CDS CO LTD

Company Overview

CDS Co., Ltd. is a small Japanese manufacturing-support company. It helps customers design products, create technical manuals and digital content, build factory-automation and robot systems, and run product-development and IT systems. It is best understood as a niche engineering-and-information-services contractor with a strong balance sheet, not as a pure software company.

Data freshness is important here. The latest clean official annual base is FY2025, ended 2025-12-31 and filed 2026-03-30. More recent data is partial and unaudited: Q1 FY2026 was disclosed on 2026-05-13. Current market data in the supplied record are as of 2026-07-14.

Using only supported figures: current market capitalization is ¥11.66 billion at ¥1,710 per share (current market data, 2026-07-14). FY2025 revenue was ¥8.83 billion, operating income ¥685 million, and net income ¥456 million (audited annual data). The supplied deterministic block marks net cash as unavailable; audited FY2025 cash was ¥4.46 billion and disclosed borrowings were roughly ¥0.62 billion, so leverage is plainly low, but I will not replace the supplied net-cash field with my own derived figure. The supplied FY2025 owner earnings of ¥1.700 billion are a model estimate, not an audited fact.

The latest clean official annual base is FY2025. More recent data is partial, delayed, unaudited, or estimated. In this case, the more recent official data are unaudited Q1 FY2026 results.

How the Company Makes Money

CDS has three operating engines. The first is Technical Information Solutions: technical manuals, product documentation, design support, translation, e-learning, videos, and related digital content. The second is FA Robot Solutions: factory-automation systems, robot applications, and educational equipment for schools and industrial training. The third is Digital Solutions: PLM, MBD, CAD/CAE-related systems, IT infrastructure, embedded software support, maintenance, and related consulting for manufacturers.

In FY2025, the segment mix mattered. Technical Information Solutions produced roughly ¥3.49 billion of sales and ¥843 million of operating profit, making it the best-margin business at about 24%. Digital Solutions was the largest revenue segment at roughly ¥4.40 billion of sales and ¥513 million of operating profit, or about 11.5% margin. FA Robot Solutions was much smaller and more volatile, with roughly ¥0.94 billion of sales and only ¥45 million of operating profit, or about 4.8% margin. That tells you a lot about the economics: CDS is not one business, and the group result is highly sensitive to mix.

The company is asset-light in accounting terms but not in economic terms. FY2025 capex was only about ¥211 million (audited annual data), so plant intensity is low. The real capital is people, customer-specific know-how, and workflow integration. That supports decent margins when utilization is high, but it also makes the business fragile when customers pause programs. This is why reported cash flow can look excellent in a downturn if receivables shrink, even while true earning power weakens.

The durable parts of the model are real but limited: long customer relationships, embedded knowledge of technical documentation and engineering workflows, and some switching friction around PLM/MBD and on-site support. But this is not a true high-return compounder. Incremental capital mostly buys more people and more customer exposure, not proprietary network effects or a software toll bridge. A well-run niche contractor can be durable; it is different from a business with deep, self-reinforcing moat economics.

Why the Stock Fell

The stock fell because earnings power broke, not because the balance sheet broke.

The first clear shock was the 2025-08-08 guidance cut (official company guidance revision). Full-year FY2025 guidance was reduced from revenue of ¥10.788 billion to ¥9.259 billion, operating income from ¥1.538 billion to ¥939 million, and net income from ¥1.058 billion to ¥628 million. That was not a small trim. It was a direct admission that major customers were cutting development budgets and delaying work.

The second shock was the FY2025 result released on 2026-02-13 and later filed on 2026-03-30. Revenue fell 15.9%, operating income fell 54.5%, and net income fell 56.8% versus FY2024 (audited annual data). The damage was concentrated where the market least wanted to see it: Digital Solutions and FA Robot Solutions. Technical Information was down only modestly. That pattern said the issue was not broad accounting noise. It was customer spending weakness in the more cyclical and customer-concentrated parts of the group.

The third shock was the 2026-05-13 Q1 FY2026 disclosure (unaudited quarterly data). Revenue fell 19.7% year on year to ¥2.104 billion, operating income fell 48.4% to ¥172 million, and net income fell 47.4% to ¥118 million. Importantly, the consolidated decline came from Digital Solutions: technical information and FA were roughly flat to slightly better, but Digital Solutions revenue fell 36.6% and operating profit fell 69.4%. So the market narrative became simpler and harsher: CDS is overexposed to a weak customer and a weak spending pocket.

The concentration issue is not theoretical. In the FY2025 annual filing, sales to Mitsubishi Motors were ¥2.86 billion, down from ¥3.87 billion in FY2024. That roughly ¥1.01 billion drop explains about 61% of the company’s total FY2025 revenue decline. I do not have a verified long price series in this record, so I am not attaching an exact peak-to-trough stock decline. The dated business events above are enough to explain why sentiment deteriorated.

