Shibaura Mechatronics (6590)

Signal panel

Investment thesis

1. Orders have stepped up. Q1 orders ¥52.0bn (+190.5%) vs ¥17.9–28.1bn a quarter in FY2026/03; back-end ¥37.1bn (+366.5%), front-end ¥14.0bn (+61.1%); semiconductor equipment 96% of orders, Taiwan 70% (Q1 短信 p.2; Q1 presentation p.2–3, p.10–12). Reported fact.
2. The guide is covered by backlog. Backlog ¥48.3bn at 2026-03-31 → ¥80.3bn at 2026-06-30 (FY presentation p.4; Q1 presentation p.2). FY sales guide ¥99.0bn minus Q1 ¥20.0bn leaves ¥79.0bn. Reported fact / inference.
3. Guidance is habitually conservative. Actual vs initial OP: FY22 +26.3%, FY23 +62.8%, FY24 +60.1%, FY25 +41.4%, FY26 +45.4% (fact sheet). FY2027/03 guide includes about ¥3.9bn more growth costs (depreciation ¥2.9bn, R&D ¥0.3bn, staff ¥0.7bn) (FY presentation p.20, p.22). Reported fact / company guidance.
4. The Q1 miss was timing. Front-end sales fell 25.8% as customers revised investment plans and moved delivery dates into Q2 or later; the company says there is no full-year impact (Q1 短信 p.2). Company guidance.
5. Price fell more than the earnings case. −42.5% from ¥6,320; 16.9x consensus FY2027/03 EPS (~¥215), 13.6x FY2028/03 (~¥267); PER band mean 19.6x (z −1.26) (fact sheet). Reported fact / external estimate.

Business and earnings

Two segments make the profit (FY2026/03, fact sheet): ファインメカトロニクス, the front-end business (wafer cleaning and etch for logic, foundry, silicon wafers and masks, plus FPD array tools), sales ¥52.2bn at 15.5% margin; and メカトロニクスシステム, the back-end business (bonders for advanced packaging, FPD module lines, vacuum deposition), sales ¥32.3bn (+36.7%) at 24.2% margin (+68% profit). Vending and ticket machines and property rental are small. Semiconductor equipment was 86% of FY2026/03 sales; its global-niche-top products rose from ¥27.1bn to ¥54.5bn between FY2022/03 and FY2025/03 (FY presentation p.32). Service revenue ¥22.8bn.

Customers: TSMC was ¥34.5bn, 39.2% of FY2026/03 sales, up from 22.7% (有報 p.27). By destination, Taiwan 40.8%, China 29.5%, Japan 24.2% (fact sheet, Regions). The AI exposure is quantified: back-end advanced-packaging bonders drove Q1 orders of ¥37.1bn, and GNT products were ¥41.6bn of the ¥49.7bn of semiconductor orders (Q1 presentation p.11). A hybrid-bonding pre-treatment tool, HSM-1000, launched on 2026-09-24 (snapshot pages/report_disclose.html); it has no revenue yet.

Q1 FY2027/03 (Q1 短信 p.1–2; presentation p.3–5): sales ¥20.0bn (−7.0%), OP ¥2.85bn (−29.9%), margin 14.2%. Back-end profit ¥3.1bn (+28.5%); front-end lost ¥0.1bn (vs +¥1.7bn) on lower sales and higher SG&A from growth investment. Q4 FY2026/03 margin was also low (13.4%), so two quarters now reflect the higher cost base.

Capital intensity is rising: capex ¥11.0bn in FY2026/03 (3.3x depreciation) for a new building and evaluation tools; inventory +19% vs sales +8.8% (fact sheet flags).

Forecast test

• FY2027/03 guidance (2026-05-13, unchanged 2026-08-06): sales ¥99.0bn, OP ¥16.0bn (ROS 16.2%), RP ¥15.7bn, NP ¥11.9bn, EPS ¥181.29, DPS ¥64; FCF +¥0.5bn; ROE 20.2% (Q1 短信 p.1; FY presentation p.21). Company guidance.
• Consensus (IFIS): FY2027/03 OP ¥19.4bn (+21.0% vs guide, +1.9% over 3 months), NP ¥14.1bn; FY2028/03 OP ¥24.0bn (+6.5% over 3 months), NP ¥17.5bn. External estimate.
• Progress: Q1 OP 17.8% of guide vs 30.4% median (−12.6 pts); sales 20.2% vs 24.4% (−4.2 pts). The median of actuals is 20.6% for OP, so the shortfall against final outcomes is small; the gap against guidance reflects how conservative earlier guides were. Reported fact.
• Orders and backlog: FY2026/03 orders ¥87.7bn (+26%), backlog ¥48.3bn; Q1 orders ¥52.0bn, book-to-bill 2.6x, backlog ¥80.3bn (FY presentation p.4; Q1 presentation p.2). Order-driven business: this is the central evidence.
• One-offs: none material.
• Revision probability: High for H2. The first test is Q2 (2026-11-05): front-end deliveries delayed from Q1 and back-end shipments from the order surge. The company has raised its guide at H1 in most recent years (FY2026/03 on 2025-11-06 and 2026-02-05).

Financial strength and capital allocation

• Balance sheet: equity ratio 56.0% at Q1; net cash ¥12.7bn at FY-end (cash ¥21.3bn, debt ¥8.7bn), cash +¥3.4bn in Q1; advances received +¥2.3bn in Q1 (Q1 短信 p.3; fact sheet). Current ratio 1.88. Gates pass.
• Cash: OCF positive every year FY2021–FY2026 (¥4.6–8.3bn); OCF/NP 0.41–0.68 since FY2023 as receivables and inventory grow with sales; FCF −¥3.5bn in FY2026/03 on capex.
• Returns: ROE 31.9% / 24.5% / 24.0% / 21.7% FY23–FY26; ROA 11.4%; IFIS ROIC 17.9%. FY26 ROE ≈ net margin 12.7% × asset turnover 0.87 × leverage 1.81 (falling as equity builds). Targets: keep ROE ≥20% and reach ROS ≥20% within Vision 2033, sales ≥¥100bn within FY2026–FY2028 (FY presentation p.36). Peers (Yahoo! Finance, actual): SCREEN ROE 20.3%, Disco 25.2%, TOWA 7.0%.
• Allocation: payout ~35% (raised from 30% from FY2025/03); DPS ¥60 → ¥64 guided (post 1:5 split on 2026-03-01) (FY presentation p.9, p.23). Buybacks are small and rare (¥0.86bn in 2023). Cash goes to capacity and R&D.

IR and management credibility

• 2023-05: Vision 2033 with phase targets; 2023-09 first integrated report.
• 2026-05-13: Phase 1 review published against its targets (SPE sales ¥41.5bn → ¥75.8bn; equity ratio 40.7% → 55.1%; payout 30% → 35%), and the ¥100bn sales target was brought forward (FY presentation p.32–36).
• Quarterly decks disclose orders, backlog, segment, product-family and destination splits.
• 2026-09-28 corrected 有報: fixes only the female hiring ratio (14.7% → 12.9%) in the human-capital section (訂正有報 p.2). Not an accounting matter.
• No results briefing for Q1 (Q1 短信 p.1).

Valuation and scenarios

65.6m shares outstanding ex treasury. Add the ¥64 dividend (1.8%).

• Own bands: forward PER 15.0x vs mean 19.6x (z −1.26); PBR 4.45x vs mean 5.36x (z −1.28). IFIS theoretical prices ¥4,756 (PER) and ¥4,379 (PBR). EV/EBITDA 12.5x forecast.
• Peers (Yahoo! Finance, 2026-10-16, company-forecast PER, actual PBR): SCREEN 22.7x / 5.25x; TOWA 23.2x / 2.15x; Disco PBR 10.1x (no forecast PER). Shibaura is 20.1x on its own guide and 16.9x on consensus.
• What the price implies: ¥3,635 is 20x the guide EPS with no beat, or about 15x a 12-month EPS that consensus already expects.
• Why the market may be right: semiconductor-equipment stocks trade on the capex cycle, and AI-packaging orders could be a one-time capacity step by one customer (TSMC 39% of sales); front-end customers are already revising plans; China (29.5% of sales) faces export-control risk; the cost base is rising ahead of revenue (margins 13–14% for two quarters).

