Screened: 2026-05-28 | Price at screen: JPY 7,770 (last close May 27) Thesis framing: Specialty batch-furnace deposition WFE; KKR spin from Hitachi Kokusai (2017); TSE IPO Oct 2023; trough-of-cycle margins; HBM batch furnace bull case
| Metric | Value |
|---|---|
| Current price | JPY 7,770 (May 27 close) / JPY 8,124 (intraday May 28) |
| 52-week high | JPY 8,279 |
| 52-week low | JPY 2,607 |
| Distance from 52W high | -6.1% (essentially at 52W high) |
| YTD return | +31.1% (JPY 5,928 Jan 5 -> JPY 7,770) |
| IPO price (Oct 2023) | ~JPY 1,840 |
| Return since IPO | +322% |
| Market cap | JPY 1,897B (~EUR 11.6B at 163 JPY/EUR) |
Trip wire: YTD 31.1% does NOT exceed 80% threshold, so full screen applies. But note the stock is within 6% of its 52W high after a +322% rerate from IPO — this is the functional equivalent of a "peak" for Graham screening purposes.
Kokusai Electric is the global leader in vertical batch thermal processing equipment for semiconductors — specifically diffusion furnaces, oxidation systems, LPCVD, and ALD. The company claims 70-80% global market share in vertical batch thermal, a product class where batch economics (processing 100-150 wafers simultaneously) are structurally advantaged over single-wafer tools for certain applications.
Key products: Batch diffusion/oxidation furnaces, ALD systems, LPCVD. Used in DRAM, HBM, NAND, logic nodes.
Corporate history: Hitachi Kokusai Electric semiconductor equipment division -> KKR LBO in 2017 -> renamed Kokusai Electric -> TSE IPO October 2023 at JPY 1,840. KKR has fully exited as of May 2026 (confirmed in news: "KKR completed exit from Kokusai Electric," May 23 2026).
Fiscal year: March 31 year-end (FY2025 = April 2024 - March 2025).
HBM (High Bandwidth Memory) stacks require precise thermal oxidation and nitridation at each bonding layer — vertical batch furnaces are the tool of choice due to wafer-level uniformity requirements and throughput. As HBM generations progress (HBM2E -> HBM3 -> HBM3E -> HBM4), layer count increases, multiplying the thermal processing steps per wafer stack.
Why Kokusai specifically:
Assessment of bull case vs current price: The stock has already re-rated from JPY 1,840 (IPO, Oct 2023, at ~11x trailing P/E) to JPY 7,770 (May 2026, at ~51x trailing P/E). The HBM batch furnace narrative drove this 4x re-rating. The thesis is not wrong — it is already priced in and then some. The question is not "will HBM benefit Kokusai?" (yes) but "does JPY 7,770 compensate you for the execution risk, China risk, and cycle risk?" At 51x trailing P/E the answer is no.
Based on company segment disclosures and analyst coverage (Nikkei, Bloomberg, Morgan Stanley Dec 2025 upgrade):
Risk: If China WFE spending softens OR US export controls expand to batch thermal equipment, Kokusai's recovery reverses sharply. At 40-50% China exposure, this is an existential revenue risk, not a marginal one.
| Metric | Value |
|---|---|
| Shares outstanding | 233.6M |
| Float shares | 219.8M (94.1% of SO) |
| KKR overhang | RESOLVED — KKR fully exited May 2026 |
| Avg daily volume (10-day) | 9.15M shares |
| Daily turnover (% of float) | 4.2% — very liquid |
Assessment: The initial concern about thin float is unfounded. With 219M share float and 9M daily volume, this is a highly liquid mid-cap on the TSE. The KKR overhang that depressed the stock in late 2025 (lock-up expiry Sep 2025) is fully resolved.
FX consideration for EUR portfolio: DEGIRO does support TSE trading, but EUR 3,000 (~JPY 489,000, ~63 shares) incurs FX conversion costs, JPY settlement risk, and adds EUR/JPY volatility to an already high-beta position (beta = 2.25). This is a meaningful execution friction for a small satellite.
