← Back to 52-Week Low Lens

PLAID INC

Company Overview

PLAID is a Japan-listed customer-experience software company. Its core product, KARTE, helps enterprises collect first-party customer data from websites and apps, analyze behavior in real time, and personalize communications across channels such as web, app, email, chat, and LINE. In plain terms, it sits in the layer between customer behavior and customer-facing action.

The latest clean official annual base is FY2025. More recent data is partial and unaudited: a reviewed half-year filing through 2026-03-31 inside a 15-month transition year after the company changed its fiscal year-end from September to December. That matters here because the headline annual FY2026 guidance is for 15 months, not a clean 12-month comparable period.

Growth has been driven by two concrete things. First, enterprise account expansion: customers with more than ¥10m of ARR rose from 222 at Q2 FY2025 to 246 at Q2 FY2026. Second, higher revenue per account: monthly subscription revenue per customer rose from ¥949k at Q4 FY2023 to ¥1.127m at Q4 FY2024 and ¥1.347m at Q4 FY2025. That is land-and-expand, not just logo accumulation.

This is not meaningfully different from FY2025 P/E because the business is light on capital expenditure. The bigger gap is period choice, not capex: TTM earnings are weaker than FY2025 because H1 FY2026 included heavy hiring and acquisition-related integration while reported profit fell.

Capital efficiency is better judged through incremental economics than through reported ROIC alone. ROE moved from about 11% in FY2024 to about 28% in FY2025. Reported ROIC screens very high in FY2025, but that figure is flattered by the company’s light asset base and working-capital structure. The more useful fact is that ARR, large-account count, and revenue per customer all improved without a corresponding explosion in total customer count. That suggests incremental commercial investment has earned good returns so far.

How the Company Makes Money

PLAID makes most of its money from recurring software subscriptions. The core pricing logic is not seat-based. It is tied mainly to customer activity, especially MAU, and in some modules to data volume. That is important because it lets revenue scale with the customer’s digital usage and complexity, not merely with software seats.

In H1 FY2026, product revenue was ¥6.28bn out of total revenue of ¥7.77bn. Service revenue was ¥1.40bn and other revenue was small. So this is still primarily a software business, even if services are becoming more visible. FY2025 also ended with subscription revenue at 82.9% of sales, down only modestly from 84.8% at Q4 FY2023.

The real economic engine is enterprise land-and-expand. Total customer count moved from 638 at Q4 FY2023 to 659 at Q4 FY2024 and then 635 at Q4 FY2025. On the surface that looks flat. But average monthly subscription revenue per customer rose sharply, and customers with more than ¥10m ARR kept increasing. That means PLAID is deliberately trading breadth for depth.

Profits actually come from the product layer, not from one-off consulting. The services layer matters because it helps customers implement KARTE, widen usage across departments, and convert data into outcomes. That can be good if services deepen software attachment. It becomes bad only if software growth starts depending on labor-heavy delivery. So far, the numbers still say the product is the center of gravity.

The best evidence that this has been a good business is pricing power plus retention. The company implemented a price increase on core products from January 2025. After that, management still reported NRR of 108.8% at Q2 FY2025, 107% at Q1 FY2026, and 109% at Q2 FY2026. Q3 FY2025 revenue churn was described as broadly flat. That does not prove an untouchable moat, but it does prove customers were not fleeing after a price increase.

Why the Stock Fell

Near-52-week-low context. Using the 2026-06-11 close of ¥513, the stock is down about 61% from its 52-week high of ¥1,309 and only modestly above the 52-week low of ¥495. Roughly ¥33bn of market value has been erased from the peak.

The decline is easy to explain in plain English. PLAID finally showed strong audited FY2025 profitability, then immediately told the market it would lean back into hiring, marketing, and service capability expansion during a messy 15-month transition year. After that, the next two official periods showed the same pattern: good revenue growth, weaker reported profit.

Q1 FY2026 revenue grew 15.9% year on year, but operating income fell 26.1% and net income fell 31.2%. H1 FY2026 revenue grew 18.6%, but reported operating income fell 30.9% and net income fell 51.5%. Management emphasizes adjusted operating profit, but the market clearly cared more about reported earnings and cash conversion. So do I.