Bull case

• Q2 recovers the delayed front-end sales and ships the back-end backlog; FY OP is raised toward consensus ¥19.4bn or above at H1.
• Packaging orders broaden beyond one customer as CoWoS-type and hybrid-bonding capacity spreads (HSM-1000).
• ROS returns toward the 20% target as new capacity fills.

Bear case

• Front-end delays become cancellations; FY OP lands at the guide.
• TSMC pauses packaging capex after a step-up, as in past cycles.
• New US or Japanese export rules cut China sales.
• The AI-capex trade unwinds further; the stock de-rates to ~13x even if earnings grow.

Catalysts

What would change the rating

• Upgrade (higher confidence) if Q2 OP is ≥¥4.5bn with orders ≥¥30bn, or the FY guide is raised to ≥¥18bn.
• Downgrade to Hold if Q2 orders fall below ¥20bn and backlog shrinks, or front-end delays extend into H2 without recovery.
• Thesis invalidated if the FY OP guide is cut, or TSMC orders are cancelled.

Monitoring

Scorecard

Fact-sheet flags

Computed by factsheet.mjs. Each flag is a prompt to investigate, not a conclusion.

• ▲amberinventory-outpacing-sales inventory grew 19.0% in 2026/03 (11,095 to 13,204) against sales growth of 8.8%
• iinfonet-cash Net cash of 12,690 in 2026/03 (BS_NET_DEBT -12,690)
• ▲amberprogress-behind-op 2027/03 Q1 op progress 17.8% vs median 30.4% (-12.6 pts): behind seasonal norm
• ▲amberprogress-behind-rp 2027/03 Q1 rp progress 18.6% vs median 31.2% (-12.6 pts): behind seasonal norm
• ▲amberprogress-behind-np 2027/03 Q1 np progress 16.3% vs median 33.7% (-17.4 pts): behind seasonal norm
• ▲amberfcf-negative Free cash flow -3,505 in 2026/03 (OCF 4,631, ICF -8,136, capex 10,964)
• iinfocapex-heavy Capex 10,964 is 3.30x depreciation 3,326 in 2026/03
• iinfoconsensus-above-guidance 2027/03 consensus OP 19,367 is +21.0% vs company guidance
• iinfoper-cheap Forward PER 14.95 vs 1-year mean 19.56 (+1σ 23.23), z -1.26: cheap vs own history
• iinfofar-below-52w-high Price 3,635 is -42.5% from the 52-week high 6,320

Sources and limitations

• 2026-05-13 FY2026/03 決算説明資料: data/6590/2026-10-17/filings/2026-03_FY_説明会_20260513.pdf p.4, p.9, p.20–23, p.32–36.
• 2026-06-15 有価証券報告書: filings/2026-03_FY_有報_20260615.pdf p.27 (major customer).
• 2026-08-06 Q1 FY2027/03 決算短信: filings/2027-03_Q1_短信_20260806.pdf p.1–3.
• 2026-08-06 Q1 決算説明資料: filings/2027-03_Q1_説明会_20260806.pdf p.2–12.
• 2026-09-28 訂正有価証券報告書: filings/2026-03_FY_訂正有報_20260928.pdf p.2.
• IFIS snapshot 2026-10-17 and fact sheet data/6590/analysis/2026-10-17/factsheet.md; disclosures and broker news from pages/report_disclose.html, pages/report_topix.html.
• Price move on 2026-10-15 and results date: https://finance.yahoo.co.jp/quote/6590.T, https://finance.matsui.co.jp/stock/6590/index.
• Peers: https://finance.yahoo.co.jp/quote/7735.T, /6315.T, /6146.T (2026-10-16).
• Limitations: cause of the 2026-10-15 fall not found (no company disclosure; likely sector-wide); Q2 order and shipment data unknown; H1 guidance not disclosed; peers' operating margins not compared on matching definitions.
• This is research, not individualised investment advice.