Numbers

The Numbers

Figures converted from JPY at historical FX rates — see data/company.json.fx_rates for the rate table. Ratios, margins, and multiples are unitless and unchanged.

Hamamatsu Photonics is a structurally moaty, AA-/A+ rated Japanese photonics franchise that just lived through its worst earnings cycle in a decade — operating profit fell from a peak of $394 M in FY2022 to $109 M in FY2025, and the FY2025 7.6% operating margin sits roughly a third of its 5-year average. The market is already paying ~42× trailing earnings for the depressed result, so the single variable that re-rates or de-rates the stock is operating-margin recovery: every 100 bp of OPM mean-reversion equals roughly $14 M of operating profit, and management's own FY2026 plan only delivers a 30 bp improvement to ~7.7%. Until margins move, the AA-/Stable balance sheet, the still-rising R&D and capex, and the freshly-issued $135 M buyback are the only things underwriting the share price.

Snapshot

Stock is back at its 2017 level after a peak-to-trough drawdown of ~70% from the May 2023 split-adjusted high of $26 (¥3,760 at the FY2023 rate). The 1-year total return of +73% looks heroic only against a $7 (¥1,122) April 2025 low — measured from FY2023, the stock is still down materially in a market where photonics peers benefitting from AI-optical demand have re-rated explosively.

Quality scorecard — durable, but not cheap-on-quality

The franchise is unambiguously durable — top-tier domestic credit, 70%+ equity ratio, never lost money over the cycle, no public bonds outstanding. But the shape of that quality has shifted: in FY2022 Hamamatsu sat on $850 M of cash with essentially no debt; by FY2025 short-term borrowings have ballooned to $362 M to fund both the $305 M NKT Photonics acquisition and $135 M of treasury cancellations, and the equity ratio has dropped 7 points in a single year.

Revenue and earnings power — the cycle in one picture

Revenue dropped only 4% from the FY2023 peak ($1,484 M → $1,433 M), but operating profit was cut by 71% over the same window. Two years of mix deterioration — falling PMT shipments to NIH-funded medical-bio customers, intensifying Chinese price competition in silicon photodiodes, and the NKT Photonics integration pulling the laser segment to a $30 M operating loss — hollowed out unit economics while gross margin slipped to its lowest since FY2020. (Yen-USD figures partly mask the swing: a weaker yen across FY2022–FY2025 made revenue look stable in dollars while it grew in yen, then masks how much earnings have actually fallen.)

Quarterly cadence — has the bottom been seen?

The 3Q25 operating profit of $10 M was the worst quarter in the dataset and produced a $2 M net loss — the only loss-making quarter in nine. 4Q25 recovered to $26 M, but 1Q26 (the company's seasonally light quarter) printed only $15 M operating profit on $331 M sales, missing consensus by 55%. Sequential margin recovery is real but slow; FY2026 guidance of $116 M operating profit (at FY26 assumed ¥148/USD) implicitly requires the second half to print $70–80 M, comparable to early FY2024.

Cash generation — earnings are real, even when they collapse

Cash conversion is the bullish counterweight to the P&L story: 5-year CFO averages $284 M against $214 M average net income — a 131% conversion that survived the FY2025 earnings collapse (CFO $255 M held essentially flat at the same level it printed in FY2024 of $267 M). The damage instead shows up in FCF: capex stepped up from ~$116 M (FY21) to ~$209–217 M (FY23–FY24) — a doubling of intensity that crushed FCF to $21 M and $50 M in those two years. FY2025 capex is undisclosed in the kessan tanshin but new building/structures growth implies another year above $200 M.

Capital allocation — a buyback chapter has begun

The capital allocation profile has changed character in two short years. FY2024 swallowed $305 M for NKT Photonics — the largest acquisition in company history — financed entirely with new debt and on-hand cash. FY2025 launched the company's first material buyback ($135 M, with ~5% of float retired), and a second $128 M (¥20 B) authorisation runs Nov 2025 through Sept 2026. Dividends have been held flat at $0.26/share since FY2023 even as the payout ratio climbed to 80.3% in FY2025 — management has formally adopted DOE (dividend-on-equity) as a floor to signal payout stability through earnings cycles.

Per-share economics — split-adjusted EPS round-trip

EPS has done a complete round-trip: the FY2025 $0.32 result is below FY2017's $0.50, eight years on. Book value per share moved from ~$5.36 (FY2017) to $7.27 (FY2025) — a 3.9% per-share book accumulation in dollars (the underlying yen book grew at 7.5%, with USD compression reflecting the JPY weakness over the period). Dividends per share have been frozen at $0.26 (post-split equivalent) for three years; the FY2025 payout ratio of 80% is the highest in the dataset and signals management is funding the dividend out of stretched earnings rather than cutting it.

Balance sheet — equity ratio breaks 70% for the first time

The balance sheet story is the most visible structural change in this period. Net cash sat above $545 M for seven consecutive years, peaked at $851 M in FY2022, and has now been drawn down to $135 M over two fiscal years through a combination of acquisition spend, peak-cycle capex, and the first material buyback. The equity ratio breaking below 75% for the first time in a decade is not yet an alarm — at AA-/Stable with no public bonds, leverage has plenty of headroom — but it does mean the balance sheet is no longer the asymmetric option it once was.

Valuation vs its own history — the critical chart

Catalog Error: Table with name pe_history does not exist!
Did you mean "pg_description"?

LINE 2: ..., pe, 31.9 as eight_yr_mean, 30.65 as eight_yr_median from pe_history
                                                                      ^

This is the chart that frames the trade. At the FY2022/2023 peak, the market priced Hamamatsu at 23×, almost a third below its long-run 30–33× band — investors were already pricing cyclical reversion. They got it. Now, on collapsed earnings, the trailing P/E is ~42×, a multiple that only makes sense if you believe FY2025 was the trough and EPS reverts toward the $0.92 band of FY2022/23. On normalized earnings the stock is ~15×, well below its history. The reader's view on this stock is essentially their view on whether Hamamatsu's 25%+ operating margin era was a one-off windfall (post-COVID semi capex) or its true earnings power.

Returns on capital — a step down, not a collapse

ROE compressed from a 16.0% peak in FY2022 to 4.4% in FY2025 — below management's stated cost-of-equity floor and the lowest in the dataset. Even FY2020's COVID trough was 8.0%. The depressed return reflects the same numerator problem visible in earnings, plus a denominator that has stayed near record-high equity (only mildly reduced by the buyback).

Photonics peers — Hamamatsu has lagged the AI-optical re-rating

The most important number in the table is the +1,428% one-year return on Lumentum and the +397% on Coherent. Photonics names with direct exposure to 800G/1.6T optical transceivers for AI datacenters were re-priced from "secular decline" to "AI infrastructure play" in twelve months. Hamamatsu's photonics product mix barely intersects with that demand pool — its industrial-segment AI exposure is the silicon photodiodes for semiconductor inspection equipment, not the transceivers themselves — and its 1-year return of +73% reflects a recovery off the April 2025 low rather than participation in the AI-optical theme. Keyence (also a sensor name with no AI-optical exposure) returned +32%, which is closer to the realistic comp.

Analyst landscape and the fair-value range

The bear-base-bull range ($7.50 — $12.45 — $21.54) widens because the equity is essentially a leveraged bet on margin recovery. Sell-side has already capitulated below $13 as the realistic 12-month anchor. The bull case requires not just "FY26 guide hits" but a return to the FY2022 unit economics (27% OPM) — which would re-rate the stock 70% from here even on a more conservative 25× multiple.

What the numbers say — a closing read

The numbers confirm the moaty franchise narrative: 70%+ equity ratio, AA-/Stable since 2018, 9 unbroken years of profits, R&D held at 8.7% of sales even as earnings collapsed, and CFO that proved indifferent to the P&L blow-up — a $260 M cash machine that kept turning. The numbers contradict the popular "Japan-blue-chip-with-defensive-margins" framing: operating margin has just round-tripped to a 7.6% level not seen since the early 2010s, ROE has fallen below cost of equity, the balance sheet has absorbed both a debt-funded acquisition and a debut buyback, and the once-pristine net-cash position has shrunk 84% in two years. The single thing to watch into FY2026 is the segment-by-segment OPM trajectory in 2Q26 and 3Q26 results: if Electron Tube and Opto-semiconductor margins claw back even 300–400 bps each on a recovery in semiconductor inspection demand and lower China price pressure, the bull thesis re-arms; if they don't, the $9 sell-side targets become the real anchor and the buyback becomes a slow-motion floor rather than a re-rating catalyst.