Excluded from today's screen — already covered in the last 7 days.
| Company | Researched on |
|---|---|
| NET PROTECTIONS HLDGS INC (7383) | 2026-03-22 |
| MEDLEY INC (4480) | 2026-03-23 |
| RAKUS CO LTD (3923) | 2026-03-24 |
| NIPPON SOUKEN CO LTD (5840) | 2026-03-27 |
| ORACLE CORP JAPAN (4716) | 2026-03-26 |
| PAL GROUP HOLDINGS CO LTD (2726) | 2026-03-27 |
| FIXSTARS CORPORATION (3687) | 2026-03-28 |
| Company | Opportunity | Core moat damage | Rationale |
|---|---|---|---|
| WRITEUP CO LTD (6580) | 2 | 8 | The trust/compliance moat is hit directly by allegations in a low-switching-cost market; partner friction and regulatory risk make the payoff strongly downside-skewed. |
| COVER CORPORATION (5253) | 5 | 4 | Brand and IP remain strong, but talent churn and merchandise write-downs weaken network density and cost leverage; upside exists, yet asymmetry is not clean. |
| AQUALINE LTD (6173) | 1 | 9 | Ad-regulation, partner attrition, and governance issues structurally damage both demand aggregation and supply density; liquidity dependence makes downside compounding. |
| NIHON KOHDEN CORP (6849) | 7 | 2 | Installed-base switching costs, service coverage, and regulatory credibility remain intact; current pressure looks mostly cyclical and timing-driven, with bounded franchise risk. |
| FIXSTARS CORPORATION (3687) Selected | 8 | 1 | Specialized performance-engineering know-how and embedded switching costs appear intact; earnings pressure looks temporary, so modest normalization can produce outsized profit recovery. |
| AIMING INC (3911) | 2 | 6 | Title wind-down and pipeline opacity expose a thin moat and shrinking scale; upside needs a new hit while current economics remain fragile. |
| SHINGAKUKAI HOLDINGS CO LTD (9760) | 1 | 8 | Closures in a shrinking student market permanently reduce local density and convenience, weakening the few real moats and leaving a poor asymmetry. |
| JAPAN HOSPICES HLDGS INC (7061) | 3 | 6 | Referral and operating moats were clearly weakened, and labor plus reimbursement risk remain structural; recovery requires several fixes before upside becomes attractive. |
| AGORA HOSPITALITY GROUP CO LTD (9704) | 6 | 2 | Property locations are intact and the guidance reset is largely normalization rather than franchise damage; labor tightness and fixed-cost leverage keep the upside only moderately asymmetric. |
| J-HOLDINGS CORP (2721) | 1 | 6 | No proven moat exists, and reset financing plus chronic losses destroy per-share optionality; even business progress would be heavily diluted. |
| ASAHI EITO HOLDINGS CO LTD (5341) | 2 | 7 | Subscale cost position and chronic losses erode channel trust and service investment; upside depends too much on external FX and input relief. |
| FIXER INC (5129) | 2 | 7 | Managed-services stickiness and channel credibility are weakening, while the SaaS moat never formed; losses and concentration make recovery highly conditional. |
| SAIKAYA CO LTD (8254) | 1 | 5 | Local brand and location assets remain, but reset warrants and listing-remediation stress create reflexive dilution and underinvestment risk that overwhelm near-term upside. |
| GMO PRODUCT PLATFORM INC (3695) | 4 | 5 | The survey moat is weakened but not broken; however, leverage and a shift toward lower-moat ad exposure limit the asymmetry until execution improves. |
| NIIGATA KOTSU CO (9017) | 3 | 6 | Route incumbency survives, but driver shortages degrade network density and reliability, while Bandai City faces a structural location challenge; upside depends on policy relief more than self-help. |
Why this company was selected: 3687 offers the best risk-adjusted asymmetry in the set: moat damage appears minimal, the core know-how and client embeddedness remain intact, and the current earnings pressure looks mainly time-based rather than structural. That creates meaningful upside if utilization, pricing, or mix normalize, without the regulatory, financing, or core-moat impairment risks that dominate most alternatives.
Fixstars is a Japanese performance-engineering company. Its core business is not generic IT outsourcing. It helps customers make compute-heavy software run faster on demanding hardware and workloads: NAND flash, autonomous-driving algorithms, GPU/AI workloads, image processing, medical and scientific computing, and some finance use cases. That specialization matters because the company's economics are much better than ordinary systems integrators.
A rough sanity check based on FY2025:
That implies an owner earnings yield of about 3.0-3.2% on the current market cap.
Is this meaningfully different from P/E? Yes. The gap comes mainly from working-capital absorption as the core Solutions business grew. Reported free cash also had a large office-related deposit outflow, but that is not sustaining capex and should not be treated as permanent economic drag.