The company also acquired CloudFit for ¥1.0bn, took on more debt, and saw its equity ratio fall from 49.3% at FY2025 year-end to 41.6% at H1 FY2026. None of that is existential, but it reinforced the market’s fear that PLAID may be becoming a lower-quality, more service-heavy, more investment-hungry story just as investors are de-rating software names that cannot show clean operating leverage.

What the Market Is Assuming

• (a) One-time / cyclical / sentiment-driven factors
  • The 15-month transition year makes comparisons messy and easier to distrust.
  • Rolling quarter-by-quarter guidance suggests limited visibility.
  • The market no longer rewards software companies for adjusted profit narratives; it wants reported earnings and cash flow.
• (b) Medium-term business headwinds
  • Hiring and marketing expense may keep growing faster than revenue for several quarters.
  • CloudFit and PLAID ALPHA may push revenue mix toward services and dilute software-quality margins.
  • Enterprise deal cycles are long; new pipeline may not convert fast enough.
  • The drop in contract liabilities and negative H1 operating cash flow may signal weaker billing quality.
• (c) Potential long-term structural threats
  • AI may commoditize parts of analytics and personalization.
  • Larger suites could bundle overlapping features and pressure pricing power.
  • Tighter privacy or data-use rules could weaken the data collection layer that KARTE relies on.

Temporary or Structural?

The structural question is not whether FY2026 margins are down. They are. The real question is whether the core value-creation mechanism is breaking: data capture, product stickiness, account expansion, and pricing power.

My essence diagnosis is straightforward: there is no hard evidence yet that retention, pricing power, or enterprise expansion have broken. The wound is in reported margin, cash-timing optics, and investor confidence. The only genuine essence risk is future AI-driven commoditization. That risk is real, but current numbers do not show it happening yet.

Is the Market Wrong? By How Much?

Probably somewhat, not massively. The market is treating a transition-year investment dip as if it were already an enduring impairment to unit economics. I think that is too harsh. But because the moat is real rather than invulnerable, and because service mix is rising, this is not a screaming mispricing.

Time-as-a-moat test.

• Could you rebuild it in 2 years with today’s market cap in cash? No. You could build something technically credible, but not the installed enterprise references, internal customer workflows, partner relationships, and trust needed to displace PLAID quickly.
• Could you rebuild it in 5 years? Partly. A well-funded competitor could build a credible challenger, especially with AI-assisted development, but would still struggle to replicate data history, reference density, and switching friction across multiple departments inside enterprise customers.
• Could you rebuild it in 10 years? Yes, probably. What still blocks you is not the code. It is trust, distribution, integrations, case-study density, and the habit of customers using the platform as a control point for first-party data and action.

This means PLAID’s moat is moderate and time-based, not permanent and physics-like.

Valuation basis. This is a FY2025-audited valuation adjusted with H1 FY2026 updates. I do not use management’s adjusted operating profit as the primary earnings base because stock compensation is a real expense and acquisition-related costs matter economically.

The bridge is simple: operating value equals normalized owner earnings divided by required yield, then I add roughly ¥3.4bn of net cash from H1 FY2026. On that framework, my base intrinsic value is about ¥25.3bn, roughly ¥4.1bn above the reference market cap of ¥21.2bn. That is about ¥99 per share of mispricing. This is an intrinsic value estimate, not a price target.

Another way to frame it: after stripping out net cash, the market is paying about ¥17.7bn for the operating business. Against my base normalized owner earnings of ¥1.2bn, that implies a roughly 6.8% operating owner-earnings yield. I think 5.5% is more appropriate if retention and enterprise expansion remain intact.

Key Facts, Estimates, and Judgments

Moat & mispricing score: 6/10. PLAID has a real moat, but not an impregnable one. The market is right to worry about AI, service-mix drift, and a more expensive go-to-market model, but the current operating data still shows strong retention, strong enterprise expansion, and intact pricing power after a price increase. What the market is probably getting wrong is treating the FY2026 transition-year margin dip as if it were already permanent. The mispricing looks modest rather than huge: enough to matter, not enough to excuse sloppy underwriting.

If I am wrong, it will most likely be because AI compresses PLAID’s differentiation faster than current NRR and enterprise-account data imply, or because services become a permanent crutch rather than a software amplifier. If I am right, the market has confused a messy transition year with a broken business.

CoffeeAnd — 52-week low lens