← Back to 52-Week Low Lens

ORACLE CORP JAPAN

────────────────────────
1. Company Overview
────────────────────────
Oracle Corporation Japan is the listed Japanese operating arm of Oracle. It sells Oracle’s database software, ERP and other enterprise applications, cloud services, support contracts, consulting, and some hardware into Japan. This is not Oracle Corp. in the U.S.; it is the Japan-focused sales and support entity, and its economics are shaped both by Japanese enterprise IT demand and by its dependence on Oracle’s global product stack and parent-company decisions.

- Economics (core numbers only):
• Market cap: about ¥1.16 trillion
• Net cash / (net debt): about ¥66.6 billion net cash at FY2025 year-end, excluding a separate ¥182.0 billion long-term loan receivable from group companies/affiliates that I would not treat as cash-equivalent
• Net income (TTM): about ¥62.8 billion
• P/E: about 18.5x on that TTM estimate; quoted trailing P/E is about 19.1x; normalized P/E is also about 19x

- Growth:
• Revenue CAGR: about 7.1% over FY2022–FY2025, from ¥214.7 billion to ¥263.5 billion
• Net income / EPS CAGR: about 5.9% over FY2022–FY2025, from ¥51.2 billion to ¥60.7 billion
• What is actually driving growth?
1. Cloud Services & License Support, the recurring core, grew from ¥157.8 billion in FY2024 to ¥174.4 billion in FY2025, up 10.5%
2. Ongoing migration of Japanese enterprises from on-premise systems toward Oracle cloud and hybrid deployments, while legacy support renewals remain sticky

- Owner earnings (sanity check, not theory):
• Net income: about ¥60.7 billion in FY2025
→ – sustaining capex: roughly ¥2.0 billion
→ ± working capital: roughly neutral over a cycle; FY2025 cash flow was helped by timing, so I would not capitalize that
= Owner earnings: roughly ¥59–61 billion
• Owner earnings yield: roughly 5.1%–5.3% on current market cap
• Is this meaningfully different from P/E? If yes, why?
Not really. This is a very asset-light business, so capex is tiny. Reported FY2025 free cash flow of ¥64.6 billion was somewhat above net income, but that does not change the core picture much.

- Capital efficiency:
• ROIC / ROE: ROE is roughly 35%–40%; ROIC is clearly very high, likely well above 50% on a rough basis, though the related-party loan asset muddies precise invested-capital math
• Is incremental capital earning high returns?
Yes. Operating profit rose from ¥72.3 billion in FY2022 to ¥86.8 billion in FY2025 while annual capex remained only about ¥0.5–2.0 billion. This is a low-capital, high-return model.

- Business quality (only what matters):
• Where do profits actually come from?
Mostly from Cloud Services & License Support. It was 66.2% of FY2025 sales and almost certainly a larger share of profits than that. This is the annuity engine.
• Why has this been a good business?
The moat is switching cost and embedded mission-criticality. Large Japanese companies do not casually rip out databases, ERP, or core middleware. Support contracts recur, migrations are expensive and risky, and Oracle’s brand, certification ecosystem, and parent IP matter. The important nuance is that the moat is real, but part of it is borrowed from Oracle headquarters rather than owned independently by the Japan subsidiary.

────────────────────────
2. Why the Stock Is Near a 52-Week Low
────────────────────────
The stock is around ¥10,825 versus a 52-week high of ¥17,875 and a low of ¥8,763. That means it is down about 39% from the high and much closer to the low than the high.

The decline makes sense if you view it as the market unwinding an earlier AI- and Oracle-parent-related enthusiasm trade. The Japanese subsidiary never had AI-hypergrowth economics. It had a high-quality, mid-single- to high-single-digit enterprise software model. Once FY2026 started with a softer first quarter, margin pressure, and less exciting guidance than bullish holders wanted, the stock rerated down hard.

Investors appear worried about three things at once: first, that growth has already peaked; second, that margins will keep getting squeezed by rising royalties and personnel-related costs; and third, that the parent-company relationship is a structural governance discount rather than a source of strength, especially after large related-party lending disclosures.

────────────────────────
3. What the Market Is Currently Pricing In
────────────────────────
(a) One-time / cyclical / sentiment-driven factors
- The stock had run too far on Oracle-parent AI enthusiasm.
- FY2026 EPS guidance of ¥490–¥505 was below bullish expectations.
- The FY2026 year-end dividend was left undecided.
- FY2026 1Q operating profit fell 4.8%, which triggered de-rating.
- Broader software and AI monetization fears hurt sentiment across Oracle-related equities.

(b) Medium-term business headwinds
- Revenue growth may settle into a 6%–10% range, not an AI-style growth rate.
- Operating costs are rising faster than investors expected.
- Hardware and services are weaker businesses than recurring software support.
- Balance-sheet cash is not as clean as it first appears because of related-party lending.

(c) Potential long-term structural threats
- Minority holders do not control capital allocation; the parent does.
- Oracle global could change economics, transfer pricing, or contract flow in a way that reduces value capture in Japan.
- Open-source and cloud-native alternatives could slowly weaken Oracle’s historical switching-cost moat in new workloads.

────────────────────────
4. Reality Check vs Market Narrative
────────────────────────
- Concern: “Growth has stalled.”
Reality: Revenue has risen for four straight fiscal years: ¥214.7 billion in FY2022, ¥226.9 billion in FY2023, ¥244.5 billion in FY2024, and ¥263.5 billion in FY2025. The latest nine-month FY2026 result was ¥206.7 billion, up 7.1% year on year from ¥192.9 billion. That is slower than hype, but not stall.

