1Summary
What this section is. The whole filing in a few paragraphs. Every figure here appears again below with the document it came from.
Sanki Engineering is a Japanese building-services contractor: it designs and installs the air conditioning, ventilation, plumbing, electrical and control systems that go inside buildings and factories, rather than building the structures themselves. Roughly five-sixths of group revenue comes from that Building Systems business. Three smaller segments — materials-handling equipment, water and waste treatment plant, and a small rental property book — make up the rest. The company celebrated its centenary in the year under review and is one year into a three-year plan running to March 2028.
The year ended 31 March 2026 was a margin year, not a volume year. Revenue rose 0.6% to ¥254,674m, while operating profit rose 27.9% to ¥27,991m and net income attributable to owners rose 37.7% to ¥23,688m. Gross margin, on the company's own disclosure, went from 18.8% to 22.0%. Orders taken grew 11.2% and the order book carried into the new year grew 19.0%, so the revenue line and the order line moved in opposite directions. The first quarter of the current year, reported on 7 August 2026, showed revenue up 7.4% and operating profit up 106.5% against the same quarter a year earlier.
Management attributes the profit improvement to work done at the point of bidding and during construction rather than to a change in market conditions, and states that every profit line was a record. They also flag what did not go well: operating cash flow more than halved as receivables built up with the higher workload, the Machinery Systems segment made a wider loss, and Environmental Systems orders fell by a third. Separately, the company acknowledges that its own estimate of its cost of equity has risen from 7–8% to 8–9% as Japanese interest rates have moved up.
2Business
What this section is. What the company actually sells and where the money comes from. Segment shares are the company's own reported splits, not our estimates.
The group runs four reported segments plus a small "other" bucket, with eight consolidated subsidiaries as of 31 March 2026.
Building Systems is the core. It covers building HVAC and sanitary work, industrial air conditioning (mainly for factories), electrical installation, and facility systems. This is where the semiconductor fabs, data centres, research facilities and large urban redevelopment projects sit. Overseas arms operate in Thailand, Shanghai and — from May 2026 — Malaysia, where Sanki bought 40% of ES Matrix, an electrical and telecoms installer serving semiconductor and data-centre customers.
Machinery Systems manufactures and sells conveyor and materials-handling systems. Environmental Systems builds water, sewage and waste-treatment facilities, largely for public-sector clients, and includes the Austrian subsidiary AQUACONSULT Anlagenbau. Real Estate rents and manages the group's own property.
How the money is earned matters here. This is percentage-of-completion contracting: revenue is recognised over the life of a project as performance obligations are satisfied, which makes the estimated margin assumed on each contract a critical accounting judgement — the company says so explicitly in its own critical-estimates disclosure. Revenue in any given year is therefore mostly a function of the order book taken in prior years and how fast it is worked off, not of orders won in that year.
On competitive position, the company describes Building Systems demand as firm, driven by large metropolitan redevelopment, data centres, semiconductor plants and research facilities, and says it carried more work into the new year than the year before. It describes Machinery Systems as flat with a difficult order environment, specifically because investment in battery-electric vehicles has become less certain and it had been targeting logistics equipment for secondary-battery plants. It describes Environmental Systems public investment as running at about the same level as the prior year. On the parent-company (non-consolidated) basis the company discloses that 70.4% of building-equipment orders were negotiated rather than competitively tendered, up from 66.8% — relevant context for the margin move, though the company does not draw that link itself.
Cross-shareholdings. These are notable. At 31 March 2026 Sanki held 21 listed stocks with a balance-sheet value of ¥31,890m, plus 49 unlisted holdings at ¥1,546m. The company states this equals 27.7% of consolidated net assets and that it targets below 20% by March 2028, having committed to halving both the number of names and the value from the March 2024 level. It sold ¥5,193m of listed shares during the year. The largest single position is Toyota Motor at ¥13,325m, followed by Imperial Hotel (¥3,377m), Ono Pharmaceutical (¥1,966m) and Canon (¥1,965m). Every holding is justified as a customer relationship in the building-equipment business. Note that the mid-term ROE and EPS targets are explicitly measured excluding gains on the sale of these shares, so the unwind flatters reported earnings without flattering the numbers management is judged on.
3Results
How to read the numbers. Japanese companies report quarterly figures cumulatively from the start of the fiscal year — Q1 covers three months, Q2 six, Q3 nine. A quarter is never annualised here. “Company forecast” is the company's own published full-year number, which it revises itself; it is not an analyst estimate.
Full year to 31 March 2026 (consolidated)
Two conventions worth explaining. Ordinary profit (経常利益) is a Japanese GAAP line that sits 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. For Sanki it runs about ¥1.3bn above operating profit, consistent with dividend income from the cross-shareholding portfolio described above. EPS has been restated for a 1-for-3 stock split effective 1 May 2026, so per-share figures here are not comparable with pre-May 2026 press coverage.
What management said
The figures below come from the company's own MD&A table in the same securities report (docID S100YH3A) rather than from the machine-verified headline extract, and are presented as the company's disclosure.
Management's account of the year is short and specific. Orders received rose 11.2% to ¥294,738m, which they attribute to winning large urban-redevelopment projects. Revenue rose only 0.6% "because small and mid-sized projects progressed on schedule" — the large redevelopment wins are, by implication, not yet converting to revenue. The order backlog carried into the current year rose 19.0% to ¥250,794m. Profit rose because of "efforts to improve profitability at the point of order and during construction." Gross profit rose 18.1% to ¥56,070m, taking gross margin from 18.8% to 22.0%; that 3.2-point move accounts for essentially the whole profit increase. Management states that every stage of profit set a record.
By segment:
Building Systems carried the group. Management credits the profit rise to "margin improvement" without decomposing it further. For Machinery Systems and Environmental Systems, management gives the same reason for both the order fall and the revenue fall: a base effect from large projects booked or delivered in the prior year. The widening Machinery Systems loss is not separately explained beyond that base effect. Real Estate revenue rose on higher tenant income but profit fell because of refurbishment work on some rental properties.
Cash flow. Operating cash flow fell from ¥29,725m to ¥13,167m. Management's explanation is that pre-tax profit exceeded the increase in trade receivables and tax payments — that is, the number stayed positive, but the working-capital build that comes with a rising workload plus a larger tax bill consumed most of it. Investing cash flow was −¥1,338m, mainly purchases of securities. Financing cash flow was −¥16,068m, on dividends and share buybacks, up from −¥11,398m.
Against the mid-term plan. Sanki reports ROE of 18.6% excluding gains on policy-shareholding sales, against a plan floor of 16.0%; EPS of ¥137 on the same exclusion and post-split share count, against a March 2028 target of ¥144 or more. Dividend-on-equity was 8.8% against a 5.0% floor. Growth investment of ¥6.6bn was made against a cumulative three-year plan of roughly ¥50bn, and 2.87m shares were repurchased against a cumulative plan of roughly 12m — 24% of the buyback plan in year one. The company also notes its market capitalisation passed ¥300bn, roughly doubling in a year, and that its PBR reached about 3x, while its own assessed cost of equity rose from 7–8% to 8–9%.
R&D spend was ¥1,978m, reported group-wide because the central R&D function works across segments.
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 will become six-month figures in November. They are not annualised.
Company forecast for the full year to March 2027: revenue ¥265,000m, operating profit ¥31,000m, ordinary profit ¥31,500m, net income ¥26,300m. Japanese listed companies publish their own full-year forecast under Tokyo Stock Exchange practice; it is management's own number, not an analyst consensus, and revising it is itself a disclosable event. Sanki disclosed a revision to its full-year consolidated forecast and its dividend forecast on 31 July 2026, ahead of the Q1 report; the figures above are from the 7 August Q1 report and therefore reflect that revision. (The 31 July release itself was not retrievable — see Sources.)
Read against that forecast, Q1 revenue is about 20% of the full-year number and Q1 operating profit about 15%. The company's own risk disclosure notes that project completions and construction progress concentrate towards the fiscal year end, so quarterly progression is not linear.
4Valuation context
What this section is, and is not. Multiples calculated from reported figures and the company's own forecast, so you can see where the shares sit. No target price is derived from them and none is implied.
At ¥2,422 per share (verified 13 August 2026), with 160,983,468 shares outstanding and a market capitalisation of ¥389.9bn (Nikkei):
- Trailing PER: ¥2,422 ÷ ¥153.51 reported EPS for the year to March 2026 = 15.8x
- Forward PER on company guidance: the ¥26,300m net income forecast against the share count implied by Q1 EPS (¥4,376m ÷ ¥28.65 ≈ 152.7m shares, i.e. ex-treasury) gives forecast EPS of about ¥172, or 14.1x. Nikkei quotes 14.0x.
- PBR: ¥389.9bn ÷ ¥121,437m net assets at 31 March 2026 = 3.2x. Nikkei quotes 3.12x, using a more recent book value — Q1 equity is higher, with the equity ratio at 60.4% versus 55.3% at year-end.
- Dividend yield: the forecast dividend for the year to March 2027 is ¥65 per share post-split, giving 2.68% (kabuyoho). The year to March 2026 also paid ¥65 post-split (¥195 before the 1-for-3 split effective 1 May 2026).
Peers, all from Nikkei company pages on 13 August 2026, on the same forecast-PER and actual-PBR basis:
Takasago Thermal is the largest listed pure HVAC contractor in Japan and the closest comparator. Taikisha combines HVAC with automotive paint-finishing plant, which is why its forecast ROE and PBR sit apart from the group. Dai-Dan is a building-services contractor of similar type concentrated in the Tokyo area. The four trade within about 1.4 turns of each other on forward PER; the PBR spread is wider and tracks the ROE spread.
5What to watch
How to use this. Specific things you can check yourself in the next filing. They are questions to carry forward, not predictions about what will happen.
KPIs, with the level that would matter
- Order backlog at the interim results. ¥250,794m at 31 March 2026, up 19.0%. Because revenue is worked off the backlog, this is the forward revenue line. The gap between +11.2% order growth and +0.6% revenue growth last year is the thing the backlog has to resolve. A backlog that stops growing while revenue accelerates is conversion; one that shrinks while revenue is flat is not.
- Gross margin. 22.0% for the year to March 2026, from 18.8%. The entire profit increase came from here, and management attributes it to bidding and site execution rather than to market conditions — which makes it something the company claims it controls. Whether the interim margin holds at or above 22% is the direct test of that claim.
- Operating cash flow. ¥13,167m, down from ¥29,725m, on receivables build and tax. Contracting margins can be reported before they are collected. Half-year operating cash flow that continues to run well below the profit line, with receivables rising faster than revenue, would matter more than the earnings figure itself.
- Machinery Systems. A ¥918m segment loss, widened from ¥614m, on orders down 23.9%. Management has restructured the unit (a 1 April 2026 reorganisation creating teams for airport and semiconductor work) and points to a large Tokyo-area airport logistics reorganisation as the opportunity, while flagging BEV investment uncertainty as the drag. Return to segment breakeven, or a further widening, is the readable outcome.
- Cross-shareholding ratio. 27.7% of consolidated net assets against a sub-20% target by March 2028. Progress here is both a capital-efficiency signal and a source of non-operating gains — remembering that the ROE and EPS targets exclude those gains, so disposals help the balance sheet and the reported bottom line but not the measured target.
Scheduled catalysts
- 30 September 2026 — record date for the interim dividend.
- 13 November 2026 — first-half (six-month cumulative) results; analyst and institutional-investor briefing on 16 November.
- 12 February 2027 — third-quarter cumulative results.
- Full-year results to March 2027, expected around mid-May 2027; not yet on the published IR calendar. This will close out year two of Mid-term Plan 2027, whose March 2028 targets are revenue ¥300bn, operating profit ¥30bn, a 10.0% operating margin, EPS of ¥144 or more, ROE of 16% or more, and cumulative buybacks of roughly 12m shares against 2.87m completed.
- Guidance revisions are unscheduled but recurrent: the company revised full-year guidance twice in the prior fiscal year and again on 31 July 2026.
6Risks the company discloses
Whose risks these are. Taken from the company's own statutory risk disclosure. This is what management chose to flag, not our assessment of it.
Sanki ranks its risks by impact and likelihood. The ones it places first are about people and inputs, not demand.
Labour. Two separate items: failure to recruit or retain its own engineers, and a shrinking pool of engineers at its subcontractors. It also lists Japan's overtime cap for the construction industry as a distinct risk — total available man-hours fall, which caps the volume of work the company can physically execute regardless of how much it wins. Its stated responses are pay increases, subcontractor training and relationship programmes, and automation of installation work.
Input costs and procurement. Materials and labour cost inflation that cannot be passed into an already-signed contract price. The company separately flags escalating Middle East tensions as a risk to oil supply, material prices and delivery lead times, and US tariff policy as a risk to both customer capex plans and material prices. Its mitigations are contractual — negotiating price-revision clauses before order, and early ordering after.
Loss-making contracts. Design changes mid-project, schedule compression, extra labour and rework generating unbudgeted cost. The company adds a specific concern about large projects: as jobs get bigger and longer, schedule management gets more complex and cost overruns, delays and quality problems become more likely. Given that percentage-of-completion accounting depends on the estimated margin for each contract, this risk and the accounting-estimate risk are the same risk.
Project scale and staffing (Building Systems). Semiconductor, automotive and data-centre projects arriving together with metropolitan redevelopment, such that site staff and subcontractor capacity cannot be allocated. Longer project durations then increase exposure to inflation and to customers changing their investment plans.
Segment-specific. Machinery Systems: intense competition and price wars. Environmental Systems: deteriorating municipal finances driving price competition, and DBO contracts — where the private operator designs, builds and runs the facility for many years — carrying long-dated inflation exposure. Real Estate: ageing buildings, occupancy and rent levels.
Financial and group-wide. Falls in the market value of the cross-shareholding portfolio and of pension assets; interest-rate movements affecting both pension funding and borrowing costs; customer credit; regulatory breach under construction, antitrust or labour law; litigation; cyber-attack and leakage of customer drawings; being late to adopt generative AI, and misuse of it; pandemic; war, terrorism and natural disaster. It also discloses seasonality as a risk in its own right: completions cluster at the fiscal year end, so funding requirements swing materially through the year.
7Sources
Why this is here. Every document behind the figures above, so any number on this page can be traced back and checked.
- EDINET annual securities report, docID S100YH3A, filed 2026-06-23 (year ended 31 March 2026). All consolidated headline figures, segment figures, risk and cross-shareholding disclosure.
- Q1 results release (決算短信) for the year ending March 2027, disclosed 2026-08-07.
- Share price ¥2,422, shares outstanding, market cap, PBR and peer multiples: Nikkei company pages, retrieved 2026-08-13 — 1961, 1969, 1979, 1980.
- Dividend forecast and history: kabuyoho and Sanki Engineering shareholder returns page.
- Results calendar: Sanki Engineering IR calendar.
- The 31 July 2026 TDnet releases on full-year and dividend forecast revision (1, 2) returned HTTP 403 and could not be read; their existence and titles are taken from search listings, and no figures from them are stated here.
---
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.