Japan’s semiconductor-equipment cycle can be read with five public gauges, checked in a fixed order: SEAJ monthly billings, global chip sales, customer capital-spending guidance, company order backlogs, and inventory days at chipmakers. The single most useful habit is also the simplest: never react to one month of SEAJ data — read the three-month average, year on year, and only trust a turn when at least two of the other four gauges confirm it. Everything else in this guide is detail around that rule.
Why this matters right now: as of early October 2026, the chip complex dominates the Tokyo tape — the Electric & Precision sector, measured by its Nomura NEXT FUNDS TOPIX-17 ETF proxy (which may deviate slightly from the official sector index), was up 5.6% over the five sessions to 2026-10-07 even as it lagged the market on the day, Disco fell 6.89% in a single session, and the Nikkei Volatility Index sat at an elevated 29.20. Moves like these come and go, and most of them carry no cycle information at all. The framework below is built to outlast them.
Why Japan’s equipment names amplify the global cycle
Japan’s listed chip complex has two layers. The equipment layer — Tokyo Electron (8035), Advantest (6857), SCREEN (7735), Disco (6146) and peers — holds structurally high global share in specific process steps: coating and developing, cleaning, test, dicing and grinding. The materials layer — photoresists, silicon wafers, specialty gases — is even more concentrated. Because chipmakers cut and restore capital spending far more violently than end demand for chips actually swings, these companies are a leveraged expression of the global cycle: a modest change in chip demand becomes a large change in equipment orders.
That leverage shows up in valuations and volatility. Tokyo Electron closed at 12,805 yen and Advantest at 41,850 yen on 2026-10-06 — both many multiples of their prior-cycle trough prices — and single-day moves of several percent are routine. Disco, for example, fell 6.89% in one session in early October 2026 on only 1.1x its 20-day average volume. The framework exists precisely so you can tell which of those moves carry cycle information and which are rotation noise.
The five-gauge dashboard
| Gauge | Source and cadence | What it tells you |
|---|---|---|
| 1. SEAJ billings | Semiconductor Equipment Association of Japan, monthly | Coincident read on equipment spending passing through Japanese vendors |
| 2. Global semiconductor sales | WSTS industry statistics, monthly | The demand side; historically leads equipment by roughly two to four quarters |
| 3. Customer capex guidance | Foundry, memory and logic earnings calls, quarterly | Forward intent — what customers plan to spend |
| 4. Order backlogs | Equipment-maker disclosures, quarterly | How much of that intent is contracted |
| 5. Inventory cycle | Chipmaker balance sheets, quarterly | Where in the cycle you are — the most reliable turning-point signal |
Gauge 1: SEAJ billings — the spine of the framework
SEAJ (the Semiconductor Equipment Association of Japan) publishes monthly sales of Japanese-made semiconductor manufacturing equipment, including exports. Three reading rules:
- Use the three-month moving average, year on year. Equipment revenue is recognized in lumps — a single fab acceptance can swing one month’s figure by double digits. Month-on-month changes in the raw series are mostly noise.
- Watch the second derivative, not the level. Historically, cycle turns announce themselves when the year-on-year rate crosses zero or when a deep double-digit decline starts shrinking. A high absolute level with decelerating growth is late-cycle; a depressed level with improving comparisons is early-cycle.
- There is no official book-to-bill anymore. SEAJ’s headline series is billings (completed sales), not bookings, so the classic book-to-bill ratio must be proxied with company-level order and backlog disclosures — Gauge 4 below.
Gauge 2: Global chip sales and the inventory cycle
WSTS (World Semiconductor Trade Statistics) monthly global semiconductor sales are the demand side of the equation. The key is the sequencing: chip sales recover first, chipmaker inventories then normalize, capex guidance follows, and equipment billings turn last. The classic four-phase read:
- Early upcycle: chip sales accelerating, inventory days falling. Equipment orders will follow — this is when equipment shares historically do best.
- Mid-cycle: sales and capex both rising, inventories stable.
- Late cycle: sales still growing but decelerating, inventory days creeping up while capex plans are still expanding. The most dangerous phase, because the headlines are still euphoric.
- Downcycle: inventory liquidation, order cancellations, SEAJ billings in year-on-year decline.
The most common analytical mistake is treating equipment orders as a demand signal. They are a supply response — a lagging echo of chip demand, filtered through customer psychology. When equipment orders boom while chip inventories at customers are rising, the boom is borrowing from the future.
Gauge 3: Customer capex and backlog quality
A handful of foundry, memory and advanced-logic customers drive most equipment orders, so their quarterly capex guidance is a direct leading input. Read it alongside the equipment makers’ own disclosures and ask three questions of any backlog number: Is it cancellable or contractually firm? Does it reflect pull-ins (customers ordering early to secure slots or front-run export-control changes), which inflate today’s backlog at tomorrow’s expense? And how long are lead times — a backlog covering several quarters of revenue is normal in this industry, so the backlog’s direction matters more than its size. For the domestic capex picture, the Bank of Japan’s quarterly Tankan survey of corporate investment plans is the structural cross-check.