- Concern: “The core franchise is weakening.”
Reality: Cloud Services & License Support, the annuity core, grew from ¥157.8 billion in FY2024 to ¥174.4 billion in FY2025, up 10.5%. Its share of total sales increased from 64.5% to 66.2%. The best part of the business is getting bigger, not smaller.

- Concern: “Margins are breaking.”
Reality: Operating profit has still climbed over time: ¥72.3 billion in FY2022, ¥74.4 billion in FY2023, ¥79.8 billion in FY2024, and ¥86.8 billion in FY2025. For 9M FY2026, operating profit was ¥67.0 billion, up 4.4% year on year. Operating margin has been high and fairly stable: 33.7%, 32.8%, 32.6%, 33.0%, and 32.4% for 9M FY2026. That is pressure, not collapse.

- Concern: “Cash flow is deteriorating.”
Reality: Free cash flow was ¥52.7 billion in FY2022, ¥67.0 billion in FY2023, ¥79.8 billion in FY2024, and ¥64.6 billion in FY2025. FY2025 was down from an unusually strong FY2024, but still comfortably above FY2022. With capex only around ¥0.5–2.0 billion per year, this remains a strong cash generator.

- Concern: “The balance sheet is becoming risky.”
Reality: This is not a leverage problem. Cash and cash equivalents were ¥91.9 billion at FY2024 year-end and ¥66.6 billion at FY2025 year-end, with no obvious net debt burden. The real issue is governance quality: FY2025 also showed ¥182.0 billion of long-term loans receivable from group companies/affiliates. By 9M FY2026, the equity ratio had improved to 59.6% and net assets had risen to ¥188.1 billion. Solvency is fine; capital-allocation quality is the question.

────────────────────────
5. Structural vs Non-Structural Diagnosis
────────────────────────
- Parent-control and related-party capital allocation
- Damaged mechanism: minority holders’ claim on surplus cash and independent capital allocation
- Does this damage the core value creation mechanism?
Not directly. Customers still buy Oracle products, renew support, and migrate workloads.
- Does it weaken the moat in an irreversible way?
It weakens minority economics, not the customer-side moat.
- Could time realistically heal this within 3 years?
Not by itself. This only improves if the parent changes behavior or repays/normalizes related-party arrangements.
- Classification: (b) Real Structural but survivable

- Risk that Oracle global captures more value directly and leaves less in Japan
- Damaged mechanism: local distribution and profit capture
- Does this damage the core value creation mechanism?
Potentially yes, if contracts or economics are redirected.
- Does it weaken the moat in an irreversible way?
It would hurt the listed subsidiary’s economics, but there is no evidence that this has happened yet.
- Could time realistically heal this within 3 years?
If it happened, only the parent could reverse it. But current operating data do not show actual damage.
- Classification: (c) Not truly structural

- Slow erosion of Oracle lock-in from open-source and cloud-native alternatives
- Damaged mechanism: renewal moat and switching costs
- Does this damage the core value creation mechanism?
Yes, if customers increasingly move core systems off Oracle.
- Does it weaken the moat in an irreversible way?
Yes, if migration becomes common. But current numbers do not show that.
- Could time realistically heal this within 3 years?
If a large installed-base migration wave began, time would not heal it quickly. But that wave is not visible today.
- Classification: (c) Not truly structural

Bottom line: the only clearly structural issue visible today is governance and capital allocation under parent control. That matters, but it is not the same as business-franchise decay.

────────────────────────
6. Time-as-a-Moat Test
────────────────────────
Assume I had the company’s current market capitalization, about ¥1.16 trillion, in cash.

- Could I realistically rebuild a competing business within 2 years?
No. I could hire a sales force and build a respectable enterprise-software reseller or services company, but I could not recreate Oracle’s installed base, product rights, certifications, or mission-critical trust inside large Japanese enterprises.

- Could I realistically rebuild a competing business within 5 years?
Still no, not on equivalent economics. I might build a niche competitor in databases, cloud infrastructure, or applications, but not a business with comparable recurring support revenues and embedded customer dependence.

- Could I realistically rebuild a competing business within 10 years?
I could build a meaningful competitor in some segments. I still could not rebuild Oracle Japan’s exact position without Oracle’s global IP and decades of installed mission-critical systems.

- What would still block you?
Oracle’s underlying product IP and compatibility
Installed base in core databases and ERP
Customer switching costs and migration risk
Enterprise trust, certifications, and partner ecosystem in Japan
Deep integration into customer workflows and compliance environments

The important conclusion is that the moat is real, but it is not purely local. A large part of it sits in Oracle global’s IP and the customer lock-in around that IP.

────────────────────────
7. Moat & Mispricing Score
────────────────────────
Score: 6/10

The market is too negative on near-term margin wobble and too quick to read normal mid-single-digit growth as franchise decay. The numbers still show a strong recurring core, high ROE, low capital intensity, and continuing revenue and profit growth. But the market is not irrational to apply a permanent discount for parent control and related-party lending; that is real structural baggage for minority holders.

In yen terms, the stock implies an owner earnings yield of roughly 5.1%–5.3% on about ¥59–61 billion of owner earnings. For a debt-light, recurring-software franchise growing around 6%–7%, a required yield closer to 4.8%–5.0% would be reasonable, but only after keeping a governance discount in place. That suggests equity value more like roughly ¥1.20–1.27 trillion, or about ¥40–110 billion above the current market cap. If you treat the related-party loan as very low-quality and apply a harsher governance discount, the stock moves closer to fair value. The market is somewhat wrong, but not dramatically wrong.

CoffeeAnd — 52-week low lens