IMPORTANT shr-017 WARNING UPFRONT: Kokusai's BALANCE SHEET passes several Graham filters because the company has been deleveraging aggressively post-KKR LBO (D/E dropped from very high levels at IPO to 0.31x now). The VALUATION filters fail catastrophically. This is the classic shr-017 dynamic: screening on trailing balance sheet data during a semi capex upcycle produces false positives. 7735.T (SCREEN Holdings) is the cited peer that "already failed Graham at peak" — Kokusai is in the same position.
| # | Criterion | Graham Threshold | Kokusai Value | PASS/FAIL |
|---|---|---|---|---|
| 1 | Adequate size | >EUR 100M mkt cap | EUR 11.6B | PASS |
| 2a | Current ratio | >=2.0x | 2.09x | PASS (marginal) |
| 2b | LT Debt < Working Capital | LT Debt < WC | LT Debt JPY 46.9B < WC JPY 90.2B | PASS |
| 2c | D/E | <1.0x | 0.31x | PASS |
| 3 | Earnings stability | Profitable 10 consecutive years | IPO 2023, only 4yr public data; all profitable FY2022-FY2025 | INCOMPLETE/FAIL |
| 4 | Dividend record | 20 consecutive years | IPO Oct 2023, no 20yr history | FAIL |
| 5 | Earnings growth | >33% per 10 yrs | EPS FY2022: 215.3 -> FY2025: 152.5 = -29% (declining cycle); only 4yr data | FAIL |
| 6 | P/E ratio | <=15x trailing | 51x trailing / 40.6x forward | FAIL (catastrophic) |
| 7 | Price-to-book | <=1.5x or P/E*P/B<22.5 | P/B = 8.65x; P/E*P/B = 548 | FAIL (catastrophic) |
Graham Score: 3/7 PASS (size, current ratio, D/E) — but the 3 passes are structural/balance sheet artifacts of LBO deleveraging, not evidence of a value situation. The 4 fails include both valuation metrics at extreme levels.
The shr-017 rule states: "High Graham screen scores on trailing data can be cyclical traps for companies entering a downcycle."
Kokusai's cycle position:
| Metric | FY2022 Peak | FY2024 Trough | FY2025 | FY2026E |
|---|---|---|---|---|
| Revenue (JPY B) | 245.4 | 180.8 | 238.9 | ~250+ |
| Operating income (JPY B) | 70.6 | 30.7 | 51.3 | ~64 |
| EPS (JPY, diluted) | 215.3 | 93.7 | 152.5 | 191.3 |
| Trailing P/E | ~11x (IPO) | N/A | 51x | 40.6x (fwd) |
Revenue is at 97.4% of prior cycle peak. The recovery is 90% complete. This is NOT a trough entry — it is close to cycle peak entry. And unlike FY2022 when the stock was priced at 11x trailing P/E (Graham-ish), it is now priced at 51x trailing P/E on essentially the same revenue base. This is definitionally the shr-017 trap.
Capex is also surging aggressively: FY2022 JPY 3.3B -> FY2025 JPY 27.7B (+8x in 3 years). This is both a signal of expansion confidence AND a FCF depressant — FCF dropped from JPY 70.3B (FY2022) to JPY 10.8B (FY2025) despite higher revenue, entirely due to capex.
Trailing EPS: JPY 128.21 (note: yfinance trailing EPS; company reports diluted EPS JPY 152.5 for FY2025 — using FY2025 reported) Forward EPS (consensus): JPY 191.26
Using FY2025 diluted EPS = JPY 152.5 (trailing) and forward EPS JPY 191.3:
| Growth (g) | IV (trailing EPS 152.5) | MOS vs JPY 7,770 | IV (fwd EPS 191.3) | MOS vs JPY 7,770 |
|---|---|---|---|---|
| g = 0% | JPY 1,296 | -83% | JPY 1,626 | -79% |
| g = 3% | JPY 2,208 | -72% | JPY 2,773 | -64% |
| g = 5% | JPY 2,818 | -64% | JPY 3,538 | -54% |
| g = 7.5% | JPY 3,581 | -54% | JPY 4,495 | -42% |
| g = 10% | JPY 4,345 | -44% | JPY 5,451 | -30% |
| g = 15% | JPY 5,871 | -24% | JPY 7,364 | -5% |
| g = 20% | JPY 7,398 | -5% | JPY 9,278 | +19% |
Break-even growth rate:
shr-003 implied growth gap:
Assessment: Even under the most optimistic forward scenario (g=20% perpetual), the stock is barely at fair value. Graham's formula was designed for mature, slow-growth companies where 10-15% is a high growth assumption. Requiring 16-26% perpetual growth to justify current price puts this squarely in growth stock territory — and growth stock that must sustain WFE cycle-peak economics indefinitely.