Confirming the cycle with Tokyo positioning data
Fundamentals set the trend; positioning data tell you how crowded it is. Three Tokyo-specific series, all free and official: the TSE daily short-selling ratio (karauri — short sales as a share of total trading value, published by JPX); standardized margin balances from Japan Securities Finance, splitting leveraged retail longs (shinyo margin buying) from stock borrowed for shorting; and gyaku-hibu — the premium charge short sellers must pay per share per day when borrowable stock runs scarce, a direct price on short-squeeze risk.
Worked example: reading one real print
Here is an actual figure and how to read it step by step. Japan Securities Finance data for Tokyo Electron (8035), as of the 2026-10-05 session: stock-lending balance (the short side) 29,100 shares, up 24,100 on the day; margin-loan balance (the leveraged long side) 841,300 shares, down 115,900.
- Scale first. The short-side balance jumped several-fold in one day — but at 29,100 shares it is still barely 3% of the 841,300-share margin-long balance. Small base, big percentage move: note it, do not panic over it.
- The long side is the signal. Margin longs fell by roughly 12% in a single session. Leveraged retail holders were taking profits.
- Add price context. The Electric & Precision sector proxy had risen 5.6% over the five sessions to 2026-10-07. Longs reducing into strength is a positioning reset, not distress selling. The same pattern appeared in Advantest (6857): short-side lending up 23,000 shares, margin longs down 11,300.
- Cross-check breadth. The TSE short-selling ratio printed 38.0% on 2026-10-06 — the bottom of its four-session range of 38.0% to 41.1% — so market-wide shorting pressure was easing even as single-name hedges went on. Meanwhile the Nikkei Volatility Index sat at 29.20, its 79th percentile over roughly the prior three and a half years: hedging demand elevated, but no panic.
Conclusion from the print: positioning in the chip complex was cooling after a strong run, while the cycle gauges — SEAJ trend, customer capex, inventories — remained the arbiter of direction. Against that backdrop, Disco’s one-day 6.89% drop on near-average volume reads as rotation inside a crowded sector, not a cycle signal. That is the entire point of the framework: it tells you which moves to ignore.
Common failure modes
- Reading one month of SEAJ data. Acceptance-based revenue recognition makes single months meaningless. Three-month average, year on year, always.
- Ignoring double-ordering. In shortages, customers over-order across multiple vendors to secure delivery slots. Reported demand at the peak overstates true demand, and the subsequent cancellation wave overstates the bust.
- Forgetting the currency. SEAJ billings are reported in yen while most end demand is dollar-based. With the yen around 158 per dollar as of early October 2026, yen-denominated growth rates flatter the underlying unit trend whenever the yen weakens year on year. Sanity-check against dollar-based global sales.
- Treating export-control pull-ins as organic demand. Policy-driven front-loading of shipments to any one geography inflates current billings and hollows out future quarters.
- Conflating share prices with the cycle. Equipment stocks typically trough while billings are still falling and peak while they are still rising. If you wait for the data to confirm, the market has usually moved first — which is why the leading gauges (inventories, capex guidance) outrank the coincident one (billings).
FAQ
Does SEAJ data lead or lag the stock prices?
It lags them. Billings are coincident with industry capex, while share prices discount the cycle roughly two to four quarters ahead. Use SEAJ to classify which cycle phase you are in, not to time entries — the leading inputs for timing are chipmaker inventory days and customer capex guidance.
What replaced the book-to-bill ratio?
Company-level disclosure. Since the industry associations’ headline series cover billings rather than bookings, practitioners reconstruct a demand-versus-shipment picture from quarterly order intake, backlog and lead-time commentary at the major equipment makers, cross-checked against customer capex announcements.
How do short-sale and margin data fit a fundamental framework?
As a crowding meter, not a forecast. JPX publishes the daily short-selling ratio and large short-position disclosures; Japan Securities Finance publishes margin balances and gyaku-hibu premium charges. When the fundamental gauges and positioning point the same way, trends extend; when positioning is extreme against an improving fundamental picture, squeezes happen. Neither replaces the five cycle gauges.
Why watch Japanese names instead of just the global chip leaders?
Concentration. Japanese vendors hold dominant share in specific process steps and materials, so their orders aggregate demand from every major chipmaker worldwide. They are a cleaner read on total industry capex than any single customer — and their Tokyo-listed shares come with unusually rich official positioning data to confirm what the cycle gauges suggest.
Sources
- SEAJ — Semiconductor Equipment Association of Japan, monthly billings statistics: www.seaj.or.jp/english
- WSTS — World Semiconductor Trade Statistics, global semiconductor sales: www.wsts.org
- JPX statistics — short selling, margin trading and investor-type data: www.jpx.co.jp/english
- Japan Securities Finance — standardized margin balances and premium charge (gyaku-hibu) rates: www.jsf.co.jp/english
- Nikkei Indexes — Nikkei 225 and Nikkei Volatility Index: indexes.nikkei.co.jp/en
- Nomura Asset Management — NEXT FUNDS TOPIX-17 sector ETFs, used here as sector performance proxies: nextfunds.jp
- Bank of Japan — Tankan survey and corporate capex plans: www.boj.or.jp/en
- Ministry of Finance Japan — weekly cross-border securities flows: www.mof.go.jp/english
Disclaimer: This is an information and analysis publication, not investment advice. See our Methodology for data sources, standards, and our corrections policy.