of which moved materially in the prior year; the quarterly release does not break this out in the extracted data, so it is stated here as an observation rather than an attribution. It is also the reason the equity ratio and the PBR below should be read with the cross-shareholding section in mind.

4Valuation context

At ¥4,080 per share — the 14 August 2026 close, retrieved 16 August — with 63,464,018 shares outstanding and a market capitalisation of ¥258,933m (Nikkei):

• Trailing PER: ¥4,080 ÷ ¥245.14 reported EPS for the year to March 2026 = 16.6x
• Forward PER on company guidance: the ¥18,000m net income forecast against the share count implied by Q1 EPS (¥2,034m ÷ ¥32.29 ≈ 63.0m shares, i.e. ex-treasury) gives forecast EPS of about ¥286, or 14.3x. Nikkei quotes 14.2x.
• PBR: on owners' equity at 30 June 2026 of ¥174,521m, ¥258,933m ÷ ¥174,521m = 1.48x. Nikkei quotes 1.47x. On owners' equity at 31 March 2026 — about ¥160.9bn — it would be 1.61x. The gap between those two numbers is the quarterly equity move described above.
• Dividend yield: the forecast dividend for the year to March 2027 is ¥119 per share, giving 2.92% (kabuyoho). The year to March 2026 paid ¥110. The company's stated policy is a dividend on equity of 4.5%.

Peers, all from Nikkei company pages retrieved 16 August 2026 and reflecting the 14 August close, on the same forecast-PER and actual-PBR basis:

The five sit within about three turns of each other on forward PER. The PBR spread is the whole story: Taikisha at 1.47x against Sanki at 3.15x, on a forecast ROE of 11.2% against 21.7%. That is close to arithmetic — PBR tracks ROE across this table with Shin Nippon the only loose fit — and it means the question about Taikisha is not whether the market has mispriced it against its peers, but whether a company earning 11% on equity can get to the 12% it has told the market it wants by 2035. Its own mid-term target is 10.3% for March 2028, which is below where it already is.

The second half of that question is the balance sheet. An equity ratio of 60.5% and ¥86bn of cash against a ¥259bn market capitalisation is a low-geared balance sheet even by the standards of Japanese contracting, and a low-geared balance sheet is a mathematical drag on ROE. The mid-term plan's own capital allocation table commits ¥73.0bn over three years, of which ¥38.0bn is growth investment (including ¥22.0bn for M&A, split ¥7.0bn Japan, ¥7.0bn North America, ¥5.0bn India, ¥2.0bn Europe, ¥1.0bn ASEAN), ¥3.5bn infrastructure, ¥16.5bn dividends and ¥15.0bn buybacks. The funding side is ¥51.5bn of pre-investment operating cash flow, ¥5.0bn from selling policy shareholdings and ¥16.5bn from cash on hand or borrowing. Whether that is enough to move an equity ratio of 60.5% is the readable test.

Reviews of the peers, on the same filings basis as this one: Sanki Engineering (1961), Takasago Thermal Engineering (1969) and Shin Nippon Air Technologies (1952).

5What to watch

KPIs, with the level that would matter

1. Order backlog at the interim results. ¥154,516m on the parent basis at 31 March 2026, up from ¥135,391m, with building HVAC up from ¥52,895m to ¥84,023m. Orders grew 26.8% against revenue growth of 3.6%; that gap has to resolve into revenue or it was a timing artefact. The company's own FY2027 plan row assumes orders fall to ¥330.5bn, so a second year of order growth would be the surprise, not the base case.
2. Gross margin on completed works. 19.3% for the year, from 16.3%. The entire profit increase came from here. Q1 operating margin was 5.8% (¥3,547m on ¥61,399m) against 8.2% for the full prior year, which is consistent with seasonality but is the first datapoint against the margin claim. Whether the interim margin holds near 19% is the test.
3. Operating cash flow. +¥64,700m after −¥21,219m. The signal is not the level but whether it normalises: a third consecutive year of a large swing in either direction would say the receivables cycle on large overseas projects is the dominant financial fact about this company, not its margin.
4. Paint Finishing Systems profit. ¥4,365m of segment profit on ¥103,088m of revenue — a 4.2% margin against Environmental Systems' 11.4%. Orders grew 37.6% and 84% of them are overseas. This segment is the reason the group's ROE is half its peers'; it is also the one where a North American automation acquisition has just been made. Segment profit converging toward the Environmental Systems margin, or not, is the single most consequential line in the group.
5. Cross-shareholding ratio. 21.1% of net assets on the company's definition — 18.7% on listed holdings over consolidated net assets — against 15% or less by March 2028. ¥1,988m sold against a ¥5.0bn three-year plan. The ratio rose despite the selling, so watch the yen sold, not the ratio.
6. The buyback. Announced 7 August 2026: up to 1.5m shares, 2.37% of shares outstanding excluding treasury, capped at ¥5.0bn, running 10 August 2026 to 31 January 2027. That is the full annual ¥5.0bn of the plan's ¥15.0bn three-year buyback, launched in the first half.