What the Market Is Assuming

The market is not pricing CDS as if it were near liquidation. At the supplied market value, the shares trade at about 1.31x FY2025 book and, on my calculation using current market data and company FY2026 guidance, about 17.6x guided FY2026 EPS. That means the market assumes FY2025 was a trough year and that the business can recover meaningfully from here.

• One-time or cyclical factors: Some of the FY2025 damage does look temporary. Technical Information Solutions held up relatively well in FY2025, with revenue down only 0.8%, and Q1 FY2026 showed slight growth and better profit there. FA Robot Solutions was hit by delayed school projects and weak manufacturing capex, which is painful but not obviously permanent.
• Medium-term headwinds: Digital Solutions is the real problem. In FY2025 its revenue fell 20.3% and operating profit fell 43.5%. Q1 FY2026 got worse, not better. That makes the market’s recovery assumption dependent on a digital rebound that is not yet visible in reported numbers.
• Long-term threats: Customer concentration and limited moat are the structural concerns. The top three customers accounted for 39.1% of FY2025 sales, and Mitsubishi Motors alone was 32.4% (audited annual data). Management’s March 2026 business briefing also showed that automotive still represented 76.3% of digital-segment sales (management update). This is not a diversified software annuity.

The best multi-period quantitative check is the order book. In FY2025, total orders were only 71.2% of the prior year and ending backlog was only 68.1% of the prior year (official annual filing data). The split is revealing: Technical Information orders were 97.0% of prior year and backlog was 104.2%; FA orders were 62.4% and backlog 73.2%; Digital orders were 57.6% and backlog 56.6%. So the market is effectively assuming that the weakest segment stabilizes before weak order data turns into another bad year.

There is one more subtle assumption embedded in the stock: that the expiration of the Mitsubishi-related service-level volume floor is manageable. The FY2025 annual report states that the minimum order-volume protection under MCOR’s service contract ran through 2026-03-31. As of the market date in this report, that date has passed. I do not have newer official disclosure on replacement terms. That evidence gap matters.

Temporary or Structural?

The cleanest diagnosis is this: the earnings collapse is mostly TIME, but it revealed an ESSENCE problem. CDS is not structurally broken in the sense of balance-sheet impairment or irreversible franchise collapse. But it is also not a deep-moat business temporarily misunderstood by the market. The downturn exposed how dependent the group is on a few customers and on discretionary engineering and IT budgets.

So the key distinction is this: the recent earnings damage is repairable, but the company’s business quality is lower than a quick screen suggests. The market’s fear is not irrational. CDS can recover without becoming a much better business. That is why “temporary problem” does not automatically mean “cheap stock.”

Is the Market Wrong? By How Much?

Moat-and-mispricing score: 4/10. The balance sheet is strong enough to limit permanent-loss risk, but the moat is modest and the stock already prices in a fair amount of recovery.

The supplied deterministic block gives two very different messages. Worth-dead says there is a real asset cushion: NCAV is ¥5.624 billion, or 48.2% of the current market cap. That is meaningful downside support. Worth-alive says FY2025 owner earnings were ¥1.700 billion, implying a 14.6% owner-earnings yield and a payback under seven years. I do not accept that at face value. FY2025 operating cash flow was boosted by a large receivables unwind as revenue fell; that is cash release from a shrinking working-capital base, not clean recurring earning power.

My valuation therefore uses normalized owner earnings rather than the raw FY2025 screen output. I use 6,819,639 shares, based on the latest official shares outstanding excluding treasury stock.

The strict asset floor is lower still: NCAV works out to roughly ¥825 per share (model estimate based on FY2025 balance sheet and current market-date share count). That matters because it limits existential downside. But the key investment question is not whether CDS is solvent. It is whether the market is paying too much for recovery in a low-moat, concentrated service business.

If I mechanically capitalized the supplied deterministic owner earnings of ¥1.700 billion at a 10% yield, I would get about ¥17.0 billion of equity value, or roughly ¥2,490 per share. I reject that result. FY2025 free cash flow was flattered by working-capital release as receivables fell sharply. In a contracting service business, that kind of cash flow is often a harvest signal, not proof of durable earning power.

My conclusion is that the market is not treating CDS like a cigar butt. It is valuing the company closer to a successful recovery case. That may still work if digital spending normalizes and the Mitsubishi dependency eases. But based on the current evidence, I think the stock is roughly fairly priced to somewhat optimistic, with the base-case intrinsic value sitting about 20% to 25% below the current market value.

Key Facts, Estimates, and Judgments

• Strongest falsification check for a cautious view: if H1 and H2 FY2026 show genuine Digital Solutions recovery, with revenue and margins improving without another large working-capital release, then my normalized earnings base is too low.
• Evidence that would improve the thesis materially: official disclosure that post-2026-03-31 Mitsubishi-related business remains secure, plus a continued reduction in customer concentration and a stabilization of digital orders/backlog.
• Evidence that would worsen the thesis materially: another guidance cut, continued digital deterioration after the contract-floor expiry, receivables rebuilding while profits stay weak, or a dividend cut caused by weak operating performance rather than a deliberate capital-allocation reset.