| Metric | Value | Notes |
|---|---|---|
| EV/EBITDA | 33.7x | Expensive for cyclical |
| EV/Revenue | 8.0x | High for equipment maker |
| Dividend yield | 0.62% | JPY 48/sh; trivial |
| Payout ratio | 28.8% | Conservative but yield is irrelevant |
| FCF yield | 0.57% | Minimal due to capex surge |
| ROE | 14.5% | Solid but not exceptional |
| Beta | 2.25 | Very high; 2x market volatility |
| Gross margin | 42.6% | Strong for WFE; structural moat |
| Operating margin | 21.5% | Below FY2022 peak of 28.8% |
| Current ratio | 2.09x | Passes Graham barely |
| D/E | 0.31x | Post-LBO deleveraging complete |
| Metric | Value |
|---|---|
| N analysts | 10 |
| Recommendation | Buy (mean 1.6/5) |
| Mean price target | JPY 7,549 |
| High PT | JPY 10,000 |
| Low PT | JPY 3,900 |
| Current price vs mean PT | -2.8% BELOW mean PT (price already above analyst consensus!) |
Morgan Stanley upgraded Dec 2025, citing AI-driven WFE recovery as "firmly underway." The upgrade is already reflected in the stock price — current price trades AT the analyst consensus mean. The stock is not "underfollowed" at 10 analysts; the WFE/HBM thesis is fully discovered.
This check applies if the stock were ever to reach a Graham-attractive price. For current price, this section confirms the verdict.
| Category | Status | Detail |
|---|---|---|
| Price vs 52W high | RED | -6.1% from 52W high — near peak |
| YTD momentum | RED | +31.1% YTD — momentum buyer territory |
| Valuation (trailing P/E) | RED | 51x vs Graham 15x threshold |
| Valuation (P/B) | RED | 8.65x vs Graham 1.5x threshold |
| Graham IV break-even | RED | 16-26% perpetual growth needed |
| Analyst mean PT | RED | Price already above consensus mean PT |
| China revenue risk | RED | 40-50% China, export control escalation risk |
| Cycle position | RED | Revenue at 97% of prior peak, P/E at 4x IPO level |
| EPS trend | AMBER | Recovery underway but not back to peak; FY2026 looks ok |
| Balance sheet | GREEN | D/E 0.31x, current ratio 2.09x, net debt minimal |
| FCF trend | RED | FCF decimated by capex surge (JPY 70.3B -> JPY 10.8B) |
| Float/liquidity | GREEN | 219M shares, 9M daily volume, KKR overhang cleared |
Red flag count: 8 RED, 1 AMBER, 2 GREEN — this is not an entry scenario at any level.
User instruction: avoid duplicating NVDA/AVGO/TSM/ASML/AMAT/LRCX/KLAC.
Kokusai Electric is NOT a direct duplicate of these names — it is a Japanese pure-play on batch thermal equipment, not covered by major ETF WFE baskets at high weight. However, the economic exposure is correlated with ASML (litho), AMAT (CVD/ALD), and LRCX (etch/ALD) — all benefit from the same HBM/AI capex cycle. Adding Kokusai would ADD WFE beta to the portfolio, not diversify it.
Current portfolio (AGN.AS, RI.PA, AIG, AKE.PA) has zero semiconductor exposure. Adding a high-beta (2.25) WFE play at cycle peak and near-52W high would be a directional bet on the AI capex cycle continuing, not a Graham value add.
Kokusai Electric fails 4 of 7 Graham defensive filters on the metrics that matter most — valuation (P/E 51x vs threshold 15x, P/B 8.65x vs threshold 1.5x), earnings history (IPO 2023, only 4 fiscal years), and dividend record (no 20-year history possible). The 3 passes are balance sheet metrics that reflect post-LBO deleveraging, not a value situation.
This is a textbook shr-017 scenario. The financial statements look reasonable (profitable, delevered, recovering margins), but:
Using forward EPS JPY 191.3 with Graham formula and requiring 20% margin of safety:
| Growth assumption | IV | 20% MOS entry | 33% MOS entry |
|---|---|---|---|
| g = 5% | JPY 3,538 | JPY 2,831 | JPY 2,371 |
| g = 7.5% | JPY 4,495 | JPY 3,596 | JPY 3,011 |
| g = 10% | JPY 5,451 | JPY 4,361 | JPY 3,652 |
A realistic WFE company with genuine batch thermal moat could justify g=7.5-10% long-term if the HBM thesis proves durable beyond the current cycle. Watchlist trigger: JPY 3,500-4,000 (requires -55% to -50% drawdown from current price).
At JPY 3,500-4,000, Kokusai would be priced at approximately:
Per shr-004 (Growth Stock Addendum), a stock failing all Graham filters can still pass if implied growth (from current P/E) <= achievable growth. Here: