← Back to 52-Week Low Lens

CYBOZU INC

Company Overview

Cybozu is a Japanese software company that sells collaboration and workflow tools, almost entirely as cloud subscriptions. Its flagship product is kintone, a no-code/low-code business application platform used to build internal workflows, databases, approvals, and reporting apps without heavy custom coding. The company also sells Cybozu Office for SMB groupware, Garoon for larger organizations, and Mailwise for shared email management. The important point for investors is that Cybozu is no longer a transitional software vendor; it is now a mostly recurring-revenue SaaS business with a domestic Japanese core.

Fact: Revenue rose from ¥15.7bn in 2020 to ¥37.4bn in 2025. Net income rose from ¥1.4bn to ¥7.1bn over the same period, although the path was lumpy because Cybozu invested heavily in 2021-2022. Concrete growth drivers were only two. First, kintone kept expanding into larger accounts; by June 2025 it was used by 47% of TSE Prime companies, up from 44% at the end of 2024. Second, Cybozu implemented a cloud price revision from November 2024, lifting ARPA across products without visible churn damage.

This is not meaningfully different from the P/E-based earnings yield. That is what you would expect from a subscription software business where depreciation roughly tracks maintenance infrastructure needs and annual billing provides some working-capital support. In 2025, operating cash flow was ¥10.7bn and free cash flow was ¥7.6bn, both stronger than net income.

Capital efficiency is high. Reported ROE was 31.1% in 2024 and 48.1% in 2025. Estimate: true operating ROIC is harder to calculate precisely because Cybozu carries excess cash and benefits from favorable subscription working capital, but it is clearly well above 20% and likely above 30%. Incremental capital is earning high returns: from 2023 to 2025, revenue increased by about ¥12.0bn while operating cash flow increased by about ¥6.1bn, with no meaningful financial leverage.

How the Company Makes Money

Cybozu now makes its money mainly from cloud subscriptions, which were 92.1% of 2025 sales. The economic engine is kintone, which produced ¥21.7bn of sales in 2025, or roughly 58% of total company revenue. The other major products were Cybozu Office at ¥6.8bn, Garoon at ¥6.2bn, and Mailwise at ¥1.1bn. On-premise and other revenue is now small enough that it matters mainly as cash flow, not as the growth story. Profit does not come from overseas operations; the core profit pool is domestic Japanese cloud software, especially kintone.

The business has been good for reasons that are concrete rather than abstract. Customer workflows get embedded inside kintone as custom apps, approvals, forms, and data structures, which creates real switching costs. Cybozu also has a broad partner channel: the official ecosystem includes 500+ partners, which matters in Japan because software adoption is often tied to implementation support, SI relationships, and local trust. Pricing power also appears real. kintone’s gross revenue churn was 0.94% at the end of 2024 and 0.92% at the end of 2025 even after the price revision. Net revenue retention for kintone improved from 109.4% to 121.9%, although part of that jump is price-driven. Cybozu also runs its own cloud infrastructure rather than being purely a wrapper on third-party IaaS, which helps cost control, though it also requires ongoing server investment.

Why the Stock Fell

The stock fell because expectations were reset, not because the reported business collapsed. Cybozu traded as high as ¥4,160 in August 2025 and later touched a 52-week low of ¥1,990 in March 2026. That is roughly a 52% decline from high to low. The key fact is that this happened after a year in which revenue rose 26.1% and net income nearly doubled. So the market was not reacting to weak trailing numbers. It was re-pricing what those numbers mean.

In plain English, investors appear to believe that 2025 overstated Cybozu’s real earnings power. The November 2024 price increase made 2025 look exceptionally strong. Then management guided 2026 revenue growth down to 12.7% and operating profit growth to only 4.1%, while openly saying it would increase hiring, advertising, and data-center spending. That combination told the market two things: first, the 2025 surge was not a clean new base rate; second, management would spend part of the windfall rather than let margins keep expanding. Add fears about AI competition and skepticism around overseas expansion, and the multiple compressed hard.

What the Market Is Assuming

(a) One-time / cyclical / sentiment-driven factors

• 2025 was flattered by the November 2024 cloud price revision.
• The stock had run too far ahead of fundamentals before the slowdown guide arrived.
• Investors are extrapolating a post-price-hike growth comedown into a broader de-rating.