Scheduled catalysts

• 10 November 2026 — first-half (six-month cumulative) results; analyst briefing on 13 November.
• 9 February 2027 — third-quarter cumulative results.
• 31 January 2027 — end of the announced buyback window.
• Full-year results to March 2027, expected around mid-May 2027; not yet on the published IR calendar. This closes year two of the three-year plan, whose March 2028 targets are revenue ¥336.5bn, ordinary profit ¥22.7bn, net income ¥15.8bn, ROE 10.3%, an equity ratio of 40% or more, DOE of 4.5% and a cross-shareholding ratio of 15% or less.
• Guidance revisions are unscheduled. None has been disclosed since the 7 August Q1 release.

6Risks the company discloses

Taikisha lists twelve risks. Unlike its peers, it puts demand first rather than labour.

Private capital investment. Named first, and named specifically per segment: for Environmental Systems, a fall in Japanese-company investment overseas; for Paint Finishing, continued shrinkage of domestic vehicle production and a global downturn in car sales. It adds a distinct risk that it could be late to adapt to how carmakers change production equipment for carbon neutrality, and lose customers as a result. Its stated response is to accelerate development for changing paint plant and to push automation into industries beyond four- and two-wheel vehicles.

Overseas operations and control of overseas subsidiaries. Unexpected legal and regulatory change, political instability, FX exposure on foreign-currency contracts both for receipts and for payments to suppliers, translation of subsidiary accounts, customer insolvency, and overseas subsidiaries missing their business plans. Given that more than half of orders are overseas, this is the risk that most distinguishes Taikisha from the domestic peers. It hedges contract FX with forwards and runs pre-order credit screening.

People. The construction overtime cap that took effect in April 2024 reducing total engineer hours, an ageing domestic workforce, and — separately — failure to develop or retain local core staff overseas. Responses are modularisation, front-loading of site work, digital productivity and a global talent portfolio system starting in ASEAN.

Serious accidents and quality defects. Contractual non-conformity after handover, for which it carries a warranty provision that could prove insufficient.

M&A. Named as its own risk: impairment if a post-acquisition result diverges from the acquisition case. An investment committee screens deals before the board votes, and PMI is monitored afterwards. This is a live risk rather than a generic one given the North American acquisition made during the year.

Materials and labour cost inflation that cannot be passed into a signed contract price; legal compliance under construction, antitrust and labour law; information leakage through cyber-attack or insider action; climate change, including carbon tax cost, lost competitiveness if decarbonisation technology lags, and — physically — lost site productivity and suspended work as average temperatures and extreme-heat days rise; human rights across the supply chain; and large-scale natural disaster, including the second-order effect on customers and the wider economy.

Notably, Taikisha does not disclose seasonality as a standalone risk, though its peers do and its own quarterly progression shows it.

7Sources

• EDINET annual securities report, docID S100YCPQ, filed 2026-06-17 (year ended 31 March 2026). All consolidated headline figures, segment and order figures, risk, R&D and cross-shareholding disclosure.
• Q1 results release (決算短信) for the year ending March 2027, disclosed 2026-08-07, via EDINET DB.
• Share price ¥4,080, shares outstanding, market cap, PBR and peer multiples: Nikkei company pages, retrieved 2026-08-16 — 1979, 1961, 1969, 1952, 1980.
• Dividend forecast and history: kabuyoho.
• Results calendar and IR topics: Taikisha IR calendar and IR topics.
• Buyback terms (1.5m shares / ¥5.0bn cap / 10 Aug 2026–31 Jan 2027) are from the 7 August 2026 disclosure as reported in market coverage; the company's own PDF was not retrieved directly.
• Comparisons with peers are drawn from their own securities reports — see the Sanki Engineering (1961), Takasago Thermal Engineering (1969) and Shin Nippon Air Technologies (1952) reviews.

---

This is a review of reported results. It contains no investment recommendation, no price target and no valuation model. Multiples are stated as arithmetic on disclosed figures and a stated share price.