Incremental capital appears to be earning high returns. From FY2022 to FY2025, revenue rose from JPY 6.31bn to JPY 9.62bn and net income from JPY 1.08bn to JPY 1.95bn, while fixed capital remained modest and the balance sheet moved to net cash.
Where profits actually come from is very clear. In FY2025:
So the profits come overwhelmingly from the core Solutions business, not from SaaS.
Why has this been a good business?
The caveat is equally clear: this is still a people-and-know-how business with customer concentration, not a self-propelled software annuity.
Fixstars has two operating engines.
The first and much more important one is Solution. Customers pay Fixstars to accelerate, optimize, and sometimes build software for specialized compute environments. This includes NAND flash firmware-related work, autonomous-driving algorithms, image processing, AI workloads, and other performance-sensitive systems. This is mostly project revenue, but it is not commodity consulting. The company's value proposition is that faster code lowers compute cost, improves product performance, and can make impractical workloads commercially usable.
The second engine is SaaS. This includes:
Management's strategic goal has been to use Solution work as the entry point and build more recurring revenue on top. That logic is sensible. The problem is that, so far, the SaaS segment is still too small to matter for profits and still loss-making enough to matter for consolidated margins.
The stock has not fallen because the business blew up. It fell because expectations compressed.
From its 52-week high of about JPY 2,315 in early January 2025 to around JPY 1,243 in late March 2026, the shares fell roughly 46%. That is a major de-rating.
The decline tracks two things. First, FY2026 guidance signaled a clear slowdown in profit growth: management guided only about 0.8% operating profit growth and about 17.8% net income decline, citing HQ relocation costs, double rent, higher rent, wage increases, and continued SaaS investment. Second, the first quarter of FY2026 confirmed the near-term pressure: revenue still grew 9.3%, but operating profit fell 10.1% year on year.
In plain terms, the market stopped paying for optionality and started discounting current earnings quality. Investors appear to be worried that the company is shifting from a very high-margin specialist engineering business into a lower-quality mix of labor-intensive services plus unproven SaaS bets.
Concern: "Core growth is breaking."
Reality: the numbers do not show that.
Concern: "Margins are structurally collapsing."
Reality: the consolidated margin is under pressure, but the core margin is not.
Concern: "Cash flow quality is weak."
Reality: cash generation has been lumpy, not broken.
Concern: "Customer concentration is worsening."
Reality: concentration remains a real risk, but the trend is actually improving.
Concern: "SaaS is either the next growth engine or a total value trap."
Reality: the disclosed evidence is mixed.
1. Customer concentration
2. Failure of the SaaS transition
3. Commoditization of optimization know-how by vendors/tools
There is no evidence yet that this is happening. In fact, core demand and Solution margins have kept rising. But it is the main long-term threat to watch.
4. Data-center capital allocation
Bottom line:
The current pain is mostly TIME, not ESSENCE. The core specialist engineering franchise is still growing and still highly profitable. The structural risks are real, but they are either survivable or not yet showing up in the numbers.
Assume you had the company's current JPY 41.8bn market cap in cash.
Could you rebuild a competing business within 2 years?
No. You could hire engineers and buy hardware, but you would not have the customer trust, delivery history, or embedded knowledge inside semiconductor and mobility codebases.
Could you rebuild it within 5 years?
Partially. You could build a capable AI/GPU optimization boutique and maybe buy small teams. But displacing an incumbent in long-cycle, mission-critical projects would still be hard.
Could you rebuild it within 10 years?
Probably yes, in broad outline. This is not an irreplaceable network monopoly. But even with money, time would still be needed to build trust, references, vendor relationships, and a culture of low-level optimization.
What would still block you?
The market is partly wrong on diagnosis but not wildly wrong on price.
At about JPY 41.8bn equity value and JPY 5.1bn net cash, enterprise value is about JPY 36.7bn. That is roughly:
That is not a distressed valuation. It is a valuation that already assumes quality, but not perfection.
My judgment: the market is over-penalizing temporary cost pressure and under-crediting the resilience of the core Solution franchise. But the error looks small, probably on the order of low-single-digit billions of yen, not tens of billions. The stock looks more misdiagnosed than deeply mispriced.
The moat is real, but it is narrower than a software platform moat and wider than a normal IT staffing moat. The market is getting the business mix wrong: it is treating consolidated margin pressure as if the core franchise had weakened, when the actual damage sits mostly in elective SaaS investment and temporary cost pressure. But the market is not getting the price dramatically wrong, because the stock still implies only a modest owner-earnings yield and leaves little room for SaaS disappointment or customer losses. This is a decent business with mostly non-structural issues, but only a moderate, not dramatic, mispricing.
CoffeeAnd — 52-week low lens