(b) Medium-term business headwinds

• 2026 revenue growth slows from 26.1% to 12.7% as price-revision comparisons normalize.
• Margin expansion stalls because management plans heavier spending on hiring, ads, and infrastructure.
• Growth may be more ARPA-driven than seat-growth-driven.
• Overseas investment, especially in the U.S., is still loss-making and may absorb cash without attractive returns.

(c) Potential long-term structural threats

• AI-enabled workflow tools from Microsoft, Salesforce, Google, and others could weaken kintone’s differentiation.
• Cybozu is still heavily concentrated in Japan.
• The business is increasingly concentrated in one product, kintone.
• The higher minimum-seat structure may narrow the small-customer acquisition funnel over time.

Temporary or Structural?

Diagnosis: this is mostly TIME, not ESSENCE. The market is treating a mathematically inevitable post-price-hike slowdown as though it proves deterioration in the core engine. The core engine is recurring domestic workflow software with low churn, positive expansion, high cash conversion, and no balance-sheet fragility. That engine is intact.

The only serious structural threat is competition from AI-enabled workflow platforms with much larger ecosystems. But that threat is still a future possibility, not an observed damage pattern. What is happening now is a reset from a temporarily elevated growth and margin trajectory.

Is the Market Wrong? By How Much?

Time-as-a-moat test. If I had Cybozu’s current market cap in cash, I could certainly build a competing product. What I could not quickly rebuild is the installed base, trust, and partner ecosystem.

• Within 2 years? No. You could build a decent low-code workflow tool, but not one trusted by 70,000+ cloud customers, 3.6m user licenses, and a domestic ecosystem of 500+ partners. The blockers are trust, local enterprise distribution, switching costs from embedded workflows, and implementation capacity.
• Within 5 years? You could build a credible niche competitor. You still would not easily replicate Cybozu’s domestic brand, its installed workflow base, or its position inside Japanese enterprise and public-sector style deployments. The partner network would still matter.
• Within 10 years? Possibly, yes. Software moats are rarely permanent. But even in 10 years the hard parts would still be customer habit, workflow migration pain, SI/channel depth, security trust, and the local product-market fit Cybozu has accumulated since its cloud transition.

So the moat is real, but it is not magical. It is stronger than the current price implies, yet weaker than a global platform monopoly.

Judgment: the market is somewhat wrong, but not massively wrong. At today’s price, investors are paying roughly 13x earnings for a net-cash, mostly recurring-revenue software business with low churn, high returns on capital, and still positive double-digit sales growth. That is too cheap if the right lens is “2025 was partly boosted by pricing, but the moat remains intact.” It is not obviously cheap if the right lens is “AI competition will commoditize low-code workflow tools faster than the market expects.”

My own reading is closer to the first lens. If Cybozu deserves an equity owner-earnings yield of roughly 6.5% to 7.0% rather than the current 7.8% to 8.0%, the gap is roughly ¥10bn to ¥20bn of market value. That is a meaningful mispricing, but it is not a heroic one. The market is discounting too much essence into what is still mostly a time problem.

Key Facts, Estimates, and Judgments

Moat & Mispricing Score: 7/10. Cybozu’s moat is real: sticky internally built workflows, low churn, strong domestic brand trust, and a 500+ partner ecosystem. The market is correctly recognizing that 2025 was helped by a price revision and that 2026 growth will slow. What it is getting wrong is the leap from slower to damaged. ARR, churn, enterprise penetration, cash flow, and leverage do not show a damaged core. The stock looks undervalued by roughly ¥10bn to ¥20bn, not because the business is perfect, but because the market is pricing it as if the moat has already started to crack.

• Facts: 2025 revenue was ¥37.4bn, net income was ¥7.1bn, cloud revenue was 92.1% of sales, kintone revenue was ¥21.7bn, operating cash flow was ¥10.7bn, free cash flow was ¥7.6bn, year-end cash was ¥11.7bn, and there was no meaningful debt. Kintone churn stayed at 0.92% through 2025, and TSE Prime penetration rose from 44% to 47% in H1 2025.
• Estimates: owner earnings are roughly ¥7.4bn to ¥7.6bn; true operating ROIC is likely above 30%; current equity owner-earnings yield is roughly 7.8% to 8.0%; fairer yield is roughly 6.5% to 7.0%.
• Judgments: the issue is mostly TIME, not ESSENCE. The main real structural risk is AI/hyperscaler competition, but current operating data do not yet show damage. Overseas losses matter for optionality, not for the core thesis. Executive stock-based compensation exists, but the disclosed cap is small relative to earnings and is not thesis-changing.

CoffeeAnd — 52-week low lens